Paul Kunst, Jos Lemmink, Bernd Stauss Auth., Paul Kunst, Jos Lemmink, Bernd Stauss Eds. Service Quality and Management

Kunst/Lemmink/Stauss (Eds.
)
Service Quality and Management
GABLER EDITION WISSENSCHAFT
Focus Dienstleistungsmarketi ng
Herausgegeben von
Universitätsprofessor Dr. Michael Kleinaltenkamp,
Freie Universität Berlin {schriftführendl
Universitätsprofessor Dr. Dr. h.c. Werner Hans Engelhardt,
Ruhr-Universität Bochum,
Universitätsprofessor Dr. Anton Meyer,
Ludwig-Maximilians-Universität München,
Universitätsprofessor Dr. Hans Mühlbacher,
Leopold-Franzens-Universität Innsbruck und
Universitätsprofessor Dr. Bernd Stauss,
Katholische Universität Eichstätt
Der Wandel von der Industrie- zur Dienstleistungsgesellschaft ist de

facto längst vollzogen, er stellt jedoch mehr denn je eine Heraus-
forderung für Theorie und Praxis, speziell im Marketing, dar. Die
Schriftenreihe will ein Forum bieten für wissenschaftliche Beiträge
zu dem bedeutenden und immer wichtiger werdenden Bereich des
Dienstleistungsmarketing. In ihr werden aktuelle Ergebnisse der
betriebswirtschaftlichen Forschung in diesem Bereich des Marketing
präsentiert und zur Diskussion gestellt.
Paul Kunst/Jos Lemmink/
Bernd Stauss (Eds.)
Service Quality
and Management
Springer Fachmedien Wiesbaden GmbH

Die Deutsche Bibliothek - ClP-Einheitsoufnohme
Service quality and management / Paul Kunst ... (eds.). -

wiesbaden: Dt. Univ.-Verl.; Wiesbaden: Gabler, 1999
(Gabler Edition Wissenschaft: Focus Dienstleistungsmorketing)
Alle Rechte vorbehalten
© Springer Fachmedien Wiesbaden 1999

Ursprünglich erschienen bei Betriebswirtschaftlicher Verlag Dr. Th. Gabler GmbH,
Wiesbaden, und Deutscher Universitäts-Verlag GmbH, Wiesbaden, 1999.
Lektorat: Ute Wrasmann
Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede
Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist
ohne Zustimmung des Verlage.~ unzulässig und strafbar. Das gilt insbeson-
dere für Vervielfältigungen, Ubersetzungen, Mikroverfilmungen und die
Einspeicherung und Verarbeitung in elektronischen Systemen.
http://www.gabler-online.de
http://www.duv.de
Höchste inhal~iche und technische Qualität unserer Produkte ist unser Ziel. Bei der Produktion und
Verbreitung unserer Bücher wollen wir die Umwelt schonen. Dieses Buch ist deshalb auf säure-
freiem und chlorfrei gebleichtem Papier ,gedruckt. Die Einschweißfolie besteht aus Polyäthylen
und damit aus organischen Grundstoffen, die weder bei der Herstellung noch bei der
Verbrennung Schadstoffe freisetzen.
Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in diesem

Werk berecntigt auch ohne besondere Kennzeichnung nicht zu der Annahme, daß solche Namen
im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei zu betrachten wären
und daher von jedermann benutzt werden dürften.
ISBN 978-3-8244-6780-8 ISBN 978-3-322-90981-7 (eBook)

DOI 10.1007/978-3-322-90981-7
Table of Contents
Table of Contents
pages
Introduction 1-3
Section I:
Service Quality and Satisfaction
Dynamics in Service Quality Perceptions - A Longitudinal

Approach to Evaluate the Outcome Dimension in Service
Quality Perceptions
Claudia Klausegger and Dieter Scharitzer 5-24
National Customer Satisfaction Indices:

A Critical Investigation from an Application Perspective
Ursula Hansen and Thorsten Hennig-Thurau 25-53
The Link between Customer Satisfaction and Dealer

Satisfaction - The case of the German Car Industry
Alexander Decker 55-88
Section 11:
Measuring Service Quality by Applying the Criticallncident
Technique
Criticality of Critical Ineidents in Customer Relationships

Bo Edvardsson and Tore Strandvik 89-110
Critical Ineidents in Dutch Consumer Press: Why Dissatisfied

Customers Complain with Third Parties
Jan A. Schulp 111-159
Section 111:
Assuring and Improving Service Quality by Applying Quality
Instruments and Methods
Achieving Service Quality through the Application of

Importance-Performance Analysis
Robert Govers and Frank M. Go 161-185
The Sequential Ineident Technique tor Innovations (SITI) -

A Tool for Generating Improvements and Ideas in Service
Processes
Alexander Decker and Harald Meissner 187-226
Table of Contents
pages
Section IV:
Service Quality from the Intemal Perspective
Beyond the Call of Duty: The Impact of Perceived Support

on Attitudes and Behavior of Temporary Call Center Agents
Martin Wetzeis and Ko de Ruyter 227-250
Quality Management and Organizational Change

Franlioise Chevalier 251-274
Sales-Related Benefit-Analysis of Service Quality Investments -

The Moderating Role of Customer Relationship Type
Manfred Bruhn and Dominik Georgi 275-300
List of Abstracts 301-305
Biographies 307-314
Introduction
Introduction
Are services millennium-proof? Certainly not. At least not at this moment. There is
no thorough research needed to derive that conclusion, simply ask around. The evi-
dence is overwhelming. True horror stories exist of all types of services in all types
of sectors. It is even becoming a business in itself. Television shows that are based
on customer complaints about services are becoming more and more popular. As is
the case in the millennium problem, management of service companies experiences
a lot of problems in the hardware and the software of services. There are still prob-
lems in defining and developing the service, and problems in creating, realizing and
managing weil defined services.
Is there than no progress at all? We believe there iso The enormous attention for
services has its advantages as weil. In various places innovations are realized, and
what is more important are linked to theory. Only in this way learning becomes pos-
sible. Eventually innovative practice will reflect in the development of theory, and in
turn good practice will be based on solid theory. This series tries to support this pro-
cess by presenting a number of innovative practices, and examples of testing theory
in service quality marketing and management.
Interesting ideas have been presented. Mainly a marketing perspective has been
taken, but contributions from service operations, HRM or organization theory are
included in "Part IV: Service Quality from an internal Perspective" tao. This means
that our continuous appeal to devote more time to interdisciplinary research is still
valid.
Clearly progress has been made concerning the application and implications of
measuring critical incidents. Alexander Decker and Harald Meissner developed an
instrument that deals with sequential incidents and explicitly takes into account the
link with usage of that information for developing service innovations. Bo Ed-
vardsson and Tore Strandvik investigated the criticality of critical incidents in cus-
tomer relationships. Their distinction between emotions, cognition and actions as a
basis for different layers of customer reactions excels in because of the inherent
beautiful simplicity. Jan Schulp addresses critical incidents in Dutch Consumer
2 Introduction
Press. He processed a huge amount of customer complaint data of the Dutch Con-
sumer Unions and draws some interesting conclusions: Wh at about the frequent
complaints concerning employee-customer interactions? In that respect Martin Wet-
zels and Ko de Ruyter stressed the importance of supporting and training of call
center personnel in that respect. A more traditional approach is taken by Robert
Govers and Frank Go. Their Importance-Performance analysis is not new, but gives
a clear example of the usability of that instrument. In the broader environment of
quality and satisfaction measurement, on different levels studies have been under-
taken. On the national level a discussion prevails how Customer Satisfaction Indices
have to be calculated. Examples in Sweden, United States and Germany are not
without critique. Ursula Hansen and Thorsten Hennig-Thurau investigated applica-
tions and conclude that comparability of the indices between branches are not guar-
anteed. Claudia Klausegger and Dieter Scharitzer use a longitudinal approach in the
measurement of service quality perceptions. However very strong and recommend-
able, too few researchers adopt such an approach for obvious reasons. Alexander
Decker investigates the link between Customer and Dealer Satisfaction. Indeed an
important issue since future EC regulations will definitely forbid exclusive distribu-
tion. In that respect, the relationship between dealer and car manufacturer is worth-
while studying. Manfred Bruhn and Dominik Georgi already did that and come to the
conclusion that it is important to discriminate between different types of relation-
ships. Finally, Franltoise Chevalier took a truly organizational point of view by her
views on change management and quality management. She concludes that most
change programs don't work as weil as they were supposed to because they are
guided by theoretical approaches by people, and thus usually unable to co pe with a
variety of requirements.
An important task for editors is to look for and find interesting contributions to the
field and help authors to publish their thoughts. This task is greatly supported by the
European Institute for Advanced Studies in Management (EIASM). They create the
opportunity for academics to meet, and discuss developments in their respective
disciplines. Therefore we thank Gerry van Dyck, director of EIASM and Graziella
Michelante, program manager who helped us to realize the Quality Management in
Services Workshops. Workshops need hosts. The Quality Management in Services
Introduction 3
Workshops VII and VIII had excellent hosts. We like to thank the chairpersons in
Kristiansand, Norway, Hans Chr. Garmann Johnsen and Harald Knudsen (Agder
College) and the team of the Chair of Services Management at the Catholic Univer-
sity of Eichstätt in Ingolstadt, Germany, in particular Frank Hoffmann and Alexander
Decker. Without their efforts this volume could not have been published.
In addition, we are indebted to Alexander Decker for his effort in realizing this book.
On the one hand he managed the contacts between editors and authors. On the
other hand he was in charge for bringing the various drafts of the papers to a formal
homogeneaus manuscript. By far he exceeded our expectations. Not at all an easy
job!
Paul Kunst
Jas Lemmink
Bernd Stauss
Dynamics in Service Quality Perceptions - A Longitudinal Approach
to Evaluate the Outcome Dimension in Service Quality Perceptions
Claudia Klausegger and Dieter Scharitzer
Contents
Abstract 7
Background information 7
2 Research Goals 9
3 Research Design and Hypotheses 9

3.1 Conceptual Framework 9
3.2 Empirical Design of the Survey 14
4 Preliminary Results - Hypothesis Testing 15

4.1 Results of the Cross-Sectional Survey after Traineeship 15
4.2 Longitudinal Analysis of Subjects 16
4.3 Detailed Analysis of the Dynamics of Economics 19
5 Summary and Research Outlook 21
References
Oynamics in Service Quality Perceptions 7
Abstract
Ouring the past decade much conceptual and empirical research has been done re-
garding customers' perceptions of process- and outcome-related dimensions of
services. In analyzing a number of empirical studies, we noticed that most of the re-
search designs were based on cross-sectional measurement approaches. Those
attempts usual/y neglect the potential dynamics in the customers' evaluation of the
services de/ivered over time.
This article presents the results of an empirical longitudinal project based on data
col/ected at two different points in time. In our research an ex-post evaluation of the
educational quality of an Austrian Business Col/ege was done by means of impor-
tance and performance ratings from the students' point of view at two different times
after finishing the undergraduate course.
1 Background information
Research techniques to measure process- and outcome-related dimensions of serv-
ice quality have been topics of high interest in marketing research and practice dur-
ing the past decade. Today, we can find a broad variety of instruments and methods
for the evaluation of services. Still, most of the empirical projects published focus on
a static perspective of analysis, i.e. customer satisfaction is measured only at one
certain point in time in the delivery or consumption process by means of a cross-
sectional study design. In the literature it is shown that aspects of process and time
play an important role regarding services. However, following different evaluation
phases in service quality models such as those of Meyer/Mattmüller (1987) it can be
stated that, due to the process- and outcome-related dimensions of services, not
only the service characteristics themselves, but also the customers' perceptions
during and after the service-delivery and consumption process may vary in line with
the time when the measurement takes place (see Haller 1995). As a result, customer
satisfaction research should be more focused on longitudinal study designs to cover
also aspects of time and context-related changes in the customers' quality percep-
tions empirically. In practice, longitudinal study designs are rarely used due to the
8 Claudia Klausegger / Oie/er Schari/zer
difficulties that arise in trying to find identical respondents at !wo different points in
time.
In our research quality perceptions of students of an Auslrian "Fachhochschule"

(henceforth referred 10 as Advanced Business College) have been investigated. Af-
ter havlng read quite a number of literalure on service qualily and service satisfac-
tion we found thai, due to the methodological and praclical problems of measuring
oulcome-related service qualily dimensions, most empirical papers only focus on
importance and performance indicators related 10 struclure and process elements of
the services investigated, which are known as search or experience qualities (Zeit-
haml 1981). Very few studies, however, cover outcome-related service attributes or
an ex-post analysis of service performance. The Advanced Business College under
scrutiny has applied different quality management and quality assurance techniques
concerning the recruiting and training of staff. Moreover, several feedback instru-
ments and course evaluations by means of questionnaires are done regularly during
and at the end of the courses. Till now, Ihere has, however, been no evaluation of
the oulcome dimensions of the service delivery processes. Consequently no student
ex-post evaluation of the whole undergraduate course program has been done so
far. These outcome or ex-post analyses of the educational services can reasonably
be done only after job experience in practical work has been acquired by the stu-
dents. This is due to the fact that it is hard for students to evaluale Ihe outcome-
related service dimensions of their education as long as theyare still involved in the
process of studying and have consequently, not goi practical evidence to determine
whelher their education is useful and practice-oriented. Since every single student
has to do a ten-months traineeship after finishing the undergraduate course pro-
gram, we hypothesize thai the practical experience resulting from it might cause an
important change in contextual variables so thai the sludents' perceptions of the
quality of the undergraduate course program will change over time.
Dynamics in SeNice Quality Perceptions 9
2 Research Goals
Two major goals in our research project are:
• To add a new evaluation instrument to the methods already used in the

quality-assurance program of the Austrian Advanced Business College un-
der scrutiny in the form of an ex-post analysis of students' perceptions of the
undergraduate course program before and after the traineeship.
• To investigate the dynamics in service quality perceptions due to an inter-
vention and a change in contextual variables from the first satisfaction
measurement to the second.
The longitudinal study design is considered as a useful and integral part of customer
satisfaction measurement techniques besides cross-sectional analysis. If the educa-
tion imparted by the Austrian Advanced Business College showed a lack of useful-
ness and practical applicability, we would expect changes in student ratings con-
cerning the quality of education being imparted in the undergraduate course pro-
gram and in the students' priority ratings of the subjects. At the same time the ques-
tionnaire used in the ex-post analysis of the students' perceptions would be tested
for its test-retest-reliability.
3 Research Design and Hypotheses

3.1 Conceptual Framework
In the early days of customer satisfaction research authors were mostly dealing with
the theoretical background as weil as modeling and measurement aspects of service
quality. They started a discussion of how to delimit and newly define terms already
known from the consumer behavior and customer satisfaction literature, such as
"satisfaction", "attitude" as weil as the term "service quality" which had just come up
at that time. In this context, issues as to how it was possible to measure service
quality, which were considered as a multidimensional bundle of different attributes from
the view point of external customers were investigated. It was found that it makes good
sense to distinguish between the providers' (internal) specifications of the services and
10 Claudia Klausegger / Dieter Scharitzer
the extemal view of customers, whichneed, however, not be perfectly equal (Scharitzer
1994a, 1994b). From a measurement point cf view, papers were dealing mostly with the
application of questionnaires and different types of scales (Le. expectations, per-
formance, disconfirmation, or value). The pioneer works in this field of research are
known quite weil and also practical use of different research designs has proven the
concepts and research instruments to be valid and useful. Nonetheless, today we
can state that the issue of service quality has given rise to a renaissance of cus-
tomer satisfaction in marketing research both in theory and practice.
Although there is common agreement on the fact that process and time are two of
the main characteristics to describe services (Meyer/Mattmüller 1987, Stau 55 1991),
most empirical research designs have so far neglected these essential aspects as
they are only based on a static perspective of the customers' service quality percep-
tions. It seems that in empirical work the time of data collection is set more or less
arbitrarily and no attempts are made towards longitudinal measurement designs that
would fit the underlying theoretical concepts in customer satisfaction literature better.
This is remarkable since although models cover the dynamic aspects of the custom-
ers' evaluations of services, empirical designs do not so. For example, the models of
Fisk (1981) or Meyer/Mattmüller (1987) differentiate the various stages in the con-
sumption and evaluation process from the customer's point of view. Moreover, Grön-
roos (1983) states the importance of the distinction between process- and outcome-
related dimensions in the perception of service quality.
Recently, there has been growing criticism on the conventionally applied cross-
sectional approach in service quality measurement. Several authors put forward
good arguments why more longitudinally oriented research designs should be used
for monitoring dynamics in customer perceptions towards services. Moreover, they
stress that in doing research also the contextual factors should be given more atten-
tion. The focus should, therefore, be on the following requirements
• Differentiating carefully the phases in the service production and consump-

tion process during observation.
Dynamics in Service Qua/ity Perceptions 11
• Approprialely selecling dimensions and items to adequately describe the

short glimpse of reality that is investigated from the customer's point of view
at the time when the interviews are made.
• Using appropriate questionnaires and scaling techniques to be able to cover
the dynamics in the customers' perceptions during, for example, a service
episode or transaction. Moreover, potential changes in the relevance of
certain quality dimensions and the impact of critical moments in service en-
counters as weil as how they add to agiobai judgement of service quality
should be investigated.
• Moreover, contextual factors should be given particular attention at the time
of data collection as they have an influence on the customers' expectations
and the service quality rating. Due 10 prior factors like, experience, mass or
individual communication, learning from service relations, episodes and en-
counters or context variables like socio-demographic, cultural or social influ-
ences dynamics may be assumed to exist in customer perceptions of service
quality.
Several key works may be quoted to stress the importance of the above-mentioned
requirements. Among these, for example, the conceptual works of Liljander/Strandvik
(1994) and Stauss/Seidel (1995) may be named, which however, mainly concentrate
on some detailed terminological input in order 10 account for the process character
of services. A distinclion is made between the terms relationship, transaction, epi-
sode and service encounter.
There are not many empirical results available yet, to learn more about the dynamic
aspects of service quality perceptions and customer satisfaction. Some of them that
have had an influence on the research should be mentioned here:
1) An increasing number of research projects have been done during the past
years regarding to customer satisfaction barometers or indices that monitor
different parameters of customers' expectations, service performance
evaluation and intentional behavior towards different services of an economy
at a macro-economic level. The mosl prominent research activities in this
context are known as the "American Customer Satisfaction Index (ACSI)"
(Fornell 1992, Anderson/Forneli/Lehmann 1994) and "Das Deutsche Kun-

denbarometer" in Germany (Meyer/Dornach 1995). Although the study de-
signs and measurement instruments applied differ very much from each
other, both instruments have produced sufficient data so that comparisons
could be made on the changes and dynamics in the indices. Nevertheless, it
has to be stated that this kind of market research may not be classified as a
"real" longitudinal design proper, since the respondents are selected by ran-
dom sampling and, therefore, differ from one survey to the other. So no
tracking of the intra-individual changes in a person's perception and the rat-
ing cf a service could be done.
2) Haller (1995) presents a completely different research design in her investi-
gation of the dynamics in students' evaluations of a distance learning
course. She analyzed the relation between students and their university de-
partment at three different times, stating the hypothesis that several service
encounters and episodes may effect changes in the importance and per-
formance ratings of the services provided by the university in the students'
perceptions over time.
3) Further research work has been done on how customer perception in differ-
ent service encounters is aggregated along different stages of the service
delivery and consumption process. By way of example, the paper of Dana-
her/Mattsson (1994) should be mentioned here, in which the authors investi-
gate four different service encounters during a hotel' stay by means of a se-
quential data collection design. The individual evaluation steps were carried
out immediately after the guest to be interviewed had experienced a service
encounter. In the last stage he was also asked for agiobai satisfaction rating
of the service delivered. The work of StausslWeinlich (1995) dealing with
process-oriented measurement techniques of service quality points in a
similar direction. They present an example of sequential analysis of service
encounters in a number of episodes and encounters that took place during a
stay in a holiday resort.
4) Other authors stress the importance of contextual variables that influence
customers' perceptions and ratings at the time, when data collection takes
place. So, for e.9., when Parasuraman et al (1988) refer to their measure-
Dynamics in Service Quality Perceptions 13
ment instrument SERVQUAL and propose that data collection in view of ex-
pectations and performance ratings should be carried out at one single step
after service delivery took place, the validity of the results must be heavily
criticized especially for the expectation rating (Grönroos 1993). Generally
measuring expectations and their dynamics seems to be a problem in itself.
Referring to the dynamics and the influence of process and time on the
building up of expectations, changes due to interventions or third party influ-
ences, such as by means of communication have, however, been investi-
gated already.
For practically applying the results obtained by satisfaction measurement in market

research it is essential to know about dynamics in perception processes when qual-
ity measurement from a customer's perspective and analyses regarding the individ-
ual phases of services delivery are carried out.
As a result, criticism raised about the instruments and the results of service quality
measurement is justified in so far as the commonly applied cross-sectional analyses
only relate to the time when the assessment is carried out and to its underlying con-
text. However, it is usually even possible up to a point in cross-sectional studies to
describe the dynamics assumed to be interest in the perceived quality in the form of
coefficients of change ("00 you think that company XYZ has performed beUer/worse
during the past 12 months?").
Moreover, relying on ex-post analysis for the !wo data collection stages in this proj-
ect yielded one further advantage: Usually, course evaluations by means of ques-
tionnaires are limited to process-related service quality dimensions only. There is
criticism about the fact that at that time students have no real experience regarding
the use and applicability of the subjects being taught in practice. As a result, ques-
tions referring 10 outcome-related qualities of the course program are not included in
the questionnaire as a rule. Alternatively students are asked what might be useful in
practice in their opinion.
3.2 Empirical Design of the Survey

In our study, empirical work was carried out in cooperation with an Austrian Ad-
vanced Business College. The curriculum in this type of business-education institu-
tion is made up of an undergraduate course program lasting for two years. The un-
dergraduate course is followed by a ten-month traineeship where students are sup-
posed to get job experience before they return to school for another year, with a
more specialized course program being offered in different areas of business ad-
ministration.
These service characteristics as weil as its time-table made the Advanced Business
College under scrutiny suitable for a longitudinal empirical study. By means of a da-
tabase it was possible to distribute pre-coded questionnaires to all students of an
age-group, so that we could identify and link perception ratings collected in two dif-
ferent stages of data gathering.
The sampie comprises of eighty students who finished the first two years of their
educational program in the summer of 1996 and then did their traineeship till Sep-
tember 1997. As figure 1 shows, these students had to fill in a standardized ques-
tionnaire at two different times: The first stage of data cOllection took place immedi-
ately after finishing the undergraduate course program, the second field survey was
taken in autumn 1997 after most students had done at least 10 months of their train-
eeship.
The questionnaire covered both image dimensions and items for perception ratings
concerning different subjects of the undergraduate course program. We used identi-
cai questions and scales in both data collection stages to eliminate potential biases
inherent in the instrument.
Curriculum of the
Austrian Advanced Business College
I
Undergraduate Course
(\wo years)
J L ____•• 1" stage of data eolleetJon (SeptJOet. 96):

\/ immediately after finishing the
undergraduate course program
Traineeship
(10 months)
J L ____•• 2"" stage of da/a eolleetJon (July/Aug. 97):

V after 10 months of traineeship
Specialized Course Program

(one year)
Fig. 1: Longitudinal study Analysis
4 Preliminary Results - Hypothesis Testing

4.1 Results of the Cross-Sectional Survey after
Traineeship
As far as the subjects of the undergraduate course program are concerned, the re-
sults of the second stage of data collection are quite satisfactory for the manage-
ment of the Advanced Business College. After 10 months of traineeship particularly,
the business-administration-related subjects (accounting, marketing and computer
science) were assessed highly positive.
Figure 2 shows the results of the quality portfolio from the students' point of view
(with the dimensions "perceived quality" and "perceived importance of subjects") af-
ter finishing traineeship. The students were asked to assign a rank from 1 to 10
(1 =highest priority) to each subject in such a way that the highest priority reflects the
highest preference for the subject ("In case you are a manager of the Austrian Busi-
ness College, which priority would you give to the ten subjects of the undergraduate
16 Claudia Klausegger / Die/er Schari/zer
course listed below and how many hours would you assign to each subject?"). The
perceived quality of the subjects was measured by a six-point rating scale, with the
data being taken exclusively from 2nd survey.
As figure 2 clearly shows only the subject "Iogistics" got very poor ratings. This find-
ing obtained in the ex-post evaluation confirmed the results of student evaluations
done immediately after the courses. Moreover, the poor ratings of "Iogistics" has al-
ready been discussed in some quality circles to find ways for impravement. The
other subjects of the undergraduate course received satisfactory results.
++
o English
o Computer Seienes
- o Marketing
o Economics 00 Law
~ PersonnellOrganization, .1::::::
Personafity Training Courses

:c
::J Finance 0 0 Accounting
"'
Ö ....................... ··············_······_··_······T····_···_·············.. _............_..............
.~
<ii
::J
o
-
o Materials Management & Logistics j
+/- ++
Ilmport.ne. 01 subj.cts I
Fig. 2: Quality Portfolio of the Undergraduate Course Subjects
4.2 Longitudinal Analysis of Subjects

Relying on our hypothesis that the practical experience gained during traineeship
might cause a significant change in perceiving the undergraduate course subjects in
retrospect, the evaluations of the subjects differ considerably. English and law
showed only slight changes in the perceptions fram the first to the second stage of
data collection. In contrast, the results of the other subjects exhibited a major differ-
ence and in the case of the personality training courses, personnel manage-
ment!organization, marketing and economics showed a significant difference.
Figure 3 shows the arithmetic means of the subjects for the first and second stage of
data cOllection, respectively.
Above all, courses in marketing, computer science, personnel management!

organization as weil as personality training courses seem to be assessed better as a
result of having acquired practical experience during traineeship. In contrast, we
also experienced negative changes in the students' perceptions, such as in subjects
as accounting, law, economics an finance. Especially the changes in the perceived
importance of economics are highly significant and have changed for the worse. Due
to the poor results in the second stage of data collection, a more detailed analysis of
this subject will be presented in chapter 4.3.
Before the main changes are scrutinized, another interesting detail has to be pointed
out. With regard to the importance of subjects ranking we expected the business-
administration-related subjects to be on top of the list. However, on the contrary, the
results show a high perceived importance of additional skills in languages and com-
puter science as weil as in personality training. These results are in accordance with
the desired positioning of the Advanced Business College under scrutiny, which'tries
to make up for educational deficits of the universities.
In summary the above hypothesis could be verified, which means that the assumed
dynamic factors in service quality perceptions definitely exist in this research. Ten
months of professional practice caused significant changes between the two stages
of data collection in the case of four subjects.
18 Claudia Klausegger / Dieler Scharilzer
-1 st Survey (n=57)
-2nd Survey (n=60)
Marketing·
Economics""·
Computer Selene.
Personnel Management/Org."
Personality Training Courses"I~~l'::.-=::ok~~:= ~

Logisties ...............I.........QIooI...,
12345678910
very important not important
lw.4of-'OIIIf1e1n«: •.. . O,CI$; -.-0. ,01 ; - ... CI,l)a1
Fig. 3: Importance ot Subjects trom the student's point ot view
10 ~""""""""""~~""""""""""""""""""""~""""""-'not importanl
9
a
7
6
5
4
3
2
0 ~ ........................~............................~........................~ve~ important
1st Survey 2nd Survey
-Economk:s
" Personality Training Courses
"'Marketing
OI:-. · .,o,:- . · ...Ot
)e", lof''''''''C.M U : •• • • • -Personell Management/Organisation
Fig . 4: Significant Differences in Ihe Importance ot subjects

Figure 4 once again shows selected subjects, where statistically significant changes
have taken place over aperiod of about 10 months. Using an ex-post research de-
sign that also comprises survey data gathered when the students had already got
some distance to the service experience, i.e. the service encounters they underwent
when attending courses, we could prove that the students' ratings are now more fo-
cused on the outcome dimensions, applicability of subjects being taught of the edu-
cation delivered.
4.3 Detailed Analysis of the Dynamics of

Economics
As pointed out in figure 3, economics showed the largest negative change in the
students' perceptions. In investigating the reasons for this phenomenon we found
that courses in economics showed very good marks in the course evaluations (with
an average rating of 1.9 on a 5-point rating scale, 1 =highly satisfactory). Conse-
quently the process qualities of this subject was highly rated in the first data collec-
tion stage of our study (Sept.lOct. 96). In an importance rating of different subjects,
students' interest in economics ranked fifth among all ten subjects of the under-
graduate course program. Besides the importance ranking , most of the quality di-
mensions referring to the subject of economics, such as quality of education, appli-
cability of skills, topicality of contents were rated high in the first survey of our inves-
tigation. However, the ratings turned significantly negative in the data collected after
the students had done their traineeship.
Figure 5 shows the results of the quality dimensions of economics between the two
stages of data collection in detail.
Quality of Education
very good very poor'
Applicability of Skins
very good very poor"
Aquired Competence
very good very poor"
Inte""t in Subject
very Interestlng not Interesting •
Topicality of Contents
up-to-date obsolet
Importance of Subjects
very Important not Important"
o 2 3 , 5 6
- 1st Survey (mean) ~2nd Survey (mean)
..ftlloI~ : ·. -O'osj-.. - O,01 ; - a. -O.oo1
Fig. 5: Evaluation of Economics after Survey 1 and Survey 2
One reason for the poor results may be attributed to the fact that the students dis-
covered a lack of practical relevance of the knowledge they had got during their ba-
sic education in economics. We suppose that also the kind of traineeship students
got might have an influence on the students' importance ratings of the subjects un-
der scrutiny. Investigating the detailed job descriptions we found out that most of the
jobs offered were in business-related areas with an emphasis on accounting and
marketing. As a result of students' experience in practice their satisfaction ratings,
quality perceptions and priorities towards the subject of economics came down sig-
nificantly during the ten months.
Figure 6 in addition shows, the results of the decreasing subject priority ascertained
by the question "Should the number of hours be extended, kept at the same level or
be reduced for this subject?".
Dynamics in SeNice Quality Perceptions 21
100%
90%
80%
70%
600;.
50%
400;.
30%
200;.
10%
O% +---~----~----+---~~~~----;
1st Survey 2nd Survey
0'
Lw'" J/llnlflc.nCfl:
a- 0.001
Cl reduce
• keep al the same level
Cl exlend
Fig. 6: Subject Priority of Economics
5 Summary and Research Outlook

The study undertaken focuses on the investigation of changes in the students' per-
ceptions of the quality of subjects of an Austrian Advanced Business College after a
ten-month traineeship. After having reviewed the relevant literature, we found that
most evaluations are being based on cross-sectional designs only, which means
one-time studies where da ta is gathered at a single point in time. While these stud-
ies yield only a "snapshot" of a situation being investigated, our proposed research
design produces a "motion picture" over time or at least aseries of snapshots of a
situation. The major advantage of carrying out longitudinal studies is that changes in
the perceptions of particular individuals can be studied and monitored during a pe-
riod of time. Therefore, a longitudinal study approach could prove to be more infor-
mative than a purely cross-sectional analysis. One reason why longitudinal studies
have, nevertheless, been applied only scarcely in social research apart from the
highly complex research design is that these studies are, in most cases, more ex-
pensive. In our case, cooperation with the Advanced Business College under scru-
ti ny enabled us to gather data in an inexpensive way; moreover the number of stu-
dents to be interviewed was only limited.
In our projeet it was shown that the time when data for customer satisfaetion research is
colleeted has to be chosen carefully. In our ex-post analysis cf service quality percep-
tions we were able to prove that student ratings of outcome qualities of the same re-
search objeet vary in certain aspeets when data colleeted at !wo different times is com-
pared. Therefore, we conclude that it is essential to allow for dynamics in service quality
perceptions, especially when a change in contextual variables occurs over time. We
furthermore think that it would be useful in future to investigate not only process-related
elements cf service performance; a broadening of evaluation instruments to cover also
outcome-related quality dimensions of services is highly recommended to meet the de-
mands of comprehensive service quality monitoring.
In our study, ex-post analysis of outcome-related quality dimensions of educational

services may serve as a useful instrument to continuously improve the contents of
the course program at the Advanced Business School under scrutiny. Gaps in the
praetical applicability and usefulness of certain subjects should be carefully ana-
Iyzed and further investigated from a praetical point of view. However, it should be
stated that the students' perceptions should not be the only criterion for evaluating a
course program and the praetical benefits of academic subjeets. The results shown
and the instrument used should be seen as a further opportunity 10 make improve-
ments on the way towards high er standards of educational services.
Since we analyzed several measurement techniques for course evaluation in the

Advanced Business College, we would strongly recommend using a multi-method
approach for getting feedback from the students and for improving service quality in
education. Different qualitative and quantitative research techniques should be com-
bined at several times to monitor episode and relationship quality. However, we
know that the presented results are limited due to a single case approach, this study
is a good attempt to fill the gap of longitudinal studies when investigating the dy-
namics in customers satisfaetion.

References
Anderson, E.W.lFomell, C.lLehmann, D.R. (1994): Customer Satisfaction, Market

Share, and Profitability: Findings From Sweden, in: Journal of Marketing,
Val. 58, No. 3, pp. 53-66.
Boström, G.-O./Wilson, T. (1993): Changing Quality Perceptions Among Providers

And Customers - Impact of a Service Enhancing Product in Professional Serv-
ice Firms, in: EIASM Conference Proceedings (Part I): 3'd Workshop on Quality
Management in Services, Helsinki, 3'd-4th May 1993, pp. 53-72.
Danaher, P.J.lMattsson, J. (1994): Cumulative Encounter Satisfaction in the Hotel

Conference Process, in: International Journal of Service Industry Management,
Val. 5, No. 4, pp. 69-80.
Fisk, R.P. (1981): Toward A Consumption/Evaluation Process Model for Services,

in: Donnelly, J.H.lGeorge, WR. (eds.): Marketing of Services, AMA's Special
Conference on Services Marketing, Orlando, pp. 191-195.
FarnelI, C. (1992): ANational Customer Satisfaction Barometer: The Swedish Expe-

rience, in: Journal of Marketing, Val. 56, No. 1, pp. 6-21.
Grönroos, C. (1993): Toward a third phase in service quality research: Challenges

and future directions, in: Swarlz, TA.lBowen, D.E.lBrown, S.W (eds.): Ad-
vances in Services Marketing and Management, Vol. 2, London, pp. 49-64.
Haller, S. (1995): Beurteilung von Dienstleistungsqualität, Wiesbaden.
Liljander, V.lStrandvik, T. (1994): The Nature of Relationship Quality, in: EIASM

Conference Proceedings (Part 11): 4'h Workshop on Quality Management in
Services, Marne La Vallee (Paris), 12-13'h May 1994.
Meyer, A.lDomach, F. (1995): Nationale Barometer zur Messung von Qualität und
Kundenzufriedenheit bei Dienstleistungen, in: Bruhn, M.lStauss, B. (eds.):
Dienstleistungsqualität, 2. Auflage, Wiesbaden, pp. 429-453.
Meyer, A./Mattmüller, R. (1987): Qualität von Dienstleistungen - Entwurf eines pra-

xisorientierten Qualitätsmodells, in: Marketing ZFP, Vol. 9, No. 3, pp. 187-195
Parasuraman, A.lZeithaml, V.A./Berry L.L. (1988): SERVQUAL: A Multiple-Item

Scale for Measuring Consumer Perceptions of Service Quality, in: Journal of
Retailing, Val. 64, No. 1, pp. 12-40.
Scharitzer, D. (1994a): Dienstleistungsqualität - Kundenzufriedenheit, in: Topritz-

hafer, E. (ed.): Schriftenreihe "Forschungsergebnisse der Wirtschaftsuniversi-
tät Wien", Wien.
24 Claudia K/ausegger / Dieter Scharitzer
Scharitzer, D. (1994b): Measuring the Quality of Educational Services - A Process-

Oriented Approach Comparing Two Measurement Scales, in: EIASM Confer-
ence Proceedings (Part 11): 4th Workshop on Quality Management in Services,
Marne La Vallee (Paris), 12_13th May 1994.
Stauss, B. (1991): Dienstleister und die vierte Dimension, in: Harvard Manager
Vol. 13, No. 2, pp. 81-89.
Stauss, B. /Seidel, W. (1995): Prozessuale Zufriedenheitsermittlung und Zufrieden-

heitsdynamik bei Dienstleistungen, in: Simon, H.lHomburg, C. (eds.): Kunden-
zufriedenheit: Konzepte - Methoden - Erfahrungen, Wiesbaden, pp. 179-203.
Stauss, B./Weinlich, B. (1995): Process-oriented Measurement of Service Quality

by Applying the Sequential Incident Method, in: EIASM Conference Proceed-
ings (Part 11): 5th Workshop on Quality Management in Services, Tilburg,
11 th-12'h May 1995.
Zeithaml, V. A. (1981): How consumer evaluation processes differ between goods

and services, in: Donelly H.J.lGeorge, w.R. (eds.): Marketing of Services. AMA
Proceeding Series, Chicago, pp. 186-190.
National Customer Satisfaction Indices:
A Critical Investigation from an Application Perspective
Ursula Hansen and Thorsten Hennig-Thurau
Contents
Abstract 27
Introduction 27
2 Overview: On the Functionality of NCSls 28

2.1 Identifying NCSI's Addressees 28
2.2 A Systematization of Different Comparisons Implied by NCSls 30
3 Formulating Criteria for the Evaluation of NCSls 31
4 A Criteria-Based Investigation into the Construction and

Application of Selected NCSls 31
4.1 The American Customer Satisfaction Index (ACSI) and the
German Customer Satisfaction Barometer (DKB) as Reference
Objects 31
4.2 A Criteria-Based Comparison of ACSI and DKB 31
4.2.1 Theoretical Foundation 31
4.2.2 Aggregation of Data 31
4.2.3 Validity of Research Design 31
4.2.4 Inter-temporal Comparability of Results 31
4.2.5 Transparency of Methods and Results 31
4.2.6 Compatibility with Economic Statistics 31
5 Summary and Implications for Further Research 31
References
National Customer Satisfaction Indices 27
Abstract
The development of National Customer Satisfaction Indices (NCSls) has recently
become a central area of research in service management and marketing. Despite
an considerable amount on scientific and societal discussion, a massive lack of
meta-theoretical research in the area of NCSI development and usage exists. Based
on the identification of several levels of NCSI based satisfaction comparisons the
authors formulate a catalogue of general criteria which function as a frame of refer-
ence for the critical analysis and evaluation of NCSls. With the American Customer
Satisfaction Index and the German Customer Satisfaction Barometer, two well-
known indices are than evaluated with regard to the criteria previously worked out.
Several proposals are generated corresponding to the results of the analysis, which
can be used as a guideline for future improvements of NCSI methodology and us-
age.
1 Introduction
become a central area of research in service management and marketing. The term
NCSls is applied for satisfaction measurement and monitoring procedures that differ
from company-related satisfaction research insofar as they incorporate customers'
satisfaction judgements referring not only to one firm, but to a large number of com-
panies in different sectors of business. Building on this, NCSls allow insights into the
state of a nation's customers degree of satisfaction with service offerings. Up to now,
NCSls are established in several countries including the United States ("American
Customer Satisfaction Index"/ACSI, see Fornel! et al. 1996) and Germany
("Deutsches Kundenbarometer"/DKB, see Meyer/Dornach 1996). Furthermore, the
implementation of NCSls is in a planning state in numerous other countries, e.g. in
Austria, Switzerland and Israel.
Unfortunately, the present academic discussion of these nationwide tools for satis-
faction measurement is limited. Until now, contributions are only made by institution-
ally involved researchers (e.g. the developers of NCSls), but are missing from other
28 Ursula Hansen / Thorsten Hennig-Thurau
(formally non-involved) academics. Corresponding to this, the scientific discussion

on NCSls is predominantly of a more or less descriptive kind, while a critical and
meta-theoretical perspective is merely taken. This absence of constructive critics
may be explained by the early phase of the lifecycle of NSCI research, but, however,
is contradictory to the outstanding theoretical and particularly societal status which is
attached to NCSI results. As a consequence, the need for supplementing the exist-
ing "insight" contributions by critical and constructive "outside" analysis, which can
lead to further refinements of the existing methodology of NCSls must be postulated.
Thus, in our paper we address the following issues of NCSI research by taking such
an "outside" perspective. First, we identify the different addressees of NCSI and
specify their interest in NCSI methodology and/or results. Second, several levels of
NCSI based satisfaction comparisons applied in praxi by the users of NCSls are dif-
ferentiated (cross-sectional vs. longitudinal, intra-branch vs. inter-branch etc.). The
results of the analysis are transformed into a catalogue of general criteria which we
use in the following as a frame of reference for the critical analysis and evaluation of
NCSls. As the American ACSI and the German DKB can be seen as highly influen-
ti al for the development of further NCSls, both are chosen as objects of our detailed
investigation into the construction and application of nationwide satisfaction ba-
rometers. The !wo indices (ACSI and DKB) are than evaluated with regard to the
criteria previously worked out. Finally, the deficiencies of both barometers identified
here are used to develop proposals for the editors of NCSls for future improvements
of their methods.
2 Overview: On the Functionality of NCSls

2.1 Identifying NCSI's Addressees
Information genera ted by NCSls is used by a multitude of groups wh ich maintain
informational relations among each other. Figure 1 provides an overview of central
addressees (and stakeholders, respectively) of NCSls and systematizes their inter-
relationships.
Increasing KI1ow1edge
transparency .baut tho
ofmarlc.a1s welfare state
Fig. 1: A systematization 01 central stakeholders 01 NSCls
Consumers and the government as decision makers in malters of economic policy

are "end users" of NCSI data. For the government, the information on customer sat-
isfaction complements existing knowledge on the welfare state of the nation. In this
way, it tri es to get a genuine and radically subjective impression of the extent of
need fulfillment in the consumption society. For individual consumers, NCSls offer a
chance to compare their own company- or branch-related satisfaction judgements
with soeietal assessments. Beyond this, NCSI results mayaiso be considered as
additional arguments in the customer's buying process. NCSI information is gathered
by independent organizations, so that it may be used by customers as a guideline
when deciding between alternative products and brands. In this case, the existence
of NCSI data can contribute to a higher degree of market transparency.
Furthermore, the results of NCSls are used by several companies and the media to
increase their extent of goal fulfillment. In the case of companies, this means that the
marketing department can get valuable information from NCSls about the degree of
achieved closeness to customers, which must be seen as crucial in many of today's
market environments. Additionally, satisfaction data from NCSls can playa key role
in marketing controlling. Finally, media enterprises try to increase their circulation by
reporting on the nation's degree of satisfaction, which corresponds to the informa-
tional needs of their customers.
2.2 A Systematization of Different Comparisons

Implied by NCSls
The calculation of satisfaction scores by NCSls enables the researcher to analyze
Ihe data in a number of ways. Comparisons between different objecls of Ihe con-
sumers' satisfaction level musl be seen as the heart of the analysis. Fundamentally,
these comparisons can be divided into intra-branch and inter-branch comparisons.
On Ihe level of intra-branch comparisons, the following ways of analysis are possi-
ble: (a) The satisfaction data of company x generated by NCSI can be compared
with its competitor within the same branch, that is company y. (b) The satisfaction
data of company x can be related to the average satisfaction score of its branch.
80th ways of analysis offer company x the possibility to define its position in the
marke!. A common way of interpretation views comparatively high satisfaction scores
as an expression of a high degree of customer andlor market orientation what is of-
len taken as a synonym for competitiveness.
Inter-branch comparisons of satisfaction scores generated by NCSls open up further

possibilities for data analysis. Especially three basic ways of analysis have to be
mentioned here. First, the average satisfaction scores of different branches can be
compared with each other, which may lead to the development of an inter-branch
satisfaction ranking. Second, the satisfaction score for company x of branch X can
be compared with the average score of branch Y. Finally, company x may be inter-
ested in comparing its level of customer satisfaction with the satisfaction score of a
specific company classified in a different branch.
To make it even more complex, each of the before mentioned comparisons based on
NCSI data can principally be done as cross-sectional analysis or, if data collection is
carried out at different points in time, as longitudinal analysis. Longitudinal data from
NCSls can be helpful with regard to both controlling companies' and other institu-
tions' activities and the proactive identification of changes in consumer behavior.
Figure 2 illustrates the variely of optional comparisons on the basis of NCSI dala.
Branch X Branc;h Y
Braneh Compllny Branch company
lovel (Bl) Iovel (CLi levil level
, I
,
I
,,
I~L
(8)-
..0(7) ".(9),-
(3)".(3),
, ,
J i
B""';'h I 'tI ,
,'x ,
J (7)0 (8)0 (9)0
CL (10)0
J (2~,. (2)" (1~,·(1), (4)".(4), (i'J..'(5),
!
I i
! I
I
, ,
J
Branch X Branch Y
,i
.' Branch
,'Ievel (Bl)
CaRlpany
l"':el(Cl)
Branch
level
campany
level ,,
, (3) =
mean(x)
Bl - I
,
I see (2)
<-> see (4) t
mean (y)
Branch
I
I
11 11
JI
"
X I
C?I= (1) = (4)= (5)=
ca19pany, (x) campany, (x) campany, (x) campany, (x)
Cl <_>
I <-> <-> <.>
" mean (x) campany, (x) mean (y) campany, (y) 1.,'
I
Fig. 2: Optional compalisons based on NCSI data
Difficult issues are in particular (however not exclusively) associated with inter-
branch comparisons. Which conditions do allow comparisons of satisfaction scores
(on the company level or on the aggregated level of means) arnong different
branches? As Anderson (1993, p. 101) stresses, these comparisons "require not
only measurement, but also an understanding of how industry differences might af-
fect satisfaction". The complexity shown above is of great importance especially,
since it severely contrasts with the common impression of NCSI data as being easy-
to-understand and self-explanatory. Furthermore, the neutrality of NCSI organizers
and the incorporation of customer judgements seem to guarantee a high degree of
objectivity. The contrast of putatively easy understanding on the one hand and com-
plexity on the other hand bears the danger of misinterpretation for several
stakeholders. For example, companies may continue to offer their products instead
of modifying thern, and consumers may be misled by NCSI information about a sin-
gle company or a whole branch. As a consequence, specific requirements have to
be formulated with regard to the provision and use of information generated by

NCSls. The following section of the paper represents an attempt in this direction.
3 Formulating Criteria for the Evaluation of

NCSls
Theoretical foundation. The validity of results from empirical research is closely
related to the theoretical basis of the included constructs. Building on this, three sub-
criteria for the construction of NCSls can be derived. (a) The measurement of the
construct customer satisfaction is to be based on a consistent theoretical framework
that must be in accordance with the current state of knowledge in marketing science.
(b) Closely related is the demand for an integration of meaningful parts of the theo-
retical framework when constructing the NCSI. This criterion is of particular impor-
tance when the NCSI is used for the development of expertise on a certain com-
pany's success and competitiveness respectively. (c) Finally, the quality of NCSI
data is combined inseparably with a valid operationalization of the incorporated con-
structs and especially the construct of customer satisfaction.
Aggregation of data. In the case of NCSls, the existence of multiple stakeholders

implies several steps of data aggregation. Thereby, data aggregation on the one
hand has to be concrete enough for analysis on the single company level and on the
other hand must enable the analysis on the branch or even national level. As a con-
sequence, the complex aggregational procedure has to be carried out according to
the following aspects. (a) The method of data aggregation must be in accordance
with the requirements of the different stakeholders which also implies validity on the
highest (national) level of aggregation. (b) To ensure the expressiveness of any
comparison on the intra-branch level, the companies need to be classified in the
"correct" branch. (c) If comparisons on the inter-branch level are intended, the or-
ganizers of NCSls must also prove the adequacy of their procedure in this area of
research.
Validity of research design. As in other areas of social research, the expressive-

ness of NCSI data is closely related to the validity of the research design used for
empirical data collection. At a more concrete level, three faeets have to be men-
tioned here that determine the validity of the research method. (a) For an adequate
interpretation of satisfaction scores of different companies or branches, information
about the significance level is needed. We know that the significance criterion itself
has its weaknesses for which it is criticized in parts of the social science literature,
but in any case it provides important information ("better-than-nothing"). (b) Another
aspect that comes along with the significance criterion is the sampie size. An in-
crease of the sampie size represents an approximation of the population and simul-
taneouslya decrease of sampling error (see Churchill 1979, p. 65, for the relation-
ship between sampling error validity in empirical research). With regard to the differ-
ent levels of analysis in NCSls, the sampie must be large enough to allow inferences
to the consumers' satisfaction with the nation's supply with products, with single
branches, and even with specific companies. (c) Eventually, an additional indicator
for the quality of the data collection process is the length of the interviewing proce-
dure. As numerous studies indicate, there exists a nonlinear relationship between
the length of the interview and the validity of data (HammannlErichson 1994, p. 95).
After exceeding a threshold value at 30 to 45 minutes length, the validity of the data
drops (Böhier 1992, p. 91; Berekoven/EckertlEllenrieder 1991, p. 100). Conse-
quently, we demand that the design of the interviewing procedure for NCSI has to
take notiee of this relationship.
Inter-temporal comparability of results. In the case of longitudinal research, the

usage of NCSI is committed to an inter-temporal comparability of results. Ta ensure
the existenee of this inter-temporal comparability, two premises have to be fulfilled.
(a) The methods used for data collection, data aggregation, and the calculation of
satisfaction scores must be left unchanged or must at least be comparable over time.
(b) In addition to the stability of the research design, the comparability of data gath-
ered at different points in time also depends on the development of several exoge-
nous variables. These variables are much more difficult (or even impossible) to con-
trol for the organizers of NCSls, but may strongly influence the development of sat-
isfaction scores over time. Thus, it must be seen as a necessary requirement to con-
sider these potential influential factors which would lead to a more experimental re-
search design.
Transparency of methods and results. As iIIustrated above, NCSls are complex

measurement instruments with a number of snares. Thus, in addition to the previous
aspects related to the generation of data, another criterion concerns the way of the
editors' communication on NCSls. An appropriate extent of transparency is needed
here because of the danger of faulty and misleading interpretations. On a more con-
crete level, a key aspect of transparency refers to the necessity of combining the
communication on NCSI data with information on the underlying limitations of meth-
ods used to generate these data.
Compatibility with economic statistics. Finally, from the existence of the govern-
ment as a stakeholder of NCSI information, another requirement is derived. To en-
sure the usability of NCSI data on the macro economic (or governmental) level, the
conceptualization of satisfaction scores has to be compatible with economic statis-
tics. For this, (a) the aggregational procedure and (b) the structuring of branches are
of special relevance.
4 A Criteria-Based Investigation into the Con-

struction and Application of Selected NCSls
4.1 The American Customer Satisfaction Index
(ACSI) and the German Customer Satisfaction
Barometer (DKB) as Reference Objects
The American ACSI and the German DKB are special types of NCSls. Both are cho-
sen here particularly for two reasons. First, the ACSI and the DKB are at the center
of scholarly discussion in the area of NCSI. Their methodology is taken as basis for
the development of other NCSls in several countries, such as Austria, Switzerland,
and Taiwan. Second, apart from their common objective of nationwide satisfaction
measurement, a number of differences between both instruments exists. These dif-
ferences illustrate the multitude of degrees of freedom in the design of NCSls.
The development of the ACSI can be traced back to the conception of the Swedish
Barometer SCSB, which was implemented in 1989, and is inseparably connected
with the names of Claes FornelJ and Eugene Anderson (NERA 1991). The ACSI was
implemented in the United States in 1994 as result frem a co-operation preject of the
American Soclety of Quality and the National Quality Research Center of the Uni-
versity of Michigan Business School. Today, the findings for selected branches are
published quarterly. The ACSI measures the customers satisfaction with private and
public service offerings, but also includes all kinds of durable and non-durable goods
(see figure 3).
ACSI: National Economy-Sectors

and Industries
Manufacturing
Nondurables
8.3°",
(SIC2)
Manufactunng
Durables
9.6 %
(SIC3)
Transportation,
Commmications,
Utilities
8.8%
(SIC4)
'T'Retail
9.3 %
(SIC5)
Finance,
Insurance
6.6 %
(SIC 6)
Services
9.0%
(SIC 7,6)
Public
Administration.
Government
12.6 0",
(SIC9)
I
Apparel
Athletic shoes
I
Automobiles
Consumer
I
Airlines,
BroadcastinglTV
Department
stores
BanksJ
commercial.
I
Hospitals
Hotels
I
GarbageJtrash
service (central
Be., electronics Electric service Discount Insurancellife Motion pictures city, suburb)
Cigarettes Household Express delivery stores Insurancel $468 IRS
Gasoline appliances Phonel Restaurants! property $ 52 8
Nempaper pe/printers long-distance fast food $ 659 8
Personal car. S 499 B U.S. Postal Service Superma rkets
Soft drinks $4508 $ 375 B
S 654 8
Size of each sectors shown as % of GOP. Sales ftgures reftect total sales for the selected U.S.
nrms. Total sales of selected flrms are 58% of GOP for these 7 sectors and 43% of national GOP.
Fig. 3: Branches included in the American Customer Satisfaction Index

Source: ASO (1998)
The German NCSI, the DKB, was founded in 1992 on the initiative of the German
Marketing Association and was sponsored by the German Post AG. Since then, it
has been carried out annually, The methodology of the instrument was developed
and revised in the meantime by the German marketing academics Anton Meyer and
Frank Dornach. The DKB concentrates on private and public services from 36
branches of industry, Nonetheless, some manifest products (e,g., computer hard-
ware) are also part of the sampie.
36 Ursu/a Hansen / Thorsten Hennig-Thurau
4.2 A Criteria-Based Comparison of ACSI and DKB

4.2.1 Theoretical Foundation
Consistent theoretical framework
With regard to the existence and design of a theoretical framework, the ACSI and
DKB differ fundamentally. The ASel is based on a complex structural model which is
shown in figure 4. The model contains five constructs in addition to customer satis-
faction of which three (perceived quality, customer expectations, perceived value)
are interpreted as antecedents of satisfaction (see for perceived value Holbrook
1994). These three variables are also treated as interrelated, with the perceived
value construct being influenced by the customer's quality perception and his/her
expectations. The two other variables (customer complaints and customer loyalty)
are conceptualized as consequences of satisfaction.
Fig. 4: Structural model of the American Customer Satisfaction Index

Source: Fomel! et al. (1996, p. 8)
However, the validity of the postulated framework must be questioned in some

points. With reference to recent studies, especially the interpretation of the cus-
tomer's quality perceptions as an antecedent of customer satisfaction must be
viewed as critical. As Hennig-Thurau/Klee (1997, p. 744) show on the basis of a
comprehensive literature review, satisfaction is appropriately regarded as "a short-

term emotional state that results from an intra-personal comparison of the customers'
expectations with the evaluation of a single [... ] service encounter", while the cus-
tomers quality perception is seen as an higher-order and more stable variable, which
is built mainly on previous experiences of (dis )satisfaction related to discrete trans-
actional episodes. Consequently, the causal relationship between satisfaction and
perceived quality is opposite to what is postulated in the ACSI model: The emotional
state of satisfaction leads to an overall, global, and cognitively dominated attitude
about service quality (see also Dabholkar 1993; Bitner/Hubbert 1994; Bolton/Drew
1991, 1994). The most problematic aspect of the Fornell et al. (1996) model is the
implicit understanding of quality perception as an quasi-objective construct. That is
clearly illustrated when StrandviklLiljander (1995, p. 119) explain the state of being
dissatisfied despite high quality by arguing that the product or service "does not fit
the customer's preferences".
In opposition to the ACSI, the DKB is not based on a theoretical framework. In addi-
tion, the organizers of the DKB do not differentiate systematically between customer
satisfaction and other related constructs. In the case of satisfaction and quality per-
ception, they justified their position by quoting an empirical investigation carried out
on their own, which shows no significant differences between both constructs
(Meyer/ErtI1996, p. 214). However, this finding must be questioned for it ignores the
close relation between the empirical operationalization of constructs and their theo-
retical conceptualization (e.g. Homburg/Giering 1996).
Integration of Meaningful Parts of the Framework
In addition to customer satisfaction, both indices analyzed here collect information

on the degree of customer loyalty and the consumers' complaining behavior. Both
indices assume a direct and linear relationship between satisfaction and loyalty. This
assumption, however, is inconsistent with recent theoretical and empirical research
which shows that the influence of satisfaction on customer loyalty is rather restricted
(Stauss/Neuhaus 1996; Hennig-Thurau/Klee 1997, p. 738 f.). Thus, researchers
have argued for the integration of additional variables. In the context of relationship
marketing literature, particularly the importance of quality perception, trust, and
commitment are seen as crucial for customer loyalty or retention. A model which in-
tegrates all of these constructs was presented by Hennig-ThurauiKlee (1997, see
also Hennig-Thurau 1998). In their model, these variables are taken as components
of relationship quality, which itself is determined by the customer's level of satisfac-
tion (see figure 5). The consideration of such a "meta-construct" relationship quality
within the construction of NCSls could contribute to increasing the level of explained
variance.
Fig. 5: Relationship quality as moderator between customer salisfac\ion and customer loyalty
Source: Hennig-Thurau/Klee (1997, p. 742)
With regard to the influence of satisfaction and loyalty on profitability and turnover
as indicators of economic success, both indices assume a significant and positive
relation between these variables. While in the case of the ACSI this ass um pt ion is
subject of empirical investigation (Anderson et al. 1997), Meyer/Dornach 1995b,
p. 40 ff.) as organizers of the DKB do not undertake detailed investigations in this
area and refer to secondary statistical data instead. To get a more differentiated un-
derstanding of the relationship between psychographie and economic target vari-
ables, an integration of the lalter into NCSls' variable system can be seen as desir-
able.
Valid Operationalization of the Incorporated Constructs
Another difference between both indices lies in the operationalization and measure-
ment of the satisfaction construct. By treating the customers' expectations as a se pa-
rate variable, the ACSI takes into account the heterogeneity of expectations con-
cerning different kinds of services and goods. This procedure enables the organizers
of the ACSI to find out about the level of performance of the branches incorporated
in the investigation. For example, Fornell et al. {1996, S. 7)came to the conclusion
that regarding the economy of the United States "customer expectations play a
greater role in sectors in which variance in production and consumption is relatively
low". Satisfaction is measured here on a multi-item scale with three items (overall
satisfaction, expectancy disconfirmation, performance vs. ideal product) because the
construct is seen (in accordance with the relevant literature) as "not directly observ-
able" (Fornell 1992, p. 12). This can be regarded principally as an adequate meas-
urement procedure. However, the authors of the ACSI do not publish information on
the reliability of the scale (e.g., coefficient alpha) which prevents us from a more
substantial evaluation. With the exception of complaining behavior, the other vari-
ables in the conceptual ACSI model shown in figure 4 are also measured on multi-
item scales (Fornell et al. 1996, p. 10).
In the case of the DKB, expectations and performance assessments are separated
only up to a certain degree. Since 1995, the DKB includes autonomous items for the
measurement of customer expectations, however, this information is not integrated
into the calculation procedure but has a complementary status only. The satisfaction
construct itself is measured on a single-item scale ("How satisfied are you with the
services or goods of XY at all?") which maybe is the most serious deficiency of the
DKB methodology. The problems associated with such a procedure have been in-
tensively discussed in literature and should not be recapitulated here. A trenchant
quote comes from Jacoby (1978, p. 66) when he states: "How comfortable would we
feel having our intelligence assessed on the basis of our response to a single ques-
tion?" (see also Churchill1979, p. 66).
Recent findings of customer satisfaction research, especially on the predictive di-

mension of the satisfaction construct, are not considered neither by the ACSI nor the
DKB. Especially the following two aspects can be expected to be fruitfuI for the fur-
ther development of satisfaction measurement by NCSI (for an overview, see Hen-
nig-Thurau/Klee 1997). First, some authors have mentioned that the influence of
customer satisfaction (and the customer's quality perception respectively) on cus-
tomer loyalty varies with the aspiration level of the customer (Stauss/Neuhaus 1996).
As a consequence, several different states of satisfaction (e.g., "progressive satis-
faction", "resigned satisfaction") can be distinguished. Each of them has a specific
relation with customer loyalty. Second, another approach tries to differentiate satis-
faction scores depending on the level of customer involvement (Bloemer/Poiesz
1989). This approach leads to the fundamental distinction between manifest satis-
faction (corresponding to high involvement settings) and latent satisfaction (corre-
sponding to low involvement settings). While in the case of manifest satisfaction a
stable and significant influence of customer satisfaction on repurchasing behavior
exists, such a relation does not exist when the customer has low or even no in-
volvement in the product s/he has to assess (latent satisfaction). As with the different
states of satisfaction, a differentiation of NCSI data with regard to the involvement
level may rise the validity of measurement enormously.
4.2.2 Aggregation of Data

Method of Aggregation
Also the method of data aggregation differs significantly. The development of the
ACSI was accompanied by an intensive scientific discussion on the permissibility of
data aggregation in the case of individual satisfaction scores (Fornell et al. 1992;
Anderson 1992; NERA 1991; Johnson/Forneli 1991; Fornell 1990). The search for
an adequate way of calculating satisfaction scores for branches and the nation as a
whole was the central issue of this discussion (Anderson 1992, p. 93 ff.) and led 10
an integration of the business administration and the economic perspective. The de-
veloped procedure was validated on empirical data of the Swedish barometer from
1989 to 1991. Furthermore, the method was modified to reduce distortions caused
by the over- or under-representation of certain branches (Fornell et al. 1992).
On the company level, the index calculation is based on the multi-item measurement
of satisfaction described above. In the next step, the company related satisfaction
scores are aggregated to a satisfaction score for the respective branch. Here, ag-
gregation implies the multiplication of each company related score with a specific
weighting factor. This weighting factor represents the percentile contribution of a
single company to the total turnover of the branch. The weighting procedure of the
ACSI is criticized by Meyer/Ertl (1996, p. 224), who argue that "the satisfaction
judgement of a Mercedes driver is of higher importance for the branch satisfaction
score as the one of a person who drives a VW Golf'. However, such a procedure is
advantageous with regard to the criterion of compatibility with economic statistics
(see below). Furthermore, branches are summed up to more abstract sectors of
industry (e.g., services, manufacturing durables) using the same weighting proce-
dure aga in. This time, the weighting factors are derived from the macro economic
system of national accounts.
In contrast to this complex and historically grown method, the data aggregation pro-
cedure of the DKB is rather straightforward. Although its organizers have taken no-
tice of the Swedish experiences (Meyer/Ertl 1996), they do without any weighting
procedure. Instead, unweighted means are aggregated on both the company level
and the branch level. To improve the quality of data interpretation, in addition to the
overall satisfaction score the customers' satisfaction with a number of attributes are
measured. These attributes differ from branch to branch (e.g., employees' friendli-
ness, quickness of task fulfillment, technical skills, see Meyer/Dornach 1995b,
p. 89), but their identification is not explained by the organizers of the DKB and no
evidence is given for their relevance for the customer's satisfaction judgement. The
customers' satisfaction on the attribute level is not incorporated into the aggregation
procedure and, therefore, has to be interpreted in isolation.
Classification of Companies and Intra-branch Comparison
The assignment of companies to certain branches forms the basis for intra-branch
comparisons of customer satisfaction which can be seen as a central component of
NCSI analysis. By the definition of branches adecision is made on the kind of com-
panies which are interpreted as competitors within the same market. So, the selec-
ti on of comparison standards for a single enterprise is strongly influenced by this
decision (see for the general relevance of relative satisfaction judgements
DicklBasu 1994).
42 Ursu/a Hansen / Thorsten Hennig- Thurau
The definition of branches and the assignment of companies to a specific branch is

oriented in the case of the ACSI to the structure of the system of national accounts
(ACSI). In the case of the DKB, no single formal criterion for structuring is given by
the organizers but the result is close to popular systematization. Especially the ag-
gregation of branches to the higher level construct sector implied by the ACSI meth-
odology raises problems. For example, a "services" sector and a "finance and insur-
ance" sector are distinguished, though both sectors are closely related with regard to
several aspects, and therefore a distinction must be seen as highly arbitrary. Fur-
thermore, the systematization of branches of both indices is partially superficial and
vague. For instance, in the case of the ACSI hausehold batteries and HiFi systems
are both assigned to a common branch named "consumer electronics", and the DKB
asks the consumers to evaluate a heterogeneaus category (and branch, respec-
tively) titled "vacational destinations".
The problems arising from these kinds of systematization are illustrated by a deci-
sion of the regional court Hamburg. According to this decision, the Techniker Kran-
kenkasse is not allowed to communicate that they are the "top" of the branch of
health insurance companies (LG Hamburg, Al. 315053/97). The court argued that
according to its view, the market offers of private health insurance companies on the
one hand and public health insurance companies on the other hand are not compa-
rable with each other.
Methodology of Inter-branch Comparisons
The problem of heterogeneity described above within the context of intra-branch

comparisons is intensified on the inter-branch level. At the DKB, the aggregated
satisfaction scores for the different branches are merged into a total rating of
branches (see figure 6, which shows an excerpt from the 1997 rating of branches).
Apart from the general dubiousness of such a ranking based on the inseparability of
the evaluation of companies and the public image of the companies offerings (e.g.,
see the case of social organizations), additional problems coming along with such a
way of analysis can be exemplified by means of the comparison between life insur-
ance companies and companies offering computer hardware. For example, while life
insurances can be characterized through a large amount of credence qualities, com-
puter hardware predominantly exists of search and experience qualities, which leads
to different internal judgement processes. Furthermore, the camparisan between
monopolistic branches (e.g., the German Mail) and branches with high competition
(e.g., parcel services) must be seen as inappropriate because of the different levels
of customers expectations (see also below).
vacational destinations 1,99

pharmacies 2,21
travel companies 2,23
broadcast stations 2,25
credit card companies 2,25
motoring organizations 2,27
car insurance companies 2,28
airtines (private customers) 2,30
garages 2,31
cellular phone companies (hardware) 2,36
mail order firms 2,37
electrical appliances companies (customer service) 2,38
health care insurance comapnies 2,39
closing traders 2,40
banking companies 2,41
social and ecological organizations 2,41
supermarketslfood shops 2,41
drug stores 2,42
computer companies (hardware) 2,48
computer companies (operating systems) 2,49
building societies 2,53
life insurance companies 2,53
energy suppliers 2,54
do-it-yourself markets 2,58
Fig. 6: DKB 1997 rating of branches Relationship quality as moderator between customer
satisfaction and customer loyalty
Source: Excerpt; DKB (1998a)
The organizers of the ACSI do not calculate a ranking of the branches' satisfaction
scores. Instead, the pros and cons of comparisons between different branches are
intensively discussed and the specific features of certain branches are worked out
(Fornell et al. 1996; Anderson 1993, S. 101). However, they keep themselves the
option for inter-branch comparisons, when they say: "We can determine how weil a
particular firm is doing relative to the best firms in its own industry, the best firms in
other industries in that sector, or the best in nation as a whole. Industries and sec-
tors can be compared with one another in a similar fashion" (Fornell et al. 1996,
p. 12). Referring to this statement, the existence of something like "the best company
in nation as a whole" shall be questioned here with vigor once again.
Anderson (1993, p. 93 f.) indicates in this context that even the intensity of the rela-
tionship between customer satisfaction and loyalty may differ from branch to branch,
which also effects the relative importance of satisfaction scores. 1 This founding is
also supported by an empirical study of the authors, whose results are described in
a prior version of this paper (Hansen/Korpiun/Hennig-Thurau 1998).
4.2.3 Validity of Research Design

Significance of Data
For the addressees of NCSI data, the validity of satisfaction scores and company
rankings based on these scores is of central relevance. As a general guideline,
rankings or comparisons of companies are permissible only if they are significant
from a statistical point of view. If the differences between the companies considered
in the comparison are not significant, one has to assume that the measured result is
an artifact and does not exist in reality. So, a follow-up study could lead to different
or even reverse results. In the health insurance companies example cited above, the
difference between the Techniker Krankenkasse as "best company" and other health
insurance companies was rather small (0.02 points on a five-point scale in 1996), so
that the competent regional court consequently has questioned the validity of any
ranking based on these data.
Both indices inspected here publish no information on the significance of satisfaction

comparisons neither on the branch nor on the company level what has to be criti-
cized. This omission prevents the user from an assessment of the validity of any
comparison or ranking as mentioned above. Closely related to this deficiency there
is another weak point of the DKB: On the single-item scale used, the difference be-
tween the best and the worst satisfaction score on the level of branches is approxi-
Interestingly, this argument can also be found at Meyer/Domach (1995a, p. 433).

mately only 0.2 points that implies the assumption that the used measurement scale
has no good selectivity due to skewness. In the case of the ACSI, these problems do
not occur to the same degree because it uses a different scale (ten-point instead of
five-point scale; see Forne1l1992, p. 13).
Sam pie Size
Both barometers, the ACSI and the DKB, are based on large sampies. In the case of
the ACSI, data from more than 45.000 respondents is used to calculate the scores
(Fornell et al. 1996, p. 13). For the DKB, a sam pie of approximately 37.000 consum-
ers is contacted (DKB 1998b). As a consequence, analysis of aggregated levels
(national level, sector level, branch level) will raise no validity problems. However,
on the level of single enterprises, the DKB calculates satisfaction scores based on
sampie sizes of 100 (and even smaller, see Meyer/Dornach 1995b, p. 136), what
limits the generalizability of the calculated results significantly. Here, an increase of
the number of respondents per company is strongly required for getting valid infor-
mation on this level of analysis. In the case of the ACSI, for the calculation of satis-
faction scores on the company level approximately 250 interviews are conducted per
firm (Fornell et al. 1996, p. 9). From a methodological point of view, this can be seen
as so me kind of lower limit for the generation of valid data.
Length of Interviewing Procedure
For both instruments, there is no information published about the average length of
the interviewing procedure. However, while in the case of the ACSI customers are
only asked to evaluate companies from one or two branches, DKB respondents have
to answer questions regarding their satisfactory state with up to nine different
branches (Meyer/Dornach 1995c, p. 171). Unlike the ACSI, wh ich includes only 25
standardized questions (17 structural, 8 demographie), the DKB covers a number of
additional questions (Meyer/Dornach 1995a, p. 442 f.) which extends the interview-
ing procedure. Furthermore, in both cases data is predominantly collected by tele-
phone interviews. This influences the relationship between interviewing length and
validity, as Churchill states: "The general rule of thumb is that long questionnaires
can be handled best by personal interview and least weil by telephone interview"
(Churchill 1991, p. 335; italics added). As a consequence, it must be supposed that

the length of the interviewing procedure of the DKB exceeds the critical threshold
thus having a negative impact on the validity of NCSI results.
4.2.4 Inter-temporal Comparability of Results

Continuity of Methods
Changes in the level of customer satisfaction over time may be used as an indicator
for increases or decreases of market performance. Therefore, they principally can be
interpreted as a key variable in a company's or the governmental controlling system.
The condition of method continuity is met by both the ACSI and the DKB. The meth-
odology of the ASCI has been slightly modified since its first use in 1994, but the
central aspects were left unchanged. In the case of the DKB, the methodological
foundation is also kept constant. Merely the number of considered firms and the
quantity of additional information have been extended continuously by the organiz-
ers of the instrument (Meyer/Dornach 1995b).
Constancy of Exogenous Variables
Even in the case of methodological continuity, a number of obstacles are still related
with longitudinal satisfaction research. Most of these obstacles have to do with the
dynamic environment, which influences the customer's state of (dis)satisfaction over
time. So, in re cent years many efforts have been made by companies (e.g., total
quality programs, relationship management tools) to reach a higher level of customer
orientation and to get "closer to the customer". All these efforts seem to have been
unsuccessful because till 1996 the degree of customer satisfaction has constantly
fallen nationwide: belween 1994 to 1996 the ASCI has lost 3,1% (NQRC 1997), and
the results of the DKB show a quite similar development. 2
While some researchers (e.g., Meyer/Dornach 1995b, p. 1 ff.) view this development
as the true expression of a societal trend and consequently postulate measures to
2 In 1997, for the first time since its renewal in 1992 the DKB has found an increase of custo-
mer satisfaction in some areas.
stop it, other scientists rather draw the attention to changes of exogenous variables
in general and to the consumers rising aspiration level in particular. To explain this
increase of the aspiration level, particularly the following arguments are stated fre-
quently:
• Today consumers find more and more pleasure in the output of NCSls. This
kind of popularization is mainly based on the coverage of NCSI results in
large parts of the media. As a consequence of the increased popularity of
NCSls, consumers are becoming more sensititive to customer satisfaction
(Köhler 1996, p. 583). Although this may be intended by the organizers of
NCSls, the implications of this rise of customers expectalions are integrated
in the NCSI methodology at most partially. In the case of ACSI, the modeling
of expectations as aseparate construct can at least be seen as starting
point.
• Changes of customer salisfaction over time can also occur when the prod-
ucts or services as reference objects are modified in a significant dimension
(e.g. as a result of technical progress). With the modification of the product
there usually goes along a justification of the interna I expectation standard
on the customers' side, which limits the validity of a comparison between
former and today's levels of satisfaction.
• Finally, customer expectations are implicitly tied to existing market structures
(e.g., monopolistic or competitive markets). If these market structures
change over time as it happens today in Germany with the telecommunica-
tions market, new competitors are appearing whose offerings also change
the customers' expectation levels. Consequently, a comparison of satisfac-
tion scores on markets with monopolistic conditions with scores on highly
competitive markets must be questioned in general for its validity.
4.2.5 Transparency of Methods and Results

As illustrated above, several stakeholders and especially the media are taking notice
of NCSI results. However, often only parts of the information necessary to develop a
holistic and therefore true understanding of the relationships investigated are noted
by the addressees. As a consequence, the actions taken by consumers or compa-

nies based on the received NCSI information may be frequently sub-optimal or, in
some ca ses, even of danger for the continuity of business.
Looking at ACSI and DKB, especially the latler has to be criticized with regard to this
criterion. Here, several general statements can be found both in the official report
and in scientific articles which do not cover the different restrictions of the methodol-
ogy discussed in parts in this paper. For example, the validity problems arising from
intra- and inter-branch comparisons and from small sam pie sizes are not even men-
tioned. Closely related, companies are explicitly encouraged by the organizers to
integrate the results of the DKB within their advertising strategy (see Meyer/Dornach
1995a, p. 450) but no additional information is given on how to handle the complex-
ity of the data. Furthermore, the importance of the satisfaction construct for customer
loyalty and economic success is not put into perspective by the authors of the DKB.
Instead, satisfaction is rather treated as some kind of "magie formula" for a com-
pany's success. In the case of the ACSI, an extensive methodological report is
available and the organizers have referred to at least some of the caveats related to
NCSI in a number of scholarly articles.
Additionally, for both barometers (DKB and ACSI) relevant information on the meth-
odology of the instrument has not been published yet (e.g., significance levels, reli-
ability coefficients). So, summing up the things mentioned above, the organizers and
communicators of both barometers' data are asked to commit themselves to a higher
degree of transparency and completeness with regard to the results and the meth-
odology of NCSls.
4.2.6 Compatibility with Economic Statistics

Because of the differing foci of both indices, it is not surprising that the ACSI has a
higher level of compatibility with economic statistics in general. With regard to the
aggregation of data, one can comment that as a consequence of the weighting pro-
cedure the satisfaction scores can easily be integrated into the national accounting
system. In the case of the DKB, no weighting is performed so that a combination of
results with economic statistics is hardly possible. With the structuring of branches it
is almost the same: while the ACSI method is directiy based on the multi-level stan-
dard industrial systematization (SIC), the less formally derived systematization of the
DKB restricts the possibilities of cooperative analysis between the micro (marketing)
and the macro (economics) level.
5 Summary and Implications for Further

Research
NSCls give information about the level of consumers' satisfaction and fulfill several
needs. At the same time, the complexity of their methodology raises a number of
questions regarding the validity of the generated results. Drawing upon the investi-
gation undertaken here, some proposals shall be formulated that may contribute to
the further development of the instruments. To go into details, the following aspects
seem worth mentioning:
Proposition 1: The measurement of customer satisfaction should be based on a

theoretical framework (DKB) that integrates recent knowledge from different areas,
but especially from relationship marketing research (ACSI). Central aspects should
be an extended conceptualization of the satisfaction construct, and the integration of
other determinants of customer loyalty, e.g. trust and commitment. In the case of the
DKB, in addition the operationalization of the satisfaction construct should be
changed into a multi-item approach replacing the single-item approach used up to
now.
Proposition 2: In the context of data aggregation, the existing systematizations of

branches have to be object of modification. The research on market structure analy-
sis (see Day/ShockerlSrivastava 1979; MyerslTauber 1977; Bauer 1989) can be ex-
pected to turn out as a fruitful source for this. As we have shown in our analysis, in-
ter-branch comparison and especially rankings of branches are accompanied by a
number of problems. Because of their questionable validity, their calculation and
publication should generally be reconsidered.
Proposition 3: The relevance of NCSls should be increased by communicating ad-

ditional information (e.g., significance level of satisfaction comparisons) with regard
to the underlying methodology. Furthermore, the generalizability of all generated

results should be tested with regard to the background of the sampie size prior to
publication.
Proposition 4: Inter-temporal comparisons of satisfaction scores should be com-

plemented by information about central exogenous variables. Here, particularly the
customers aspiration level is of great importance for the interpretation of longitudinal
data.
Proposition 5: The way of communication on NCSI results should consider the mul-
tifarious risks of misinterpretation. The organizers should prove responsibility for the
generated data by informing their addressees on the existing methodological prob-
lems and limitations of their instrument.
Proposition 6: As the actors of economic policy are NCSI stakeholders, it has to be

checked whether the compatibility of the indices' methodologies with economic sta-
tistics can be heightened. By the way, this must be taken into account when modify-
ing the systematization of branches as it was suggested in proposition 2.
References
Anderson, E.W. (1993): Firm, Industry and National Indices of Customer Satisfac-
tion. Implications for Services, in: Advances in Services Marketing and Man-
agement, Vol. 2, pp. 87-108.
Anderson, E.W.lFornell, C.lRust, RT. (1997): Customer Satisfaction, Productivity,

and Profitability: Differences Setween Goods and Services, in: Marketing Sci-
ence, Vol. 1, No. 2, pp. 129-147.
ASO (1998): American Society for Ouality, National Economy-Sectors and Indus-
tries, http://www.asq.org/graphics/fig2a.gif, 17.09.98.
Bauer, H.H. (1989): Marktabgrenzung, Serlin.
Berekoven, L./Eckert, W.lEllenrieder, P. (1991): Marktforschung: Methodische

Grundlagen und praktische Anwendung, 5. Aufl., Wiesbaden.
Bitner, M.J.lHubbert, A.R (1994): Encounter Satisfaction Versus Overall Satisfac-

tion Versus Quality: The Customer's Voice, in: Rust, RT.lOliver, RL. (eds.):
Service Quality: New Directions in Theory and Practice, Thousand Oaks,
pp. 72-94.
Böhler, H. (1992): Marktforschung, 2nd ed, Stuttgart u. a.
Bolton, RN.lDrew, J.H. (1991): A Multistage Model of Customer's Assessments of

Service Quality and Value, in: Journal of Consumer Research, Vol. 17, No. 3,
pp. 375-384.
Bolton, RN.lDrew, J.H. (1994): Linking Customer Satisfaction to Service Opera-

tions and Outcomes, in: Rust, R.T.lOliver, R.L. (eds.): Service Quality: New Di-
rections in Theory and Practice, Thousand Oaks, pp. 173-200.
Churchill, G.A. Jr. (1979): A Paradigm for Developing Setter Measures of Marketing
Constructs, in: Journal of Marketing Research, Vol. 16, No. 1, pp. 64-73.
Dabholkar, P.A. (1993): Customer Satisfaction and Service Quality: Two Constructs
or One?, in: Cravens, D.W.lDickson, P. (eds.): 1993 AMA Educators' Proceed-
ings, Chicago, pp. 10-18.
Day, G.S./Shocker, A.D./Srivastava, RK. (1979): Customer-Oriented Approaches

to Identifying Product-Markets, in: Journal of Marketing, Vol. 43, No. 3,
pp. 8-19.
DKB (1998a): Das Ranking der Globalzufriedenheit 1997, http://www.postag.de/

kundenbarometer/dkb_ranking97. html, 15.03.1998.
DKB (1998b): Was ist das Deutsche Kundenbarometer?, http://www.postag.de/

kundenbarometer/dkb_was97.html, 10.09.1998.
52 Ursu/a Hansen / Thorsten Hennig-Thurau
Fomel!, C. (1992): ANational Customer Satisfaction Barometer: The Swedish Expe-

rience, in: Journal of Marketing, Vol. 56, No. 1, pp. 6-21.
Fomel!, C./Johnson, M.D.lAnderson, E.W./Cha, J./Bryant, E. (1996): The Ameri-

can Customer Satisfaction Index: Nature, Purpose, and Findings, in: Journal of
Marketing, Vol. 60, No. 3, pp. 7-18.
Hammann, P.lErichson, 8. (1994): Marktforschung, 3rd ed., Stuttgart u.a.
Hansen, U./Korpiun, M./Hennig-Thurau, T. (1998): Nationale Kundenzufrieden-

heitsindizes als Informationsgrundlage des Dienstleistungs-Management. Eine
kritische Bestandsaufnahme, in: Bruhn, M.lMeffert, H. (eds.): Dienslleistungs-
Management, Wiesbaden, pp. 307-342.
Hennig-Thurau, T. (1998): Konsum-Kompetenz: Eine neue Zielgröße für das Man-

agement von Geschäftsbeziehungen, FrankfurtJM. u.a.
Hennig-Thurau, T./Klee, A. (1997): The Impact of Customer Satisfaction and Rela-

tionship Quality on Customer Retention: A Critical Reassessment and Model
Development, in: Psychology & Marketing, Val. 14, Special Issue an Relatian-
ship Marketing, December, Na. 8, pp. 737-764.
Holbrook, M.B. (1994): The Nature of Custamer Value, in: Rust, R.T.lOliver, R.L.
(eds.): Service Quality: New Directians in Theary and Practice, Landan,
pp. 21-71.
Homburg, C./Giering, A. (1996): Konzeptualisierung und Operationalisierung kom-

plexer Konstrukte, in: Marketing ZFP, 18. Jg., No. 1, pp. 5-25.
Jacoby, J. (1978): Consumer Research: AState of the Art Review, in: Journal of
Marketing, Val. 42, No. 2, pp. 87-96.
Koehler, R. (1996): Dienstleistungswüste Deutschland?, in: Die Betriebswirtschaft,

Vol. 56, No. 5, pp. 583-586.
Meyer, A./Domach, F. (1995a): Nationale Barometer zur Messung von Qualität und
Kundenzufriedenheit bei Dienstleistungen, in: Bruhn, M.lStauss, B. (eds.):
Dienslleistungsqualität: Methoden, Konzepte, Erfahrungen, Wiesbaden
pp. 429-453.
Meyer, A./Dornach, F. (1995b): Das Deutsche Kundenbarometer 1995 - Qualität

und Zufriedenheit, Deutsche Marketing Vereinigung e. V. und Deutsche Bun-
despost Postdienst, Düsseldorf u.a.
Meyer, A./Domach, F. (1995c): Das Deutsche Kundenbarameter - Qualität und Zu-

friedenheit, in: Simon, H.lHomburg, eh. (eds.): Kundenzufriedenheit: Konzepte,
Methoden, Erfahrungen, Wiesbaden, pp. 161-178.
Meyer, A.lDomach, F. (1996): Das Deutsche Kundenbarometer 1996 - Qualität und

Zufriedenheit, Deutsche Marketing Vereinigung e. V. und Deutsche Bun-
despost Postdienst, Düsseldorf u.a.
Meyer, A./Ertl, R. (1996): Nationale Barometer zur Messung von Kundenzufrieden-

heit. Ein Vergleich ausgewählter grundlegender Aspekte des "Deutschen Kun-
denbarometers - Qualität und Zufriedenheit" und des "American Customer
Satisfaction Index (ACSI)", in: Meyer, A. (ed.): Grundsatzfragen und Herausfor-
derungen des Dienstleistungsmarketing, Wiesbaden, pp. 201-231.
Myers, J.H./Tauber, E. (1977): Market Structure Analysis, Chicago.
NERA (National Economic Research Associates) (1991): Developing aNational

Quality Index: A Preliminary Study of Feasibility, New York.
Stauss, B./Neuhaus, P. (1996): The Dissatisfaction Potential of Satisfied Customers

- The Qualitative Satisfaction Model. Paper presented at the French-German
Workshop on Services Marketing Research, 28.2-1.3.1996 in Innsbruck, Aus-
tria.
Strandvik, T./Liljander, V. (1995): A Comparison of Episode Performance and Re-

lationship Performance for a Discrete Service, in: Kleinaltenkamp, M. (ed.):
Services Marketing: Concepts and Applications, Wiesbaden.
The Link between Customer Satisfaction and Dealer Satisfaction -
The Case of the German Car Industry
Alexander Decker
Contents
Abstract 57
Introduction and Goal of the Paper 57
2 Conceptual Background 59
2.1 Customer Satisfaction in the Car Industry 59
2.2 Dealer Satisfaction in the Car Industry 61
2.2.1 Literature Review 61
2.2.2 Definition and Conceptualization 63
3 The Relationship between Customer Satisfaction and Dealer
Satisfaction 66
3.1 Consideration of Preliminary Research Findings 66
3.2 Examination of the Cause-Effect-Relation 70
3.2.1 Dealer Satisfaction will Lead to Customer Satisfaction 70
3.2.2 Customer Satisfaction will Lead 10 Dealer Salisfaction 72
3.2.3 Dealer Salisfaction and Customer Salisfaclion are
Relaled Reciprocally 73
4 Summary: Implications for Future Research 75
References
Appendix
The Link between Customer Satisfaction and Dealer Satisfaction 57
Abstract
The satisfaction of car dealers with their suppliers has gained some remarkable at-
tention in the German car industry in recent years. A major reason for this phenome-
non may be seen in the apparent link between dealer satisfaction and customer sat-
Isfaction. Although this relation seems apparent, the positive link between dealer and
customer satisfaction is not proven. For example, there exists no empirical evidence
for the causaJity of the relation. Therefore, the goal of this paper is to gain some
theoretical insights into the relationship on the basis of preJiminary findings and addi-
tional industry related considerations.
The findings of the analysis point out, that on the basis of preliminary empirical find-
ings there is only a weak support for the assumption of a positive relationship. In ad-
dition, the direction of the connection is yet unclear. According to a more theoretical
viewpoint, it seems to be more likely that the two constructs are at least positively
related. An additional calculation will show, that none of the two constructs has a
higher impact on the other. Limitations of these findings are going to be shown as
weil as implications for future research in this area.
1 Introduction and goal of the paper

In recent years, the power balance between manufacturers and retailers in many
industries has been shifting towards the dealers. As a consequence, manufacturers
do not only worry about their customers' satisfaction, but also ca re about the satis-
faction of their dealers. The reason for this striving can be seen in the hope of
manufacturers, that satisfied dealers will be more likely to behave as intended by the
manufacturer more likely (RuekertiChurchili 1979, p. 226 with further references).
This is important to manufacturers, as with the lack of power, the ability to induce
behavior on the part of the dealer will be limited. In this context it is remarkable, that
the major part of surveys regarding dealer satisfaction in Germany are undertaken in
the car industry, although in this industrial sector the manufacturer still keeps the
power.
58 Alexander Decker
A reason for this phenomenon may be seen in the apparent link between dealer
satisfaction (satisfaction of a dealer with his relationship to the manufacturer) and
customer satisfaction. Satisfied dealers will probably have more satisfied customers.
This is of high interest for car manufacturers, as the dealership is becoming more
important for the success of the brand. In times of severe competition, rising cus-
tomer expectations and products, which become more and more similar in sense of
design and quality, dealers hold a major part of the potential to differentiate the
brand from competitors (SeweIl1992, p. 124; Thurow 1992, p. 205). Eighty per cent
of the contacts with customers in the car industry take place at the dealerships
(Spilker 1995, p 192). The majority of the contacts refer to the after-sales-service.
Therefore, dealers have a major influence on customer satisfaction with the after-
sales service. In addition, as results fram other studies have shown, this also has an
impact on global customer satisfaction (in terms of satisfaction with the car or the
brand) (Bloemer/Lemmink 1992, p. 356; Goff et al. 1997, pp. 177; Herrmann/Huber
1997, p. 15). Taking the above mentioned link into account, one way to improve
customer satisfaction scores is consequentiy, to keep the dealer body happy (Power
and Associates 1992, p. 5).
Although this relation seems apparent, the positive link between dealer and customer
satisfaction is not proven. For example, Ihere exists no empirical evidence for Ihe cau-
sality of the relation. Therefore, the goal of this paper is to gain some theoretical insights
into the relationship on the basis of preliminary findings and additional industry related
considerations as apreparation for empirical research in Ihis area.
The second paragraph provides so me terminological consideration of the two con-

structs examined in the paper. Since there exists a large body of literature on cus-
tomer satisfaction, only abrief overview, especially with respect to the car industry,
will be given (2.1). The focus will be more on the theoretical development of the
dealer satisfaction construct (2.2). The relation of customer and dealer satisfaction
will be analyzed in paragraph 3. On the basis of available research findings, first
insights will be shown in 3.1. The findings will lead to an examination of the cause-
effect-relation, which will be the conte nt of 3.2. The analysis will resuil in a set of
implications for future research in paragraph 4.
The Unk between Customer Satisfaction and Dealer Satisfaction 59
2 Conceptual Background
2.1 Customer Satisfaction in the Car Industry
For years there has been no doubt conceming the importance of customer satisfac-
ti on as the ultimate goal of industrial and service companies, because it is assumed
that customer satisfaction is a necessary condition for customer retention, loyalty
and therefore helps in realizing the economic goals of a company (Reichheld 1996;
Scheuing 1995; Stauss/Neuhaus 1997a, p. 236). As a consequence, hundreds of
academic articles have been written for more than two decades trying to explain
customer satisfaction and its related nomological network (Babin/Griffin 1998, p.
127; Perkins 1993, p. 217). In spite of the multitude of academic work, the remark of
Day (1982, p. 1) from the early 80s, that no consensus has been achieved with re-
spect to a core theory of satisfaction/dissatisfaction, is still valid today.
Researchers generally agree that customer satisfaction requires at least one indi-
vidual experience with the performance of an industrial or a service company.
Therefore, it can be described as a post-purchase evaluation. In addition, customer
satisfaction results from a subjective comparison of expected and perceived attribute
levels. In this sense, most researchers follow the disconfirmation paradigm model
(see e.g. BoltonlDrew 1991; CadottelWoodruff/Jenkins 1987; ChurchilllSurprenant
1982; Hili 1986; Oliver 1980, 1981; SwanITrawick 1981). This model holds that cus-
tomer satisfaction is related to both the size and the direction of disconfirmation, with
three potentialoutcomes: negative disconfirmation, positive disconfirmation and
confirmation/indifference (Walker 1995, p. 6).
Apart from these agreements in the literature, researchers are irresolute conceming
the nature of the components of the model (see e.g. Stauss/Neuhaus 1995, pp. 12
for a detailed description). Nevertheless, taking the current satisfaction research into
account (e.g. Bitner/Hubbbert 1994; LiljanderlStrandvik 1995; ManolOliver 1993;
Oliver 1993; WestbrooklOliver 1991; Stauss/Neuhaus 1997a, 1997b; ZeithamllBerryl
Parasuraman 1993), customer satisfaction can be assumed as a complex emotional,
cognitive and intentional construct, that represents an individual subjective evalua-
60 Alexander Decker
tion of a certain objective or relationship. It is the result of a complex comparison

process in the sense of a perceived confirmation/disconfirmation of expectations.
The multitude of objectives and relations satisfaction can refer to, will lead to a dif-
ferentiation of various satisfaction dimensions in dependence of the focused indus-
try. In this case, it is the German car industry.
Since cars are complex, valuable and expensive products, manufacturers in the
German car industry prefer to distribute their cars indirectly via a network of dealers,
who are exclusively bound by contract to the manufacturer (Diez 1995, p. 168).1 The
dealers are responsible for the sale and the after-sales service of the cars. Taking
these aspects into account, Burmann (1991) distinguishes four different types of
customer satisfaction in the car industry, considering different phases in the relation-
ship (see figure 1).2
As it can be seen trom the figure, only one of the four dimensions refers directly to the
manufacturer (product satisfaction), the other three refer to the dealers. Therefore, the
performance of the dealer body has not only an important impact on the satisfaction with
the dealership, but also an important impact on the satisfaction with the brand.
As a consequence, to secure a certain performance standard throughout the dealer net-

work, the relationship between manufacturer and dealer is based on a comprehensive
contract with rights, rules and recommendations for both parties. Since the manufacturer
still holds the power within this industry, there are more rules and recommendations for
the dealers than rights. Yet, the largest degrees of freedom for a dealer lies in the area
of the after-sales-service. Because of the characteristics of services, possibilities to
standardize the process are limited and therefore the quality of the performance cannot
be controlled in the same way as e.g. the product quality. As a consequence, dealers
have to be made to behave as intended by the manufacturer voluntarily. This is crucial to
car manufacturers for several reasons.
This so-ca/led »conlractual distribution system« is similar 10 a franchise-system.

2 A similar approach is presented by Bloemer/Lemmink (1992, p. 352). The authors idenlify
Ihree dimensions of customer salisfaction, which are the same as the dimensions in the pur-
chase and post-purchase phase in the study of Burmann (1991).
Phase Dimension
Pre-purchase phase Shopping Satisfaction
Satisfaction with the reception of and the information
about the dealerships in the area
Purehase phase: Sales Satisfaction
Satisfaction with the experiences during the purchase
process
Post-purchase phase: Product Satisfaction
Satisfaction with the car while using the product
After-Sales-Service Satisfaction
Satisfaction with the After-Sales-Service performance of
the dealer
Fig. 1: Satisfaction dimensions in the German car industry
Source: Burmann (1991, p. 251)
First, one problem of most dealerships in the German car industry is the lack of market-
ing know how and professionality (Heß 1997, p. 36). This problem can be solved in part,
if the dealers put into action the manufacturer's strategy. Second, since there is an in-
tense intra-brand competition, particularly dissatisfied dealers first try to distinguish their
dealership trom others in the dealer network with special discounts andlor special offers.
As a result, the quality of the dealers' performance and service varies enormously, due
to their lack in marketing professionality, which will lead to feelings cf uncertainty on the
side of the customers and therefore supports customer dissatisfaction.
2.2 Dealer Satisfaction in the Car Industry

2.2.1 Literature Review
While there is a lot of research done on customer satisfaction and as a result, recent
years have brought impressive gains this research field (Rust et al. 1996, p. 65), this
cannot be stated for the field of dealer satisfaction research, although there is a re-
search tradition in the Anglo-American area since the late 60s. A lot of studies were
undertaken in different industries, among them the car industry. Most of the studies
concentrated on the empirical analysis of the causes of dealer satisfaction (such as
62 Alexander Decker
power, conflict and control) and not on the development of a theoretical conceptuali-
zation. As a consequence, RuekertlChurchili (1984, p. 226) summarize the state of
the research on the basis of an intense literature review as folIows:
"In spite of the apparent importance of the concept and its central place in
a number of theoretical structures, channel member satisfaction remains a
rather primitive construct. It is almost never conceptually defined, nor ex-
plicitly operationalized, when it is used. Further, studies which attempt to
operationalize it are often poorly executed."
Since this statement is nearly 15 years old, it can be assumed that there must have been
progresses in the canceptualization of the canstruct, which is not the fact. In a large
number of the surveys caming fram the Anglo-American area, dealer satisfaction is not
still canceptually defined (se e.g. GassenheimerlSterling/Robicheaux 1989; Joseph et al.
1995; Lewis/Lambert 1991; SmithlKoenig 1985 or Yavas/Habib 1987). The same can be
stated regarding the research in the Eurape, especially with respect to several studies
undertaken in the German car industry (see e.g. BauerlMüller 1992; Diekhof 1995;
MatherlTopping 1995; Müller 1995,1996; Sewell 1992).
The majority of studies on dealer satisfaction, in which the construct is defined, use
the definition praposed by AndersonlNarus (1984, 1990) or even more frequently
SchultLittle/Pride (1985). Accordingly, dealer satisfaction is defined as
,,[ ... ] a channel member's affective attitude and feelings concerning the
domain of characteristics describing the internal environment of the chan-
nel organization and the relationship between the channel member and
other institutions in the channel arrangement. Thus, channel member sat-
isfaction is conceptualized as an affective response of individual channel
members towards salient aspects of the channel organization" (Schult
Littie/Pride 1985, S. 13)."3
3 The very simiJar definition proposed by Anderson/Narus (1984, p. 66) defines dealer satisfac-
lion as .[...) a positive affective state resulting from the appraisal of all aspects of a firm's
working relationship with another firm." Studies that use one of these !wo definitions are:
BrownlLusch/Smith (1991, p. 16), Frazier/GiII/Kale (1989, p. 57), Gaski/Nevin (1985, p. 131),
Gassenheimer/Calantone/Scully (1995, p. 10), Gassenheimer et al. (1994, p.228), Gassen-
heimer/Ramsey (1994, p. 253), Heß (1994a, p.301, 1994b, p. 265), Maier (1996); Müller
(1995; 1996); ScheerlStem (1992); Schul (1987), Shamdasani/Sheth (1995, p. 8) or Skin-
ner/Gassenheimer/Kelley (1992, p. 179).
It can be seen from this definition and through the adoption of that definition in the
aforementioned studies, that dealer satisfaction still is viewed as an uni-dimensional
emotional construct or as the affective component of an attitude. Therefore within
these studies, not even the disconfirmation paradigm is applied.
There are other studies that define dealer satisfaction rather casually by saying, it is
the relation of expected and experienced advantages of a cooperation (MeffertJ
Wöllenstein/Burmann 1996, p. 281) or by equating it to customer satisfaction (Dichtl
et al. 1997, p. 491) or job satisfaction (WiendieklMaas 1991, p. 200). As it will be
seen later, similarities exist among these constructs, but they are not identical.
There is only one definition known to the author which was proposed by Meinig
(1995, p. 2), that takes the multidimensionality of the construct into account. How-
ever, in this approach, dealer satisfaction is defined as the attitude of a dealer to-
wards the dealer-manufacturer relationship. Dealer satisfaction may be similar to an
attitude because both constructs encompass three components; however, they are
not identical (see e.g. Stauss/Neuhaus 1995, p. 16).
In summary, compared to the customer satisfaction research, previous work related

to dealer satisfaction provides only limited insights into the nature of the construct.
Therefore, the goal of the following paragraph is to gain a broader perspective on
the basis of the findings of the current customer satisfaction research.
2.2.2 Definition and conceptualization

The problem in defining and even more in conceptualizing dealer satisfaction lies in
the broad variety of industries it can refer to. Most of the industries show different
power relations, different organizational forms of their marketing channels and
therefore different characteristics in the dealer-manufacturer relationship, which can
have major impact on the components of the construct. As a consequence, the char-
acteristics regarding the German car industry have to be considered first in abrief
overview.
As mentioned above, the basis of the relationship between a dealer and a car
manufacturer is the so-called dealer contract. With the contract the manufacturer
64 Alexander Decker
licenses a number of dealers to exclusively market the brand, to secure the after-
sales and repair service and to engage in a business developed by the manufacturer
using the latter's brand name, trademarks, service marks, know how, and methods of
doing business (compare for franchise systems: Stern/EI-Ansary 1988, p. 337). The
contract regulates all rights and duties of both parties. To fulfill the requirements of
the manufacturer (e.g. corporate identity), the dealer has to invest a certain amount
of money into the relationship (e.g. costs of construction of the dealership outlet).
The relationship between the dealer and the manufacturer is long term (e.g. there is
aperiod of notice to terminate the relationship of two years), close and involving a
complex pattern of frequent (daily) interaction between the two companies. As a
consequence, dealers show high involvement. Since most of the dealerships in the
German car industry are from the middle-class-segment, there is usually only one
owner or manager to interact with the manufacturer.
As mentioned above, the findings of current customer satisfaction research will be

the basis for the definition and the conceptualization of dealer satisfaction. There-
fore, the next step is to examine whether the findings can be transferred.
Similarities between customer satisfaction and dealer satisfaction can be found, be-
cause the dealer can also be seen as a customer (SeweIl 1992, p. 121). He asks for
the products and services of the manufacturer and evaluates the performance from a
subjective point of view. In so far, it is assumed that there are a lot of equal items,
that both customer and dealer are taking into account, while evaluating their rela-
tionship (e.g. product quality or product design). Furthermore, it seems to be plausi-
ble, that dealer satisfaction also encompasses the three satisfaction components
emotion, cognition and intention. That means, analogous to customer satisfaction,
the emotional component represents an emotion towards the manufacturer, that can
be expressed through different qualities just like confidence, consent, disappoint-
ment or frustration. The cognitive component refers to an individual expectation of
the dealer with regard to the future performance capacity of the manufacturer
(Stauss/Neuhaus 1995, p. 16). While there seem to be similarities between the two
constructs regarding the emotional and the cognitive components, this cannot be
stated for the intentional component. Intentions considered with customer satisfac-
tion refer to the willingness to choose the provider again (repurchase intention), dis-
satisfied customers tend to break up a relation and not to repurchase again (for a
crilical discussion of this core thesis see Stauss/Neuhaus 1997a, p. 237, 1997b,
p. 112). With respect to the aforementioned characteristics in a channel of distribu-
tion in the car industry, dealers do not have the same intentional option as custom-
ers. Since they are contractually bound to a manufacturer for a longer period, car
dealers cannot quit from the relationship in case of dissatisfaction. For example, they
are obliged to ask for a certain amount of cars within a year. In addition, there are
sunk costs because of the high investments of the dealer in the relationship, which
are in contradiction with such an intention in the short- and medium-term. II is as-
sumed that salisfied dealers are more willing to behave as inlended than dissatisfied
dealers. Accordingly, dissatisfied dealers will probably be more willing to distinguish
themselves from the channel system and will have a higher level of conflict than
satisfied dealers (see e.g. Gaski 1984 for an overview on the satisfaction-conflict-
relation).
With respect to the long term character of the channel relationship and the high fre-
quency of interaction between both parties, it can be derived, that dealer satisfaction
has a much lower dynamic development Ihan customer salisfaction. II has been
shown, that wilh a smaller number of experiences, (customer) satisfaction shows a
larger variance. More experiences of individuals lead to more stabilized satisfaction.
Taking the aforementioned differences into account, it can be summarized, that the
current findings concerning customer satisfaction can be transferred to dealer satis-
faction. Therefore, dealer satisfaction of a car dealer will be defined as folIows:
Dealer satisfaction is a complex, emotional, cognitive and intentional con-

struct, which represents the dealer's subjective evaluation of the dealer-
manufacturer-relationship in its various aspects. It is the result of a com-
plex comparison process in the sense of a perceived confirma-
tion/disconfirmation of expectations.
66 Alexander Decker
3 The relationship between customer

satisfaction and dealer satisfaction
3.1 Consideration of preliminary research findings
As mentioned in the introduction of the paper, the importance of the dealer satisfac-
ti on construct in the car industry refers to the assumption, that there exists a positive
link between the two constructs. Nevertheless, the relation between customer satis-
faction and dealer satisfaction is not proven in detail. Most of the articles, that refer
to the link, do not provide any reference, indication or at least a theoretical argu-
mentation. Only a few support the relationship on the basis of (preliminary) findings.
To gain a first insight into the relationship, an overview of the findings will be pro-
vided.
The first remarks on the customer satisfaction/dealer satisfaction relation known to

the author refer to the publication of Power and Associates (1992, p. 5). They report
a strong positive relationship as the result of the comparison of the CSI (Customer
Satisfaction Index) and OSI (Dealer Satisfaction Index) data. As figure 2 indicates,
those manufacturers whose dealer networks view the dealer-manufacturer relation-
ship most favorably also tend to perform weil in satisfying the customer. Please note,
that in the figure the 1990 OSI results are used in conjunction with 1991 CSI results
due to the overlap of the sam pie timeframe used for each study (Power and Associ-
ates 1992, p. 5). Unfortunately, no further data or calculation has been presented.
Another survey, that gives deeper insights into the relationship, is the so-called
Dealer Satisfaction Index (OSI) of the German car industry. The research tool, simi-
lar to the J.O. Power OSI, was conceptualized as an independent method, to explore
the dealer-manufacturer relationships of all important dealer networks in Germany
(see Meinig 1995, 1997). Similar to the argumentation in the J.O. Power studies, the
nature of the 'relationship of the two constructs is explored on the basis of the calcu-
lation of a rank correlation coefficient.
1990 OSI Top Performers 1991 eSI Score

• Lexus • Tied for 1st
• Range Rover • Below Average in Truck CSI
• Acura • Tied for4 th
• Mercedes • 3rd
• Honda • Tied for 4th
1990 OSI Worst Performers 1991 eSI Score

• GMC • Below Average in Truck CSI
• BMW • Above Average
• Eagle • Below Average
• Jeep • Below Average in Truck CSI
• Suzuki • Below Average
Fig. 2: The eSI and DSI relationship in the studies 01 J.D. Power and Associates
Source: Power and Associates (1992, p. 5)
As shown in figure 3, the comparison of the ranks of various car brands within the
»Dealer Satisfaction Index (DSI)« in the German car industry (see Meinig 1995) and a
national customer satisfaction index (»Deutsches Kundenbarometer«, see Meyerl
Domach 1995)4 support the assumption of a positive relationship between the !wo con-
structs (see the brands inside the arrow, e.g. Subaru, Porsche or Mercedes-Benz with a
good ranking in both studies, Mitsubishi, Nissan or Suzuki with a medium ranking in both
studies and Saab, Seat, Peugeot or VAG with a bad ranking in both studies).5 The
Spearman Rank Correlation Coefficient in this case was r", =0.46 in 1995 (Meinig 1995,
p. 191) and even r", = 0.66 in 1997 (Meinig 1997, p. 188).
But it can also be seen in figure 3, that there are exceptions to this rule (see the
brands outside the arrow; see especially Alfa Romeo with a good ranking in CSI and
4 The »Deutsches Kundenbarometer« is an inter-industrial survey in 38 different branches. In

the survey, the satislaction with the manulacturer and the satislaction with the dealerships are
explored separately (see Meyer/Domach 1995; Reik 1995, p. 38). The calculation 01 the rank
correlation coefficient bases on the ranks w~h respect to the dealerships.
5 To get a better overview over the ranking, the ranking list is enclosed in Appendix A. The
rankings 01 the \wo surveys are not exactly corresponding to the original rankings. Since
some brands were included in only one of the \wo studies, these brands had to be eliminated
Irom the ranking (Meinig 1995, p. 188).
68 A/exander Decker
a bad ranking in OSI (17/3) or Lada with a bad ranking in CSI and a good ranking in
OSI (24/4)).
Rank eSI
~ o~ eh S~b ru
eo
j'" - 0
I'
I.t'
er~u 1/ ~ e~ ~~ s-~e
1/
V ,.,a d~
Vovo ~k da 11' 1/ ~
1/ 1/ 1'\
b .. ~ M 5 bi hi V
/
IZ
I'" .A, 1/
1/ O~i
.
::u
:J
",.
~ ~ 2 1 ~1 ~ 1 1 Vi Cl
~u uk
!I!
V 1/
V IZ
Vi :itJ~i! V I ~, IVL ~n ,a
1/ V
V V. ~o d
yu~a.,
V V o el
,
/.
V
.-:. 11'
1/
1/
"." 1'\ 1/
V
1/
ILa~a
Fig. 3: The CSI and DSI relation as a result of the rank comparison in the studies of the
Deutsches Kundenbarometer (Customer Satisfaction Index) and the Dealer Satisfaction
Index (DSI) in 1995
Source: Meinig (1995, p. 191)
To gain seme additional insights, the OSI data was compared by the author with an-
ether custemer satisfaction index, the European Customer Satisfaction Survey
(ECSS) with regard te the German results only.6 The Spearman Rank Correlatien
6 The ECSS is a survey, which is conducted by an association of the most important European
car manufacturern. In most countries, Japanese car manufacturern are enclosed in the sur-
vey, too. The concept of the ECSS is based on the concept of the J.D. Power CSI. The rank-
ing is a result of the global satisfaction scores regarding the dimensions product satisfaction,
service/maintenance, purchase/delivery, money mattem and warranty. The latter four repre-
sent the customer satisfaction with the dealer (ECSS 1997).
Coefficient regarding the years 1995 to 1997 show completely different correlations.
However, the results from the comparison with the data of the »Deutsches Kunden-
barometer« showed a higher significance. The results are presented in figure 4. 7
Year Spearman Rank Correlation Coefficient
ECSS/DSI 1995 r""=0.0985

ECSS/DSI 1996 r. =-0.0137
p
ECSS/DSI1997 r. =- 0.009
p
Fig. 4: Resulls of the rank correlation coefficient calculation with the data from the European
Customer Satisfaction Survey (ECSS) and the Dealer Satisfaction Index (OSI)
1995 -1997
Source: Calculation on the basis ofthe data from ECSS (1997); Meinig (1995, p.191, 1997,
p. 187); Meunzel (1996, p. 22).
The preliminary findings presented above indicate that the real nature of the relation
is yet unclear. A positive link may exist be!ween the !wo constructs customer satis-
faction and dealer satisfaction, but there also might be no relation at all.
The presented results are all based on the calculation of rank correlation coeffi-
cients. Since there was no further explanation for the calculation in the J.D. Power
example, only the data from !wo different studies were considered. The Spearman's
Rank Correlation Coefficient was be!ween r.p =- 0.0137 (no association) and r = sp
0.66 (a strong association). For the higher correlations, a higher significance was
obtained. This may lead to the assumption, that aleast a weak positive link exists.
However, there are some limitations to the Spearman's Rank Correlation Coefficient
(Churchill 1991, p. 864). The comparison of the coefficients is not without problems,
since the observed studies use different operationalizations of the customer satis-
7 To get a beller overview over the ranking, the ranking list is enclosed in the Appendix, too.
The ranking ot the !Wo surveys are not exactly corresponding to the original rankings tor the
same reasons as mentioned in footnote 5. Since in the ECSS the brands VW/Audi and
FiatiLancia are treated separately, the brands tied up for the same rank in the OSI ranking list
respectively.
70 Alexander Decker
faction construct. In addition, the calculations refer to a comparison of accumulated

data from several brands. Thus, important information might be excluded. Finally,
the rank correlation coefficient gives no causality indication of the relation.
Nevertheless, since no other data was available, this procedure seemed to be a rea-
sonable first step towards gaining insights into the relation. It made clear, that there
is a need for additional research in this area (compare Meinig 1995, p. 186). The
following paragraph will examine this cause-effect-relation from a more theoretical
viewpoint.
3.2 Examination of the cause-effect-relation

Initially, it was assumed, that dealer satisfaction will have positive effects on cus-
tomer satisfaction. However, the results of the preliminary findings were quiet con-
tradictory. From a theoretical point of view different cause-effect-relations could ex-
ist. Among these are the following:
• Dealer satisfaction will lead to customer satisfaction;

• Customer satisfaction will lead to dealer satisfaction;
• Dealer satisfaction and customer satisfaction are related reciprocally;
• Dealer satisfaction and customer satisfaction are related inversely;
• The two variables are not causally related at all and any empirical associa-
The Unk between Customer Satisfaction and Dea/er Satisfaction 71
The literature on internat marketing reports a similar relation between job satisfaction
and customer satisfaction. Job satisfaction is considered to be a necessary prereq-
uisite to obtain high levels of customer satisfaction (e.g. Bruhn 1995, p. 381; Töpfer
1995, p. 10). "If the business wants to satisfy the needs of its customers, it must first
satisfy the needs of its employees· (Hoffman/lngram 1992, p. 71; similar: George
1977, p. 91). Analogously, it can be argued for vertical marketing systems, that
dealer satisfaction should be a crucial prerequisite for obtaining high customer sat-
isfaction scores. For example, Joseph et al. (1995, p. 34) report in their study, that
dealers expect that partnering arrangements with the manufacturer will improve
customer satisfaction. The results of their study showed, that dealers, who perceive
to have a strong partnership with their manufacturer, are significanlly more satisfied
with their manufacturer than are dealers in »arm's length relationships« (Joseph et
al. 1995, p. 33). As a consequence, dealers in partnership-style relationships realize
many benefits - among these customer satisfaction - that dealers in arm's length
relationships do not. A reason for this can be seen in the fact, that satisfied dealers
will more likely cooperate with the manufacturer (e.g. HuntiNevin 1974, p. 186) and
probably behave as induced by the manufacturer (see chapter 2.2).
Furthermore dealers, who are satisfied with their dealer-manufacturer relationship,

will act with higher motivation (HuntiNevin 1974, p. 186). This will lead to higher ef-
fort and consequenlly, since dealers represent the point of sale/service, to higher
customer satisfaction scores. Since there is a lack of research in determining the
consequences of dealer satisfaction (Franke 1997, p. 78; Gaski 1984, p. 14), no em-
pirical data is available to get support for this argumentation. At least, for the case of
service employees, the latter relation (effort-satisfaction) has been the focus of a
survey conducted by Mohr/Bitner (1995). Mohr/Bitner (1995, p. 250) found, that the
customer's perception of the employee's effort is positively related to customer sat-
isfaction, even if the service outcome is statistically controlled (the perceived out-
come was held constant for each industry condition examined in the study). In addi-
tion, Hoffmanllngram (1992, p.68) report, that overall job satisfaction is positively
related to customer-oriented behavior.
However, there is a problem in transferring these findings onto the dealer satisfac-
tion context. Dealer satisfaction and job satisfaction may be similar (e.g. in both
72 Alexander Decker
situations there exists a mutual dependence of both parties, an ongoing nature of

the interactions and many intangible and tangible features that enable the individu-
als to form evaluations of their partner; RuekertlChurchili 1984, p. 227), but they are
not identical. The differences between the constructs refer especially to the inten-
tional component. An employee has a better possibility to terminate the relationship
in case of dissatisfaction. As mentioned above, because of the special characteris-
tics of the dealer-manufacturer relationship, this does not apply to car dealers. In
addition, employees can reduce their performance in case of dissatisfaction. A dis-
satisfied dealer who chooses to perform at a low level, would erode his own rewards
(Lewis/Lambert 1991, p. 209). As a consequence, it can be argued that the previ-
ously mentioned argumentation might be valid for dealer satisfaction, but not neces-
sarily for dealer dissatisfaction. Furthermore, the findings in the studies of
Mohr/Bitner (1995) and Hoffman/lngram (1992) refer to a calculation of correlation
coefficients as weil. That means, it is possible, that the relation may be converse.
This will lead us to the consideration of a second cause-effect relation.
3.2.2 Customer satisfaction will lead to dealer

satisfaction
Taking the aforementioned findings into account, it can also be assumed, that cus-
tomer satisfaction will lead to dealer satisfaction. That means, an increase (de-
crease) in customer satisfaction is accompanied by an increase (decrease) in dealer
satisfaction. In this sense, a possible argumentation may be presented as folIows:
since the reported positive relation between effort and customer satisfaction
(Mohr/Bitner 1995) gives no advice about the direction of the relation, high customer
satisfaction scores can be seen as a positive confirmation and therefore will lead to
higher efforts and higher dealer satisfaction.
In addition, there is another plausible explanation for this direction of the relation.
Since most of the manufacturers in the German car industry have installed perform-
ance-oriented dealer reimbursement systems, the results of customer satisfaction
surveys have become apart of these systems. Dealers, who obtain above average
customer satisfaction scores, receive an additional discount on the price of the prod-
ucts (see e.g. Meinig/RennertlHeß 1993, p. 33). Hence, this will have a positive im-
pact on the dealer's income. As recent dealer satisfaction studies have shown, the
satisfaction with financial aspects of the relationship (especially the dealer's income)
have a major impact on global dealer satisfaction (Heß 1994a, p. 309 and p. 318;
Kollenbach 1995, p. 139). Consequently, in this case, customer satisfaction leads to
dealer satisfaction, customer dissatisfaction will cause dealer dissatisfaction. As a
limitation it must be considered, that the example refers only to apart of the camp lex
dealer-manufacturer relationship.
Nevertheless, similar to the first observed cause-effect-relation in chapter 3.2.1, em-

pirical findings from job satisfaction research can be taken into account to find sup-
port for this second causality. As the studies of Crosby/Grisaffe/Marra (1994, p. 29)
and Schlesinger/Zornitsky (1991) indicate, employee satisfaction is determined, in
part, by employees' perceptions of customer satisfaction. These findings can be
seen as consistent with the above-stated exemplary assumption.
Finally, one last nation has to be considered, while investigating this second relation.
As mentioned in chapter 2.2.2, dealer satisfaction shows a less dynamic develop-
ment than customer satisfaction. As a consequence, variations in customer satisfac-
tion will not result in the same variation in dealer satisfaction at the same time.
3.2.3 Dealer satisfaction and customer satisfaction are

related reciprocally
As the description of the first two cause-effect-relations in chapter 3.2.1. and 3.2.2
have made clear both conjunctions seem to be possible from a theoretical point of
view but limitations were found for both as weil. Since the rank correlation coeffi-
cients reported in chapter 3.1 showed, that at least a weak positive relationship
could exist, it is fair to assume that the two constructs are related reciprocally (inter-
related). That means, dealer satisfaction will support customer satisfaction and vice
versa.
Ta explore which construct has astranger influence on the other with the existing
data, the following research design seems to be appropriate to gain same further
74 A/exander Decker
insights (see e.g. for the case of the relation between job satisfaction and perform-
ance Rosenstiel1981, p. 244):
first measurement of second measurement of

customer satisfaction (CS 1) customer satisfaction (CS 2)
first measurement of second measurement of

dealer satisfaction (DS 1) dealer satisfaction (DS 2)
----------------------------------------------------.~ time
Fig. 5: Design to explore the cause-effect relation between customer satisfaction and dealer
satisfaction on Ihe basis of rank correlalion coefficienls
For the calculation, data from at least two consecutive surveys is needed. A rank
correlation coefficient of the first dealer satisfaction survey data (OS 1) with the data
trom the second customer satisfaction survey data (CS 2) as weil as rank correlation
coefficient of the first customer satisfaction survey data (CS 1) with the data fram the
second dealer satisfaction survey data (OS 2) will be calculated. If the OS 1-CS 2
correlation coefficient is higher than the other (CS 1-0S 2), it is more likely, that
dealer satisfaction will lead to customer satisfaction.
For this purpose, the data presented above was used to examine the relation tol-
lowing the described procedure. The results can be seen in figure 5. 8
8 See Appendix Band C for Ihe ranking dala.

DS7CS r.p CS7DS r. p
OSI 95-ECSS 96 0,0743 ECSS 95-0SI 96 0,0426
OSI 96-ECSS 97 0,1015 ECSS 96-0SI 97 0,055
OSI95-0KB 97 9 0,5393 OKB 95-0SI 97 0,5036
OSI = Dealer Satisfaction Index

ECSS = European Customer Satisfaction Survey
OKB = Deutsches Kundenbarometer
r.p = Spearman's Rank Correlation Coeffieient
Fig. 6: Resulls of a rank correlalion coefficienl calculalion 10 explore the cause-effect relation
between customer satisfaction and dealer satisfaction
Source: Calculalion on Ihe basis ofthe data from ECSS (1997); Meinig (1995, p. 191, 1997,
p. 167); Meunzel (1996, p. 22).
As the results indicate, there are very small differences. If at all, there seems to be
same evidence for a greater impact of the causal link "dealer satisfaction on cus-
tomer satisfaction" since all three empirical established relations have slightly higher
correlation coeffieients. This can be seen as a weak support of the assumption, that
the !wo constructs are related reciprocally, especially because - again - the rank
correlation coefficient of the OSIIOKB comparison had a higher significance than the
others.
4 Summary: Implications for Future Research

Although there seems to be an apparent link between customer satisfaction and
dealer satisfaetion, no attempt has been made to explore the nature of this relation
in detail. The aim of this paper was to gain a first insight into the relationship on the
basis of theoretical considerations and same preliminary empirical findings. The pa-
per coneentrated on the ease of the German car industry.
As a result of the paper, it was shown that on the basis of preliminary empirical find-
ings there is only a weak support for the assumption of a positive relationship. In
9 Since no customer satisfaction data (Deutsches Kundenbarometer) for 1996 was available to
the author, the resulls from 1995 had to be compared with the results trom 1997.
76 Alexander Decker
addition, the direction of the connection is yet unclear. According to a more theoreti-
cal viewpoint, it seems to be more likely that the !wo constructs are at least positively
related. An additional calculation showed, that none of the two constructs has a
higher impact on the other. Limitations of these findings were shown as weil.
As it can be seen from the description, more research needs to be done in this area
in the future. Here are some of the aspects, that have to be considered:
First of all, the preliminary findings have been correlational in nature and were
based on a comparison of the ranks of different brands. Due to the calculation of
mean values over a set of items and a number of dealers and customers per brand,
it is assumed, that important information was deleted from the data. In addition, with
the comparison of different brands within one study, different situations and there-
fore different influences had an impact on the relation. As a result of these limita-
tions, a study within the marketing channel of one single brand has to be conducted
to investigate the real nature of the relation. The relation then has to be examined
through a causal longitudinal analysis.
The preliminary findings are based upon methods mainly applied in the satisfaction
research. This type of research indicates, that satisfaction can be measured validly
by single scales (e.g. from very satisfied to very dissatisfied; see Stauss/Neuhaus
1997a, p. 238). According to Stauss/Neuhaus (1995, 1997a, 1997b) not only the in-
tensity (satisfied/dissatisfied) but also the quality of satisfaction is relevant, because
- as mentioned above - satisfaction is not a one-dimensional, but a multi-
dimensional construcl. That means, that satisfaction is connected with different
emotional and cognitive states, and these different states influence future behavior
directed to a provider (Stauss/Neuhaus 1997a, p. 237). As results from several em-
pirical studies on customer satisfaction have shown, it can be expected, that indi-
viduals, who give equal satisfaction scores have different emotions and expectations
and different behavioral intentions towards the provider, and furthermore, that typical
combinations of these components form qualitatively different satisfaction types
(Stauss/Neuhaus 1997a, p. 238). It is assumed, that this connection can be trans-
ferred to the dealer satisfaction context as weil. If so, this might also have an impact
on the customer salisfaclion/dealer salisfaction relation. Therefore, il seems to be
necessary to investigate, whether there are different dealer satisfaction types, and if
they contribute to a better understanding of the relation.
Additionally, one has to keep in mind, that there are other variables as weil (like e.g.
a positive economic development), which might have an impact on the relation.
These were not explicitly considered, but should be in the future.
Finally, the last note refers to the dealer satisfaction construct. As it was shown in
the paper, dealer satisfaction can refer to a broad variety of industries. Most of them
are completely different in terms of power relations or organization of their marketing
channels. Therefore, the dealer-manufacturer relationships are characterized differ-
ently. This has an impact on the components of the construct. While investigating
the customer satisfaction/dealer satisfaction relation, this has to be considered. For
instance, in an industry, where the retailers hold more power (e.g. the food industry),
the intentional component may be more similar to the customer satisfaction con-
struct, because the retailers are not bound by contract in the same way as dealers of
the car industry.
This article attempted to give some insights into the customer satisfaction/dealer
satisfaction relation. It can be seen as a first step towards a deeper understanding.
When the implications stated above are taken together for future research together,
this field encompasses several interesting research questions.
78 A/exander Decker
References
Anderson, J.C./Narus, JA (1984): A Model of the Distributor's Perspective of Di-

stributor-Manufacturer Working Relationship, in: Journal of Marketing, Vol. 48,
No. 3, pp. 62-74.
Anderson, J.C./Narus, JA (1990): A Model of Distributor Firm and Manufacturer

Firm Working Partnerships, in: Journal of Marketing, Vol. 54, No. 1, pp. 42-58.
Babin, B.J./Griffin, M. (1998): The Nature of Satisfaction: An Updated Examination

and Analysis, in: Journal of Business Research, Vol. 41, No. 1, pp. 127-136.
Bauer, H.H.lMüller, W. (1992): Integratives Dienstleistungsmarketing - Entwick-

lungstendenzen eines Forschungsprojektes und empirische Befunde zum Au-
tomobilservice, in: WHU: Hochschulnachrichten aus der Wissenschaftlichen
Hochschule für Unternehmensführung Koblenz, no Vol., No. 3, pp. 5-16.
Bitner, M.J.lHubbert, A.R. (1994): Encounter Satisfaction versus Overall Satisfac-

tion versus Quality: The Customer's Voice, in: Rust, R.T./Oliver, R.L. (eds.):
Service Quality: New Directions in Theory and Practice, Thousand Oaks,
pp. 72-94.
Bloemer, J.M.M.lLemmink, J. (1992): The Importance of Customer Satisfaction in

Explaining Brand and Dealer Loyalty, in: Journal of Marketing Management,
Vol. 8, No. 4, pp. 351-364.
Bolton, R.N.lDrew, J.H. (1991): A Longitudinal Analysis of the Impact of Service

Changes on Consumer Attitudes; in: Journal of Marketing, Vol. 55, No. 1,
pp. 1-9.
Brown, J.R.lLusch, R.F.lSmith, L.P. (1991): Conflict and Satisfaction in an Indu-

strial Channel of Distribution, in: International Journal of Physical Distribution &
Logistics Management, Vol. 21, No. 6, pp. 15-25.
Bruggemann, A. (1974): Zur Unterscheidung verschiedener Formen der Arbeitszu-

friedenheit, in: Arbeit und Leistung, Vol. 28, No. 11, pp. 281-284.
Bruhn, M. (1995): Internes Marketing als Forschungsgebiet der Marketingwissen-

schaft, in: Bruhn, M. (ed.): Internes Marketing. Integration der Kunden- und Mit-
arbeiterorientierung, Wiesbaden, pp. 13-61.
Burmann, C. (1991): Konsumentenzufriedenheit als Determinante der Marken- und

Händlerloyalität. Das Beispiel der Automobilindustrie, in: Marketing ZFP,
Vol. 13, No. 4, pp. 249-258.
Cadotte, E.R./Woodruff, R.B.lJenkins, R.L. (1987): Expectations and Norms in

Models of Consumer Satisfaction; in: Journal of Marketing Research, Vol. 24,
No. 3, pp. 305-314.
Churchill, G.A. Jr./Surprenant, C. (1982): An Investigation into the Determinants of

Costumer Satisfaction; in: Journal of Marketing Research, Vol. 19, No. 4,
pp. 491-504.
Churchill, G.A. Jr. (1991): Marketing Research. Methodological Foundations, 5th

edition, Chicago.
Crosby, L.A./Grisaffe, D.B./Marra, T.R. (1994): The Impact of Quality and Cus-
tomer Satisfaction on Employee Commitment, in: Marketing and Research To-
day, Vol. 22, No. 1, pp. 19-31.
Day, R.L. (1982): The Next Step: Commonly Accepted Constructs for Satisfaction
Research. Paper presented at the Seventh Annual Conference on Customer
Satisfaction, Dissatisfaction, and Complaining Behavior. Knoxville.
Dichtl, E./Hardock, P./Ohlwein, M./Schellhase, R. (1997): Die Zufriedenheit des

Lebensmitteleinzelhandels als Anliegen von Markenartikelunternehmen, in: Die
Betriebswirtschaft, Vol. 57, No. 4, pp. 490-505.
Diekhof, R. (1995): Händler rügen Hersteller, in: W&V, Vol. 33, No. 37, p. 9.
Diez, W. (1995): Das Handbuch für das Automobilmarketing, Landsberg a.L.
ECSS (1997): European Customer Satisfaction Survey, Research Report.
Franke, N. (1997): Das Herstellerimage im Handel. Eine empirische Untersuchung

zum vertikalen Marketing, Berlin.
Frazier, G.L.IGiII, J.D./Kale, S.H. (1989): Dealer Dependence Levels and Recipro-
cal Actions in a Channel of Distribution in a Developing Country, in: Journal of
Gaski, J.F. (1984): The Theory of Power and Conflict in Channels of Distribution, in:
Journal of Marketing, Vol. 48, No. 3, pp. 9-29.
Gaski, J.F./Nevin, J.R. (1985): The Differential Effects of Exercised and Unexerci-
sed Power Sources in a Marketing Channel, in: Journal of Marketing Research,
Vol. 22, No. 2, pp. 130-142.
Gassenheimer, J.B.lCalantone, R.J.lSchmitz, J.M.lRobicheaux, R.A. (1994); Mo-

dels of Channel Maintenance: What is the Weaker Party to Do?, in: Journal of
Business Research, Vol. 30, No. 3, pp. 225-236.
Gassenheimer, J.B./Calantone, R.J./Scully, J.I. (1995): Supplier Involvement and

Dealer Satisfaction: Implications for Enhancing Channel Relationships, in:
Journal of Business & Industrial Marketing, Vol. 10, No. 2, pp. 7-19.
Gassenheimer, J.B./Ramsey (1994): The Impact of Dependence on Dealer Satis-

faction: A Comparison of Reseller-Supplier Relationships, in: Journal of Retai-
ling, Vol. 70, No. 3, pp. 253-266.
80 Alexander Decker
Gassenheimer. J.B./Sterling. J.U./Robicheaux. R.A. (1989): Lang-Term Channel

Member Relationships, in: International Journal of Physical Distribution and
Materials Management, Val. 19, No. 12, pp. 15-28.
George. W.R. (1977): The Retailing of Services - AChalIenging Future, in: Journal
of Retailing, Val. 53, No. 3, pp. 85-98.
Goff. B.G./Boles. J.S./Bellenger. D.N./Stojack. C. (1997): The Influence of Sales-

person Selling Behavior on Customer Satisfaction with Products, in: Journal of
Retailing, Val. 73, No. 2, pp. 171-183.
Herrmann. A./Huber. F. (1997): Kundenloyalität als Erfolgsdeterminante im Mar-

keting: Ergebnisse einer kausalanalytischen Studie im Automobilsektor, in:
Journal für Betriebswirtschaftslehre, Val. 47, No. 1, pp. 4-25.
Heß. A. (1994a): Verhaltensaspekte in vertikalen Absatzkooperationen - Die Zufrie-

denheit des Automobilhändlers, in: Meinig, W. (Hrsg.), WertschöpfungskeIle
Automobilwirtschaft. Zulieferer - Hersteller - Handel. Internationaler Wettbe-
werb und globale Herausforderungen, Wiesbaden, pp. 259-276.
Heß. A. (1994b): Konflikte in Vertraglichen Vertriebssystemen der Automobilwirt-

schaft - Theoretische und empirische Analyse, Ottobrunn.
Heß. A. (1997): Aktuelle Entwicklungen der Vertriebsgestaltung in der Automobil-

wirtschaft, in: Jahrbuch der Absatz- und Verbrauchsforschung, Val. 43, No. 1,
pp. 25-40.
Hili. D.J. (1986): Satisfaction and Consumer Services; in: Lutz, R.J. (ed.): Advances
in Consumer Research, Val. 13, Ann Arbor, pp. 311-315.
Hoffman. K.D./lngram. T.N. (1992): Service Provider Job Satisfaction and Cus-
tomer-Oriented Performance, in: Journal of Services Marketing, Val. 6, No. 2,
pp. 68-78.
Hunt. S.D./Nevin. J.R. (1974): Power in a Channel of Distribution: Sources and

Consequences, in: Journal of Marketing Research, Val. 11, No. 2, pp. 186-193.
Joseph. W.B./Gardner. J.T./Thach. S.Nernon. F. (1995): How Industrial Distribu-

tars view Distributor-Supplier Partnership Arrangements, in: Industrial Marke-
ting Management, Val. 24, No. 1, pp. 27-36.
Kollenbach. S. (1995): Positionierungsmanagement in Vertragshändlersystemen,

Frankfurt.
Lewis. M.C./ Lambert. D.M. (1991): A Model of Channel Performance, Depen-

dence, and Satisfaction, in: Journal of Retailing, Val. 67, No. 2, S. 205-225.
Liljander. V./Strandvik. T. (1995): Service Satisfaction: Exploring the Role of Emo-

tions, in: Proceedings of the Workshop on Quality Management in Services V,
Tilburg, pp. 297-318.
Maier, A. (1996): Partner oder Handlanger?, in: Kfz-Betrieb, no Vol., No. 2,

pp. 48-51.
Mano, H.lOliver, R.L. (1993): Assessing the Dimensionality and Structure of the
Consumption Experience: Evaluation, Feeling, and Satisfaction, in: Journal of
Consumer Research, Vol. 20, No. 4, pp. 451-466.
Mather, M./Topping, N. (1995): Measuring Up To Your Own Standards - Dealers'

Satisfaction with Their Supplier, in: Automotive Marketing Review, Vol. 1, No. 1,
pp. 1-6.
Meffert, H./Wöllenstein, S./Burmann, C. (1996): Erfolgswirkungen des Konflikt-

und Kooperationsverhaltens in vertraglichen Vertriebssystemen des Automo-
bilhandels, in: Marketing ZFP, Vol. 17, No. 4, pp. 279-290.
Meinig, W. (1995): Dealer Satisfaction Index 1995, Research Report, Bamberg.
Meinig, W. (1997): Dealer Satisfaction Index 1997, Research Report, Bamberg.
Meinig, W./Rennert, C.lHeß, A. (1993): Kundenorientierung als Problem des verti-

kalen Marketing in der Automobilwirtschaft, in: Marktforschung und Manage-
ment, Vol. 37, No. 1, pp. 31-37.
Meunzel, R.M. (1996): Prüfstand Partnerschaft, in: Autohaus, no Vol. No. 13,
pp. 20-22.
Meyer, A.lDomach, F. (1995): Das Deutsche Kundenbarometer 1995 - Qualität und

Zufriedenheit, München.
Mohr, L.A./Bitner, M.J. (1995): The Role of Employee Effort in Satisfaction with
Service Transactions, in: Journal of Business Research, Vol. 32, No. 3,
pp. 239-252.
Müller, W. (1995): Kundenzufriedenheit beginnt beim Hersteller, in: Kfz-Betrieb Un-

ternehmermagazin, no Vol., No. 5, pp. 38-39.
Müller, W. (1996): Angewandte Kundenzufriedenheitsforschung, in: Marktforschung

& Management, Vol. 41, No. 4, pp. 149-159.
Oliver R.L. (1980): A Cognitive Model of the Antecedents and Consequences of

Satisfaction Decisions, in: Journal of Marketing Research, Vol. 17, No. 4,
pp. 460-469.
Oliver R.L. (1981): Measurement and Evaluation of Satisfaction Processes in Retail

Settings, in: Journal of Retailing, Vol. 57, No. 3, pp. 25-48.
Oliver R.L. (1993): Cognitive, Affective, and Attribute Bases of the Satisfaction Re-
sponse, in: Journal of Consumer Research, Vol. 20, No. 3, pp. 418-430.
82 A/exander Decker
Perkins, 0.5. (1993): An Update of the Consumer Satisfaction, Dissatisfaction and

Complaining Behavior Bibliography: Revolution and Evolution, in: Journal of
Consumer Satisfaction/Dissatisfaction and Complaining Behavior, Vol. 6, No. 1,
pp. 217-279.
Power, J.D. and Associates (1992): DSI and CSI: There's a Connection, in: The
Power Report on Automotive Marketing, no Vol., No. 2, p.5
Reichheld, F.F. (1996): The Loyalty Effect, Boston.
Reik, M. (1995): Service Barometer, in: Autohaus, no Vol., No. 11, pp. 12-38.
Rosenstiel, L. v. (1981): Leistung und Zufriedenheit, in: Beckerath, P.G.I

Sauermann, P./Wiswede, G. (eds.): Handwörterbuch der Betriebspsychologie
und Betriebssoziologie, pp. 240-245.
Ruekert, R.W.lChurchill, G.A. Jr. (1984): Reliability and Validity of Alternative

Measures of Channel Member Satisfaction, in: Journal of Marketing Research,
Vol. 21, No. 2, pp. 226-233.
Rust, R.T.lStewart, G.L.lMiller, H.lPielack, D. (1996): The Satisfaction and Reten-

tion of Frontline Employees. A Customer Satisfaction Approach, in: Internation-
ais Journal of Service Industry Management, Vol. 7, No. 5, pp. 62-80.
Scheer, L.K.lStern, L.W. (1992): The Effect on Influence Type and Performance
Outcomes on Attitude Toward the Influence, in: Journal of Marketing Research,
Vol. 29, No. 1, pp. 128-142.
Scheuing, E.E. (1995): Creating Customers for Life, Portland.
Schlesinger, L.A.lZornitsky, J. (1991): Job Satisfaction, Service Capability, and

Customer Satisfaction, in: Human Recourse Planning, Vol. 14, No. 2,
pp. 141-149.
Schul, P.L. (1987): An Investigation of Path-Goal Leadership Theory and Its Impact
on Intrachannel Conflict and Satisfaction. in: Journal of the Academy of Marke-
ting Science, Vol. 15, No. 4, pp. 42-52.
Schul, P.L.lLittle, T.E.lPride, W.M. (1985): Channel Climate: Its Impact on Channel
Members' Satisfaction, in: Journal of Retailing, Vol. 61, No. 2, pp. 9-38.
Sewell, R. (1992): The Dealer as a Customer, in: E.S.O.M.A.R. (Hrsg.): International

Automotive Marketing Conference, Genf, pp. 121-155.
Shamdasani, P.N./Sheth, J.N. (1995): An Experimental Approach to Investigating

Satisfaction and Continuity in Marketing Alliances, in: European Journal of
Skinner, S.J.lGassenheimer, J.B.lKelley, S.W. (1992): Cooperation in Supplier-

Dealer Relations, in: Journal of Retailing, Vol. 68, No. 2, pp. 174-193.
Smith, L.P.lKoenig, H.F. (1985): Meta-Analysis on the Relationship between Satis-

faction and Manifest Conflict in Marketing Channels, in: Lusch, R.F. et al.
(Hrsg.): American Marketing Association, Proceedings Series No. 51, Chicago,
pp. 341-347.
Spilker, E.M. (1995): Wettbewerbsvorteile durch eine aktive und systematische

Kundenorientierung, in: Reuss, H.lMüller, W. (eds.): Wettbewerbsvorteile im
Automobilhandel, Frankfurt, pp. 191-219.
Stauss, B./Neuhaus, P. (1995): Das Qualitative Zufriedenheitsmodell (QZM), Dis-

kussionsbeiträge der Wirtschaftswissenschaftlichen Fakultät Ingolstadt, Katho-
lische Universität Eichstätt, No. 66, Ingolstadt.
Stauss, B./Neuhaus, P. (1997a): The Qualitative Satisfaction Model, in: Interna-

tional Journal of Service Industry Management, Vol. 8, No. 3, pp. 236-249.
Stauss, B.lNeuhaus, P. (1997b): The Dissatisfaction Potential of Satisfied Custo-

mers - The Qualitative Satisfaction Model, in: Mühlbacher, H.lFlipo, J.-P.
(eds.): Advances in Services Marketing, Wiesbaden, pp. 111-131.
Stern, L.W./EI-Ansary, A.I. (1988): Marketing Channels, 3'd edition, Englewood

Cliffs.
Swan, J.E./Trawick, F. (1981): Disconfirmation of Expectations and Satisfaction

with a Retail Service; in: Journal of Retailing, Vol. 57, No. 3, pp. 49 - 67.
Töpfer, A. (1995): Kunden-Zufriedenheit durch Mitarbeiter-Zufriedenheit, in: Perso-

nalwirtschaft, no Vol., No. 8, pp. 10-15.
Thurow, W. (1992): Differenzierungsstrategien in der Automobilindustrie, in: Wer-

beforschung & Praxis, no Vol., No. 6, pp. 202-205.
Walker, J.L. (1995): Service Encounter Satisfaction: Conceptualized, in: Journal of

Services Marketing, Vol. 9, No. 1, pp. 5-14.
Westbrook, R.A.lOliver, R.L. (1991): The Dimensionality of Consumption Emotion

Patterns and Consumer Satisfaction, in: Journal of Consumer Research, Vol.
18, No. 1, pp. 84-91.
Wiendieck, G.lMaas, P. (1991): Analyse und Intervention: Arbeitszufriedenheitsfor-

schung als sozialer Prozeß, in: Fischer, L. (ed.): Arbeitszufriedenheit, Stuttgart,
pp. 199-212.
Yavas, U./Habib, G. (1987): Correlates of Franchisee Satisfaction: The Case of the

Saudi Car Dealers, in: International Journal of Physical Distribution & Materials
Management, Vol. 17, No. 3, pp. 46-55.
Zeithaml, V.A./Berry, L.L.lParasuraman, A. (1993): The Nature and Determinants

of Customer Expectations of Service, in: Journal of the Academy of Marketing
Science, Vol. 21, No. 1, pp. 1-12.
The Link between Customer Satisfaction and Dealer Satisfaction 85
Appendix
A Ranking list of the »Deutsches Kundenbarometer« (KB) and the Dealer

Satisfaction Index (DSI)
B Ranking lists European Customer Satisfaction Survey« (ECSS) 1995 -

1997 and »Dealer Satisfaction Index« (OS I) 1995 - 1997
C Ranking list of the »Deutsches Kundenbarometer« (KB) 1995 - 1996

and the Dealer Satisfaction Index (DSI) 1995 - 1997
86 Alexander Decker
AppendixA
Brand KB 1995 0511995

Alfa-Romeo 3 17
BMW 10 13
CitroE!n 15 18
FiatiLancia 16 7
Ford 14 12
Honda 17 8
Hyundai 19 20
Lada 24 4
Mazda 6 5
Mercedes-Benz 5 3
Mitsubishi 9 9
Nissan 12 10
Opel 18 11
Peugeot 21 24
Porsche 2 6
Renault 4 15
Saab 20 21
Seat 22 23
Skoda 8 14
Subaru 1 1
Suzuki 13 16
Toyota 11 2
VAG 23 22
Volvo 7 19
Appendix A: Ranking list of the .Deutsches Kundenbarometer« (KB) and the Dealer
Satisfaction Index (DSI)
Source: Meinig (1995, p. 191)
Appendix B
Brand ECSS OSI ECSS OSI ECSS OSI

1995 1995 1996 1996 1997 1997
Alfa-Romeo 21 15 19 12 22 11
Audi 3 19 2 20 6 21
BMW 7 11 8 2 6 2
Citroen 10 16 12 22 8 20
Fiat 17 5 18 4 14 11
Ford 10 10 8 18 18 14
Honda 17 7 15 6 18 4
Jaguar 19 14 19 10 14 1
Lancia 20 5 19 5 21 11
Mazda 7 4 5 9 8 7
Mercedes-Benz 1 2 1 1 1 3
Nissan 7 8 8 3 8 6
Opel 10 9 12 17 8 10
Peugeot 10 22 8 15 8 9
Renault 3 13 5 6 4 16
Rover 15 3 15 16 18 19
Saab 3 18 5 14 17 7
Seat 22 21 22 19 8 17
Skoda 15 12 15 6 1 15
Toyota 1 1 2 10 1 5
Volkswagen 3 19 4 20 4 21
Volvo 10 17 12 13 14 18
Appendix S: Ranking lists European Customer Satisfaction Survey« (ECSS) 1995 - 1997 and
»Dealer Salisfaction Index« (OS I) 1995 - 1997
Source: According to ECSS (1997); Meinig (1995, p. 191, 1997, p. 187), Meunzel (1996,
p.22)
Please note, that the ranks of the DSI are not exactly corresponding to the original
ranks in the surveys. Same of the brands were not included in the ECSS. Therefore,
these brands had to be eliminated from the comparison. The ranks had to be num-
bered again.
88 Alexander Decker
Appendix C
Brand KB 1995 0511995 KB 1997 0511997

BMW 5 5 5 1
Citroen 9 12 11 14
Fiat 10 4 15 10
Ford 8 10 14 11
Honda 11 5 2 4
Mazda 3 3 1 7
Mercedes-Benz 2 2 3 3
Mitsubishi 4 6 7 2
Nissan 7 7 6 6
Opel 12 8 12 9
Peugeot 13 15 9 8
Renault 1 11 10 12
Seat 14 14 13 13
Toyota 6 1 4 5
VAG 15 13 8 15
Appendix C: Ranking list otthe »Deutsches Kundenbarometer« (KB) 1995 -1996 and the
Dealer Satistaction Index (051) 1995 - 1997
Source: According to Meinig (1995, p. 191,1997, p. 187), Meunzel (1996, p. 22)
Please note, that the ranks of the 1995 data are not exactly corresponding to the
data in Appendix A Some of the brands from the 1995 comparison were not in-
cluded in the 1997 comparison (see Meinig (1997, p. 187». To compare the data,
the ranks had to be numbered again.
Criticality of Criticallncidents in Customer Relationships
Ba Edvardsson and Tore Strandvik
Contents
Abstract 91
Introduction and Aim 91
2 Points of Departure 92
2.1 Critieal Ineidents as Snapshots 93
2.2 Critieal Ineident Studies - An Overview 94
2.3 Interaetion Levels Between the Customer and the Service
Provider 95
2.4 Layers of Reaetions 96
2.5 ADefinition of CI in a Relationship Context 98
3 Our Tentative Framework for Analysis 99
4 Using our Tentative Framework for Re-Analyzing So me

Empirical Studies 103
5 Future Research 106
References
Criticality of Criticallncidents in Customer Relationships 91
Abstract
In this article we challenge the traditional seNice encounter oriented view on critical
incidents. Instead we propose an approach based on a customer relationship view.
The main research questions are: "In what way do critical incidents affect the cus-
tomer relationship? How critical are critical incidents? When and why are critical inci-
dents critical?". We develop and present a tentative framework for analyzing critical
incidents in customer relationships. Our framework is built on two main contextual
dimensions; (1) the situational context which can be divided into an internal, actor-
based and an external, environmental context and (2) the time dimension which
captures the history, the present state and the future of the relationship.
1 Introduction and aim

The Criticallncident Technique (CIT) has widely been used to measure and analyze
customer perceived service quality defects and customer delight (see e.g. Bitner et
al. 1985, 1990; Edvardsson 1988, 1992). The focus has been on the interaction as
such and has not included situation specific or relationship related variables. It has
been argued that incidents that the customer remembers constitute factors that are
important for customer perceive quality and thus should be recognized in order to
achieve higher customer satisfaction, customer loyalty and customer relationship
profitability.
The managerial usefulness of a focus on critical incidents (CI) is considered to be

the diagnostic value of the information which can be used for identifying areas with
potential for improvements in the service offering and service production process
(Olsen 1992, Stauss 1993). Furthermore it has been argued that these often repeti-
tive, negative critical incidents are built in when new services are developed (Ed-
vardsson 1996). Data on critical incidents thus may be used both as a basis for de-
veloping belter services and when improving quality of existing services.
But how critical are critical incidents studied in the established, traditional way? We
will argue in this paper that it is necessary to consider critical incidents with a larger
scope than before if we want to understand how critical critical incidents are. Most
92 Bo Edvardsson / Tore Strandvik
studies focus on the critical incident per se and in few studies there has been an ex-
plicit customer relationship foeus and a context related framework for analysis and
interpretation. Such a contextual oriented perspective does not only take an interest
in the incident as such, but also in the outcome or impact, that is what the incident
leads to concerning the specific customer-service provider relationship.
Which effect do critical incidents have on a customer's relationship with a service

provider? Naturally, there are often several different and interacting factors that de-
termine the criticality of a critical incident in terms of the future and development of a
customer relationship (Johnston 1995). Some conditions seems to be more critical
than others are and some factors are satisfiers and others dissatisfiers. The main
aim with this article is to develop an analytical framework where we capture not only
the critical incident as such but also relevant dimensions of the context. The ambi-
tion is to contribute to a beUer understanding of the dynamics of the influence of
critical incidents on the customer relationship. Our discussion is grounded in both
service and relationship marketing literature as weil as in earlier empirical studies
that broadly have employed the critical incident approach.
The article is structured in the following way. First we describe our points of depar-
ture, both theoretical and empirieal. In the following section we develop and present
a tentative framework for analyzing critical incidents in a relationship context. Our
framework or conceptual model, which is widening the scope of critical incidents, is
built on two main dimensions; (1) the situational dimension which is divided into an
internal and an external context and (2) the time dimension which depicts the history
and future of the relationship. We use our frame of reference for a discussion of the
data collection methods used and results obtained in some previous CI-studies. Our
focus is on the effect of the critical incident on the relationship with the customer.
Finally, we present some avenues for further research in this area.
2 Points of departure
In this seetion critical incidents are characterized, the change from a transactional
view to a relational view in marketing is commented on, and adefinition of critical
incidents in a relational view is presented. This involves seeing relationships analyti-
cally as different levels of interaction between the customer and the service pravider.
In order to analytically depict customer reactions a three-Iayer model, including
emotions, cognition and customer action is proposed.
2.1 Critical incidents as snapshots

Critical incidents are defined as interaction incidents, which the customer perceives
or remembers as unusually positive or negative when asked about them. Customers
recall them and tell them as stories (Stauss 1993). The moment of interaction, i.e.
the service encounter between the customer and the firm has been intensively stud-
ied in service marketing research. The interaction has been considered the essence
of the service, the moment-of-truth, from the customer's point of view (Bit-
ner/BoomsfTetreault 1990). Typically, researchers have focused on finding the
service quality dimensions or determinants by using traditional content analysis to
analyze these incident-based stories. A critical incident represents thus an event
that differs from what the customer considers normal or expects in the service in
question. These differences are studied as such without a consideration of their
context. On an aggregate level, factars that can be found to cause either positive or
negative incidents to a large extent among the customers are considered important
fram a managerial point of view. It is implicitly assumed that these incidents will have
an effect on the customer's evaluation of the service, the relationship and ultimately
on their repurchase decisions.
The focus in traditional CI-studies is always on one single incident, not on particular
combinations of incidents or the accumulation of incidents over time. It is often im-
plicitly assumed in critical incident (CI) studies that all CI are important and that
negative incidents reflect defects in service quality that should be corrected. It is
also assumed that customers are similar in their sensitivity to experience critical in-
cidents and in their reactions to these incidents. Furthermore, the context in which
the incident take place is not considered in the analysis We argue that the reason
why critical incident studies have focused on single incidents alone is based on the
transaction-oriented perspective that has been applied in service quality and service
management research. Only the relationship perspective will suggest that all critical
94 Bo Edvardsson I Tore Strandvik
ineidents may not affeet the eustomer relationships in a similar way. Same ineidents
may have a more severe effeet than others. It should also be noted that the effeet of
positive eritieal ineidents might be different from the effeet of negative eritieal inei-
dents, as they represent satisfiers and dissatisfiers respeetively (Johnston 1995). A
single eritieal ineident may have a different effeet if it is a first time experienee rather
than a frequently oeeurring ineiden!. Different eustomers may have a different sensi-
tivity to eritieal ineidents. The eritieality of the ineident may partly reside in the eus-
tomer.
2.2 Critical Incident studies - an overview

Roos (1996) and Roos/Strandvik (1996) analyzed studies that have employed same
variant of eritical ineident teehnique aeeording to how the initial state, trigger, proe-
ess, and final outeome were depieted. It ean be observed that most of the studies
have foeused on determinants of the eritieal ineident without reeognizing the under-
Iying processes and the eontext. There are same early exeeptions, for example, Ed-
vardsson (1988) and (1992). In order to study switehing behavior Keaveney (1995)
applied CIT and she was able to identify same switching determinants. In the ease
eritieal ineidents have been seen as processes they have been eonsidered as
bracketed by the partieular episode studied (i.e. Olsen 1992). They thus represent a
pieture of a micro-proeess from a relationship perspective. Olsen struetured the inei-
dent into the triggering factar, further critical steps, and the final outeome of the inci-
den!. He, however, does not classify the incidents based on the triggering factor but
on what he interprets to be the eore problem. The SIT (Sequential Ineident Teeh-
nique) (StausslWeinlieh 1995) is also based on an episode perspeetive although it
systematieally leads the eustomer through the whole proeess of interactions when
the eustomer uses a service. A study on relationship strength by StrandviklLiljander
(1994) is taking a different perspeetive. There the eritieal ineidents are based on
eustomers' relationships with their bank and the interview retrieves retrospectively
the incidents and the outeome. In Roos' (1996, 1998) study of customers switching
behavior in supermarkets retrospective interviews were used to reveal the processes
behind dissolution of a eustomer relationship. A similar approach has been used by
Stewart (1997) in her a study of the dissolution of bank relationships.
Criticality o( Criticallncidents in Customer Relationships 95
2.3 Interaction levels between the customer and

the service provider
A relational perspective on the interaction between the customer and the service
provider brings forward a need to differentiate between different analysis units or
levels of interaction. LiljanderlStrandvik (1995) suggested a division into act, epi-
sode and relationship levels. This model has been further refined by Holmlund
(1996, 1997) to include also a sequence of episodes representing an epoque in the
relationship, and by Wrange (1997) to include a step as a smaller part of an act. It is
thus suggested that from an analytical point of view it would be helpful to consider
the whole relationship as hierarchically structured interaction units.
Holmlund (1996) has explicitly discussed the dynamics of connections between

these units and concluded that very little is known about these processes (figure 1).
She also suggests that each interaction level can be divided into a process dimen-
sion and an outcome dimension which both influence the customer's total evaluation
of the interaction. For example, an episode contains both a process of interaction
and an outcome of the interaction.
....-
Relationship
level
Sequence
level
Episode
level
Action
level II II I I I II
=good_ lity
• :: neutraJ quality
• =poor quality
Fig. 1: Different interaction levels in a business relationship

Source: Holmlund (1996, p. 136)
A customer may be dissatisfied with the process but if the outcome is satisfying the
customer may still consider the episode acceptable. It is, however, clear that the ef-
fect of critical incident, which can be defined as a step, action or episode is depend-
ent on the current overall evaluation of the whole relationship, how often the critical
incident has occurred before and the customer's evaluation of wh at may happen in
the future. The relationship can in other words be seen as an ongoing process
where new experiences are added to and shape the overall evaluation in a dialectic
way. In order to understand the criticality of critical incidents different hierarchical
units of interaction could be used as a framework. This approach would represent
one way to structure the time dimension and process in a customer relationship.
2.4 Layers of reactions

Another aspect of the context in which the relationship and the critical incident is
embedded is the actor-internal repertoire of emotions, cognition and actions. A
continuing relationship contains all of these elements. In the literature the role of
several concepts in a customer relationship like perceived service quality, perceived
sacrifice, perceived value, commitment, bonds, trust and relationship strength have
been discussed (see LiljanderlStrandvik 1995; Storbacka/StrandviklGrönroos 1994).
In a model summarizing many of the interrelationships between perceptual concepts
and customer actions also the role of critical incidents have been indicated (Stor-
backa/StrandviklGrönroos 1994). In their model critical incidents are considered as
episodes that have an effect on the relationship in co nt rast to how critical incidents
have been defined traditionally. These critical episodes are proposed to be a factor
beside customer satisfaction, customer commitment, bonds and customer perceived
alternatives that has an influence on relationship strength and relationship longevity.
Emotions Negative No emotions Positive

emotions Neutral emotions emotions
Cognition Negative No cognition Positive

cognition Neutral cognition cognition
Actions Negative
No actions
Positive
actions actions
Fig. 2: Different layers of reactions
There has recently been a discussion about different layers of concepts like satis-
faction, commitment and loyalty. Stauss/Neuhaus (1997) have suggested that satis-
faction should be recognized as having a quality, containing en emotional, a cogni-
tive and an intentional component instead of being considered as one-dimensional
concept. Several types of loyalty have been identified. Oliver discusses four different
levels of loyalty, affective loyalty, cognitive loyalty, conative loyalty and action loyalty
(Oliver 1997). Commitment is identified as an important factor in customer relation-
ships (Morgan/Hunt 1996). Commitment is considered to consist of both an attitudi-
nal part and behavioral activities. LiljanderlStrandvik (1995) have suggested that in
customer relationships not only the strength but also the direction of commitment
should be taken into account. Commitment could thus, be categorized into three
categories, positive commitment, indifference and negative commitment. Negative
commitment would for example, depict situations when the customer strives to end a
relationship but is hindered by exit barriers.
In figure 2 a three-Iayer model (emotion, cognition and action) is proposed as a way

10 depict a cuslomer's overall evaluation of a relationship Positive actions refer to
positive word-of-moulh, part-time marketing activilies and repeat purchasing. Nega-
live aclions refer to negative word-of-mouth, de-marketing and deliberate harmful
actions againsl the other party. No aclions may indicate thaI no relationship exists.
This is however a matter of definition. LiljanderlSlrandvik (1995) have suggested
that a minimum requirement for a relationship is that the customer has made a con-
tract with the service provider (bank, telephone services, electricity etc.) or thaI the
98 Ba Edvardsson / Tore Strandvik
customer has made a purchase at least twice. This definition is, however, only one
way to define a relationship.
A customer may in the similar manner have no cognitive activity regarding the serv-
ice provider, or may have a positive or negative perception (refers to concepts like
service quality, service value, satisfaction) of the company and service. Even if the
customer has a negative perception of the service provider he/she may repeatedly
buy the service. In these cases there are other factors (bonds as exit barriers) (Lil-
janderlStrandvik 1995) that prevent the customer from switching. It is also possible
to find customers that are indifferent towards the service provider and show no inter-
est in the service but are loyal customers making repeat purehases
(StrandviklLiljander 1994). Different configurations of action and cognition states are
therefore proposed to be possible.
Customer emotions can in a similar way be divided into three categories, positive
emotions, neutral or no emotions, and negative emotions (LiljanderlStrandvik 1997).
We propose that this three-Iayer model could be used as an analytical tool to depict
the effect of critical incidents.
2.5 Adefinition of CI in a relationship context

If the focus is on the relationship rather than on the service a different definition of a
critical incident is warranted. A critical incident can thus be defined as an interac-
tion or factor that changes the customer's evaluation of the relationship. The
origin of the change may be in 1) the interaction, 2) in a change in the customer, 3)
in a change in the service provider, or 4) in a change in the environment comprising
both competition and other factors. It is quite possible to focus on the subgroup of
critical incidents that are caused by the interaction itself. A critical incident should
also be perceived by the customer as diverging from the normal or expected and is
thus subjectively defined. A critical incident may lead to a strengthening of the rela-
tionship or a weakening or even dissolution of the relationship. Roos/Strandvik
(1996) have suggested that the termination of a relationship may have different du-
ration, some relationships are ended quickly others dissolve over a longer period of
time.
The criticality of a critical incident would be based on the customer's reactions and
the effects on the three layers, emotions, cognition and actions. Changes could be
detected only in emotions, cognition or action but also in combinations of these. The
most severe effect of a critical incident can be defined as a change in customer ac-
tions.
From a management point of view it is not only the effect on one relationship that is
in focus. The same CI may happen often, which considering the whole customer
base, on an aggregate level thus has an overall effect on customers. This aggregate
level perspective may represent another view on the criticality of critical incidents
that is directly related to a managerial perspective. A certain type of critical incidents
may, for example, happen frequently or concern an important customer segment,
thus, increasing the criticality of the incident from a managerial point of view.
(StrandviklFriman 1998)
3 Our tentative tramework tor analysis

In this section we present our tentative framework for describing and analyzing Cis.
We widen the scope of the analysis by considering the context in wh ich the Cis take
place. Re-contextualisation, which opens up for re-analyzation of data and results
fram CI-studies, is the basis for our framework. We have chosen to use the relation-
ship between service provider and the customer as the mega-context and within this
context we focus on two distinct but at the same time interdependent context dimen-
sions: the time dimension and the situational dimension (StrandvikITörnroos 1997).
External context
Com petition
Type of service
Other factors
Critical
Relationship Incident Relationship
history in a Relationship future
Length of relatlonsh Ip Probability of CI
Recency and Irequency Levels of Interactton occurrence
01 Cllmportance Proeess elements Importance 01
of relationship Layers of reactlon relation.hip
Internal context
Commitment
Involvement
Satisfaction, quality
Bonds
Perceived alternatives
Fig. 3: A tentative framework for analysis

Source: Adapted from StrandvikfTömroos (1997)
The time dimension has to do with the history, present time and future of the rela-
tionship (see Figure 3). Important aspects of the time dimension are the character of
the relationship e.g. (1) the length of the relationship, (2) the importance and inten-
sity of the relationship to the customer and (3) experience of critical incidents, both
in terms of number of criticality and recency. The customer mayaiso include future-
oriented aspects into the evaluation of a relationship. If we want to evaluate the criti-
cality of a critical incident, one way would be to measure the cognitive, emotional
and behavioral changes over time.
The second dimension is the situational dimension, which has to do with the inter-
nal and external conditions of the relationship. The situational context is changing
over time and we argue that critical incidents should be analyzed within this context
at the time when the incident occurred. The criticality of critical incidents may differ
over time and between customers. In our frame-of-reference, dialectic processes
take plaee in the interseetion between these two contextual dimensions. When we
switch from the episode level to the relationship level we see different aspects of a
critical incident and our understanding will inerease.
Customers are human beings and the reactions in a critical incident situation may be
very different. A combination of eognitive, emotional and behavior-based processes
may be triggered. The outcome may be observed in terms of full or partial switching
but mayaiso result in a negative eognitive and emotional reaclion which in the future
may result in a changed behavior. One way to arrive at increased understanding of
the eritieality of CI might be to focus on course, character and outcome of these pro-
cesses (Edvardsson 1988, 1992). Roos/Strandvik (1996) and Roos (1998) have
suggested that termination processes can be analyzed by foeusing on the initial
state of the relationship, the trigger starting the termination process, the termination
process in itself and the outcome of the process. We propose that the same frame
can be used more generally to study the effect of critical incidents regardless if the
outcome is dissolution of the relationship or a strengthening of the relationship.
The initial state can be described along different dimensions capturing the cus-
tomer's eurrent satisfaction level, perceived quality, eommitment, involvement, bonds
and perceived alternatives. The trigger can be generalized to reasons for a eritical
incident to arise. These impulses can reside in the interaction, in the customer, in
the firm or in the environment. This trigger leads to a process initiated by the eritieal
incident that can take several different paths. Eaeh proeess ultimately ends in an
outcome. The outcome is the factor in focus in this study. In prineiple each critieal
incident can lead to no change, a change in positive direction or a change in nega-
tive direction on each of the three layers (emotions, cognition and action).
To summarize, the approach suggested in this paper may highlight important as-
pects, which will help us to re-contextualize and thus creating a better basis for ana-
Iyzing CI. The reasons and faetors influencing a eustomer's perception of critieal in-
eident in a relationship are thus suggested 10 be based on the customer's pereeption
of the relationship over time and perception of the internal and external eontext. The
following aspects (figure 4) of each contextual dimension were discussed and intro-
duced in figure 3. They represent elements of the framework that we consider im-
portant to focus on in further empirical research.
Actor based variables (internal context)

1) Customer commitment (Ioyalty, relationship strength)
2) Involvement (need, activities, importance)
3) Satisfaction, perceived service quality
4) Perceived alternatives
5) Bonds
Environment based variables (external context)

1) Competition
2) Type of service I industry logic (e.g. discrete, continuous)
3) Other situational factars (e.g. infrastructure, pOlitical, legal and social
conditions)
Variables related to the interaction in the relationship

1) Levels of interaction
2) Initial state-trigger-process-outcome
3) Layers of reactions
Variables related to the history of the relationship

1) Length of relationship
2) Recency and frequency of critical incidents
3) Importance of relationship
Variables related to the future of the relationship

1) Probability of CI occurrence
2) Importance of relationship
Fig. 4: Contextual aspects of critical incidents
Each variable represents an aspect that may influence the criticality of a critical inci-
den!.
Different types of services: A distinction between episode dominated and relation-

ship dominated services can be made according to whether the service is of discrete
or continuous nature (LiljanderlStrandvik 1994). Concerning discrete services cus-
tomers make aseparate decision each time which service provider to use. The
service mayaiso be of a continuous nature (the customer makes a contract about
service delivery with the service provider). There are also services, wh ich fall in-
between discrete and continuous services.
They have a long duration but are of a discrete type, e.g. holiday travel packages,
educational services and medical services. These services can consist of many epi-
sodes and acts that taken together can be described as a relationship. A distinction
can also be made according to usage frequency. Services of a discrete type, which
are seldom used, can be said to be episode dominated. A relationship perspective
will become more important the higher the frequency of purchases or contacts.
We may assume that the importance of a relationship approach when analyzing

critical incidents is more important when we have to da with relationship dominated
services compared to episode dominated services. Gur hypothesis is that individual
critical incidents are more critical in episode dominated service contexts than in
other service contexts. Relationship domination is, however, not necessarily only
linked to the service per se but mayaiso be linked to individual customers and dif-
fers between customers.
4 Using our tentative framework for re-

analyzing so me empirical studies
In this section we will use our tentative framework for re-analyzing so me previous
empirical CI-studies. We will concentrate on two different types of services accord-
ing to the categorization above specifying relationship domination. Airline services
represent discrete services with a high frequency when considering business travel-
ers. Bank services are continuous services based on a contract between the cus-
tomer and the bank.
What do we see and capture and what do we not capture? Since we have no spe-
cific customer related data on the history of the relationship, e.g. previous critical
incidents in this specific service provider relationship, it is not possible to capture the
time dimension in the relationship.
104 Ba Edvardsson / Tore Strandvik
When it comes to the situational dimension, in the studies below, we have some
context-related data, e.g. on the competitive situation in the market. We start with
studies with data from the domestic airline market in Sweden and the U.S. The
Swedish domestic airline market in 1992 may best be characterized as a monopoly
and, at the time for the data-collection for the 1996 study de-regulated but still domi-
nated by one airline, Scandinavian Airline System (SAS). In the U.S. on the other
hand, the market has been characterized by keen competition for at least two dec-
ades and most often there are a number of alternative choices to many destinations.
The criti- Domestic Airlines in Domestic Retail bank in

calityof airline in the U.S. airline in Sweden
critical in- Sweden Sweden
eidents on
the rela- 320 CI 241 CI 200 CI 272 CI
tionship Edvardsson Edvardssonl Edvardsson Olsen 1992
1992 Bejou 1994 1996
Broken 3% 15% 2% 20%
Weaker 13% 44% 15% 35%
Unchanged 80% 40% 78% 40%
Stronger 4% 1% 5% 5%
Fig. 5: The crilicalily ot critical incidenls according 10 some empirical studies
Figure 5 above shows that between 3-15% of the critical incidents have resulted in a
broken relationship between the customer and the airline. The results show that only
2-3% of the customers in the Swedish studies has chosen some other airline or
transportation alternative. The situation in the U.S. is very different with many com-
peting airlines. Here we also find that 15% of the studied critical incidents resulted in
a broken relationship with the service provider. If we add the Cis where the relation-
ship are weaker (44%) after the reported incident, we find that 59% of the incidents
had a negative effect on the relationship compared to 17% in the Swedish sludy. If
we relale this data to our frame-of-reference we might draw the conciusion thai Ihe
criticality of critical incidents seems to be related to the situation al dimension. We
Critica/ity of Critica//ncidents in Customer Re/ationships 105
da, however, not know to wh at extent internal and external situational variables drive
this outcome.
If we move to the retail banking market in Sweden the situation is different compared
to the domestic airline market. For many years there have been same big competing
banks and a number small local or regional banks. The Swedish financial market
was deregulated same years aga and now there are a number of foreign banks
competing in the market. The consumer market is characterized by keen competi-
tion. The results from the study in table !wo above show that the relationship tend to
break more frequently (20%) than is the case in the Swedish domestic airline market.
We also see that the relationship is perceived to be weaker in many cases (35%).
Ifwe look at the Swedish financial market, we see that a number of companies oper-
ate in the consumer market and aggressive competition also come from non-
traditional actors such as insurance companies and retail-companies e.g. IKEA and
ICA. Märtenssan (1985) studied banks in Sweden and found that 10% (4-23% de-
pending on which bank) of the customers have chosen their present bank due to
dissatisfaction with the former bank. Anderssan (1985) in his CIT-study found that
almost 25% had switched bank as a result of a critical incident. StrandviklLiljander
(1994) on the other hand focused on relationship strength in retail banking and
found that customers may have experienced negative critical incidents but still have
a very positive overall evaluation of the bank and stay in the relationship.
If we try to answer our main question: How critical are critical incidents we may con-
clude that in the financial market summarize the categories "Broken" and "weaker"
we find the more than half (55-59%) of the critical incidents in strang competition
market situations result in a weaker or broken customer relationship. On the other
hand a number of relationships remain unchanged despite a critical incident. Based
on data collected in traditional CI studies it is not possible to understand the reasons
or the effects on the customer relationship. Beside effects on customer actions (pur-
chases) there may be emotional or/and cognitive effects that are not detected. The
reasons for an effect may lie in situational factars or in the accumulated experience
within the relationship or in expectations for the future. A new approach to collect
data is therefore needed.
5 Future research
In thls paper we have discussed some aspects of the criticality of critical incidents.
We have argued for the fruitfulness and necessity of widening the scope of the
analysis by including contextual factors. We have developed a tentative framework
for analyzing critical incidents where the history and future of the relationship is in
focus together with the situation al context. When trying to reanalyze data from previ-
ous CI-studies we lack data on the context and therefore we have not, so far, been
able to test or validate our suggested model in a sufficient way. In future research we
want to collect data on the contextual dimensions to be able to include this part of
the analysis and thus take us one step further in our search for an answer on the
question: How critical are critical incidents?
We need further empirical research and data from different service industries and
customer groups before launehing these as generic contextual dimensions. Further
more we need to develop methods for how to identify the variables or aspects to be
measured within these main dimensions and how to best collect and analyze the
data.
Consequently we suggest that if critical incidents are studied by interviews the inter-
view should cover the initial state of the relationship before the incident, the trigger
starting the critical incident, a description of the micro-process representing the inci-
dent and the outcome of incident. It should be noted that outcomes can be depicted
as emotions, cognition and actions. There are different starting points and ap-
proaches concerning critical incidents in customer relationships. One field of re-
search where critical incidents are embedded in a process defined according to the
outcome is studies concerning relationship termination. In this case the focus is on
relationships where the effect can be seen in customer actions. A related approach
is to focus on customers' switching behavior where both the dissolution process and
the initiation of a new relationship are in focus. Another approach is employed in
relationship strength studies where customers are asked to retrospectively reflect on
their current customer relationships and critical incidents within these. The aim may
then be to understand the nature of relationship strength and factors that make the
customer sustain the relationship. A focus on the criticality of critical incidents would,
however, represent a distinet type of studies that could employ similar methods but
would have a different aim. One purpose with such studies is to detect faetors that
cause changes in customers' evaluation of their relationships. Another purpose is to
identify customer segments that react differently to similar incidents. The level of
customer involvement in the service er industry combined with the level of customer
commitment to the relationship may provide an important internal cantext that ex-
plains customer reactions to critical incidents. The role of bonds can also be ex-
plored in these this type of studies. Interesting issues are also the effect of changes
in the external context, for example, campetitors' actions or substantial changes in
the conditions in the market place. As we see it, the proposed framework gives pos-
sibilities to specify a number of new research avenues that would contribute to the
current knowledge about customer relationships.
Another area which we would like to see more research in is the internal services-
cape which has to da with the individual customer and the cognitive, emotional and
behavior reactions "inside", the internal landscape and how the external service
arena is described and perceived. Maybe the criticality to a great extent is a function
of a pers ans own norms, standards and emotional status. This would imply that a
critical incident is not only related to a particular relationship but rather an element in
the customer's all experiences and relationships. The context may be defined by the
customer, rather than by the service provider or the researcher. The relevant setting
may be the customer's history and future rather than the history and future of a spe-
cific relationship. The situation al context may weil be the customer's combination of
a number of interdependent service offerings and service provider alternatives rather
than the service in focus. Here we need definitely more research.
References
Andersson, G. (1985): Kritiska händelser i banken, Paper presented at Nordic

Conference on Service Management, Lund University, Lund. (In Swedish).
8ejou, D.lEdvardsson, 8. (1994): An Empirical Investigation of Service Break-

downs in Swedish and U.S. Airlines: A Critical Ineident Approach, AMA Confer-
ence, Summer 1994.
8itner, M.J. (1990): Evaluating Service Encounters: The Effects of Physical Sur-
roundings and Employee Responses, Journal of Marketing, Vol. 54, pp. 69-82.
8itner, M.J.l8ooms, 8.H./Tetreault, M.S. (1990): The Service Encounter: Diag-

nosing Favorable and Unfavorable Ineidents, in: Journal of Marketing, Vol. 54,
January 1992, pp. 71-84.
8itner, M.J.lNyquist, J.D./8ooms, 8.H. (1985): The Critical Incident as a Tech-

nique for Analyzing the Service Encounter, in: Boch, T.M.lUpah, G.D.I Zei-
thaml, VA (eds.): Service Marketing in aChanging Environment, Chicago,
pp. 48-51.
Edvardsson, 8. (1988): Service Quality in Customer Relationships: A Study of Criti-

cal Incidents in Mechanical Engineering Companies, in: The Service Industries
Journal, Vol. 8, No. 4, pp. 427-445.
Edvardsson, 8. (1992): Service Breakdowns, A Study of Critical Incidents in an Air-

line, in: International Journal of Service Industry Management, Vol. 3, No. 4,
pp. 17-29.
Edvardsson, 8. (1996): Kvalitet och tjänsteutveckling, Studentlitteratur, Lund.
Holmlund, M. (1996): A Theoretical Framework for Perceived Quality in Business

Relationships, Swedish School of Economics and Business Administration, Re-
search Reports from the Swedish School of Economics and Business Admini-
stration 1996, Helsinki.
Holmlund, M. (1997): Perceived Quality in Business Relationships, Publications of

Swedish School of Economics and Business Administration No 66, Helsinki.
Johnston, 8. (1995): The Determinants of Service Quality: Satisfiers and Dissatisfi-

ers, in: International Journal of Service Industry Management, Vol. 6, No 5,
pp. 53-71.
Keaveney, S.M. (1995): Customer Switching Behavior in Services Industries: An

Exploratory Study, in: Journal of Marketing, Vol. 59, No. 2, pp. 71-82.
Liljander, V.lStrandvik, T. (1995): The Nature of Customer Relationships in Serv-

ices, in: Swartz, TA/Bowen, D.E.lBrown,S.w. (eds.): Advances in Services
Marketing and Management, Vol. 4. London, pp. 141-167.
Criticality of Critica//ncidents in Customer Re/ationships 109
Liljander, V./Strandvik, T. (1997): Emotions in service satisfaction, in: International

Journal of Service Industry Management, Vol. 8, No 2, pp. 148-160.
Morgan, R.M./Hunt, S.O. (1994): The Commitment-Trust Theory of Relationship

Marketing, in: Journal of Marketing, Vol. 58, No. 3, pp. 20-38.
Märtensson, R. (1985): Consumer Choice Criteria in Retail Bank Selection, in: The
International Journal of Bank Marketing, Vol. 3, No. 2, pp. 64-74.
Oliver, R.L. (1997): Satisfaction - A Behavioral Perspective on the Consumer, New

York.
Olsen, M.J.S. (1992): Kvalitet i banktjänster: Privatkunders upplevda problem med

banktjänster - en studie med hjälp av kritisk-händelse-metoden. Doktorsavhan-
dling vid Stockholms Universitet och Centrum för ijänsteforskning (CTF), Karl-
stad.
Roos, I. (1996): Customer switching behavior in retailing, Swedish School of Eco-

nomics and Business Administration, Working Papers 327, Helsinki.
Roos, I.lStrandvik, T. (1996): Diagnosing the Termination of Customer Relation-

ships, Working Papers No. 335, Swedish School of Economics and Business
Administration, Helsinki.
Roos, I. (1998): Customer switching behavior in retailing, Research reports 41,

Swedish School of Economics and Business Administration, Helsinki.
Stauss, B. (1993): Using the Critical Incident Technique in Measuring and Manag-
ing Service Quality, in Scheuing, E.lChristopher, W. (eds.): The Quality Hand-
book With Contributions from 57 International Experts, pp. 408-427.
Stauss, B./Weinlich, B. (1995): Process-Oriented Measurement of Service Quality

By Applying the Sequential Incident Method. Proceedings from Workshop on
Quality Management in Service V, May 11-12, 1995, Tilburg, The Netherlands.
Stauss, B./Neuhaus, P. (1997): The Qualitative Satisfaction Model, in: International

Journal of Service Industry Management, Vol. 8, No. 3, pp. 236-249.
Stewart, K. (1996): Relationship breakdown - an exploration in retail banking, in:

Sheth, J.N.lSöllner, A. (eds.): 1996 International Conference on Relationship
Marketing, Development, Management and Governance of Relationships,
March 29-31,1996, Humboltdt-Universität Berlin.
Strandvik T. (1994): Tolerance Zones in Perceived Service Quality, Research Re-

port No 58, Swedish School of Economics and Business Administration, Hel-
sinki.
Strandvik, T.lLiljander, V. (1994): Relationship Strength in Bank Services, in:

Sheth, J.N.lParvatiyar, A. (eds.): Research Conference Proceedings Relation-
ship Marketing: Theory, Methods and Applications.
Strandvik, T./Storbacka, K. (1996): Managing Customer Relationships, in: Ed-

vardsson, B.lBrown, S.w.lJohnston, R.lScheuing, E.E. (eds.): Advancing
Service Quality: Agiobai perspective, International Service Quality Association,
Warwick, pp. 67-76.
Strandvik, T./Tömroos, J.-A. (1997): Discovering Relationscapes - Extending the

Horizon of Relationship Strategies in Marketing, American Marketing Associati-
on, Special Conference, New and Evolving Paradigms: The Emerging Future of
Marketing, Dublin, Ireland, 1997. pp. 381-394.
Strandvik, T./Friman, M. (1998): Negative Criticallncident Mapping, Proceedings of

QUIS6 Conference, June 29-July 2, 1998, Norwalk, Connecticut, USA.
Wrange, K. (1996): Process oriented approach to relationship termination behavior,

unpublished working paper, Swedish School of Economics and Business Ad-
ministration, Center for Relationship Marketing and Service Management, Hel-
sinki.
Wrange, K. (1997): Relationship Termination Intentions - A Study of Hotel Custom-

ers, unpublished lic. dissertation, Swedish School of Economics and Business
Administration, Helsinki.
Critical Incidents in Dutch Consumer Press:
Why Dissatisfied Customers Complain with Third Parties
Jan A. Schulp
Contents
Abstract 113
Introduction 113
1.1 Dissatisfaction and Complaining Behavior 115
1.2 Complaining and Consumerism 117
1.3 Perceived Justice: Considering the Feelings of Customers 119
1.4 The Role of Ethics in Dealing with Customer Dissatisfaction 121
2 Methodology 121
2.1 Description of the Data 121
2.2 Stekeligheden are Criticallncidents 124
2.3 Stekeligheden are a Secondary Source: Strengths and
Limitations 125
2.4 Reliability of the Classification 126
2.5 Outline of the Research Process 127
3 Results 128
3.1 Classification of the Critical Incidents by Industry 128
3.2 Classification by Primary Causes of the Incidents 130
3.3 Understanding the Process Leading to a Complaint 133
112 Jan A. Schulp
3.4 Classification of the Complaint Triggers 134

3.4.1 Attributes of the Promotions 135
3.4.2 Employee and Management Behavior 138
3.4.3 Customer Behavior 140
3.5 The Costs and the Benefits 144
4 Discussion 146
4.1 Disconfirmation of Expectations 146

4.2 Unresponsiveness 147
4.3 Perceived Justice: Stability and Controllability 148
4.4 Perceived Justice: Clear Dishonesty and Wounded Feelings 148
4.5 The Answers to the Research Question 150
5 Managerial Implications 150
References
Appendix
Criticallncidents in Dutch Consumer Press 113
Abstract
Customer complaints published in the main Dutch consumers' monthly were studied
using the Critical Incident Technique. Out of 750 incidents two sampies on brand
promotions and retail promotions of 52 and 137 items respectively were studied in
more detail.
Dissatisfiers in the incidents have been classified according to attributes of the pro-
motions and the employee behavior. Customer behavior is described and classified
in so far as it leads to complaining with the Consumentenbond. Dissatisfaction is
mainly about the intangible aspects of the transactions. An aggravating faetor in
promotions is the enhanced expectation which is disconfirmed. Dissatisfaction is
caused both by attributes of the promotions and the interaction between employees
and customers due to negative customers affects and negative perception of service
fairness. Differences between the outcome in retail promotions and brand promotions
clearly reflect the difference in personal contaet intensity. The incidence of com-
plaining with the Consumentenbond without previous complaining with the seNice
supplier is highlighted as a managerial problem. In designing a promotion, ethical
viewpoints should be taken into account.
1 Introduction
Complaint management and service recovery have received extensive attention in
service management research (HartiHesketUSasser 1990; Kelley/Hoffman/Oavis
1993; Bitner/BoomslTetreault 1990; Stauss/Seidel 1996; Oe Ruyter 1994; Olsen
1992; Edvardsson/Luukkonen 1996). The subject is important for several reasons. A
big majority of dissatisfied customers utter no complaint whatever. If an alternative is
available, they switch silently to another service provider. If no alternative is avail-
able, they stay on in spite of their dissatisfaction. More active dissatisfied customers
complain with the service provider. In studies on customer complaint behavior and
complaint management, most of the attention is focused towards these direct com-
plaints to the service provider (Bitner/BoomslTetreault 1990; HartiHesketUSasser
1990; Olsen 1992; Kelley/HoffmaniDavis 1993; De Ruyter 1994; Stauss/Seidel 1996;
114 Jan A. Schulp
Edvardsson/Luukkonen 1996; Oliver 1996; Edvardsson 1997; TaxlBrown/Chandra-

shekaran 1998). Attention to complaining with third parties in general receives less
attention (Richins 1983; Bearden/Oliver 1985; Singh 1990; Oliver 1996). Yet, com-
plaining with third parties can have serious consequences for firms that have made
their customers dissatisfied. It has been reported that negative experiences are told
to other people by 85% of dissatisfied customers (Richins 1983, as quoted by
Stauss/Seidel1996, p. 45). Stauss/Seidel (1996) also stress that this communication
is all the more credible because the complainer does not aim at a direct profit for his
own. HeskettiSasserlSchlesinger (1997, pp. 105-107) have calculated the negative
impact of service terrorists upon the profitability of service companies.
Complaining with third parties can have really serious consequences, if the com-
plainant goes to court, and even more if mass media are involved. There is one well-
documented case of a manufacturer of health mattresses who went broke after an
attack of the TV program 'Ook dat nog!' (Gemeenschappelijke Pers Dienst 1992).
Word of Web is a fairly recent phenomenon with a high damage potential. It is an
important medium, but difficult to control (HeskettiSasserlSchlesinger 1997, p. 109).
Consumers' organizations are important third parties to which dissatisfied consum-

ers refer. Mostly the consumers get individual help, but certain conflicts between
customers and service providers are published in the organizations' journals. Publi-
cations of this kind are an underestimated and underexploited source for the under-
standing of the complaining behavior with third parties and on the causes of con-
sumer dissatisfaction. The author has performed two studies of 750 incidents pub-
lished in the column Stekeligheden, in the Consumentengids (The Consumers
Guide), the monthly of the Nederlandse Consumentenbond , the Dutch Consumers'
Association (Schulp 1997, 1998). Stekeligheden is a Dutch word best translated as
prickles. The central research quest ion for the present paper is: Wh ich factors in
the incidents reported have made the customer so dissatisfied that she has
sought publicity with her complaints?
In order to provide a theoretical framework for the consumers' behavior when com-
plaining with third parties, a literature review will be given on complaining behavior
and on perceived justice and ethics in consumerlservice provider interactions.
1.1 Dissatisfaction and complaining behavior

The earliest and still authoritative customer satisfaction models are based upon the
difference between expected and experienced service quality: the confirma-
tion/disconfirmation paradigm. Examples are the Gränroos/Gummesson model
(Gränroos 1992, pp. 36-43) and the gap-model of Parasuraman/Berry/Zeithaml
(1985). These models still have a great value in analyzing and explaining customer
satisfaction and dissatisfaction.
Once the customer has perceived service quality as unsatisfactory, he or she can
take several ways of action. In the words of Oliver (1996, p. 359):
"Answer: To do nothing, absolutely nothing.
Queslion: Whal is Ihe mosl frequenl cuslomer response 10 bolh salisfying and dis-
salisfying consumplion outcomes?
So complaining, lei alone complaining wilh Ihird parties, is a relalively rare phe-
nomenon; yel it is very importanI."
Oliver (1996, ch. 13) describes consumer dissatisfaclion as hard 10 remedy. If

somelhing goes wrong during Ihe initial service contacl, much of Ihe ulilily 10 the
cuslomer is losl. In Ihis perspeclive, il is surprising thai nonelheless a service recov-
ery Ihal is salisfactory to Ihe cuslomer will creale a posilive attilude towards Ihe
service provider, even if the full ulilily of Ihe service is not reslored (Hart/HesketU
Sasser 1990; ClarklKaminski/Rink 1992 as quoled by Oliver 1996, pp. 365-369).
This is one reason why not many dissalisfied consumers will complain and even less
cuslomers will complain 10 Ihird parties.
Oliver has developed two models for the explanalion of complaining behavior: a be-
havioral model, analyzing personal and cultural traits making Ihe dissalisfied cus-
tomer more or less prone to complain and an economic model, analyzing the poten-
tial utilily of complaining 10 the dissatisfied cuslomer (figures 1 and 2 respeclively).
The contributors to the Stekeligheden have all vanquished their reluclance to com-
plain with a Ihird party: the Consumentenbond. All faclors of Oliver's behavioral
116 Jan A. Schulp
model are present: they have the ability and motivation. The members of the Con-
sumentenbond know their own association, they have access to its services and they
have the communication skills to write to the Consumentenbond. In Dutch culture,
complaining is certainly considered acceptable. Complaint proneness is even higher
amongst Consumentenbond members than in the Dutch population in general; the
desire for astronger position against suppliers is precisely what membership is for.
Dissatisfied customers can go to several formal institutions: many lines of trade have
their own arbitration boards; there is anational ombudsman, there are several very
popular TV programs in which malperforming service providers can be exposed (Ook
dat nog; Breekijzer), and last not least, there is the Consumentenbond. Although
many people are willing to confront a malperforming service provider, they perhaps
dare not to da so directly due to implicit or explicit, real or supposed threat and in-
timidation from the service provider.
• Knowledge of Skills
• Access to Channels Ability
• Communication Skills
• Cultural Norms
• Formal Institutions
• Complainig Identity
• Willingness to Confront
•Threat to Intim idation
Fig. 1: The deeision 10 eomplain: Ihe behavioral model

Souree: Oliver (1996, p. 363)
Oliver's economic complaint model (see figure 2) will be used to analyze the costs
and benefits of complaining, especially with the Consumentenbond. From the data it
will be inferred what role perceived complaining costs and benefits played in the be-
havior of the complainants.
• Direct Monetary Loss

• Indirect Losses
• Time Perceived Costs
• Effort
• Product Importance
• Money Back
• Replacement
• Extra Reimburement
Perceived Benefits
• Correct Problem
• Apology
• Catharsis
• Firm's Repuatation
• Threat to Business Probability of Success
• Perceived Efficacy
Fig. 2: The decision 10 complain: Ihe economic model

Source: Oliver (1996, p. 362)
In his review of complaining behavior, Oliver devotes a paragraph to the risks of

fraudulent complainers - a small risk (pp. 369-370). The assumplion Ihat complain-
ers are fraudulent also plays a role in so me cases in the present research.
1.2 Complaining and consumerism

In this section the literature review by Oliver (1996, p. 375) has been used exlen-
sively. The following section is a literal quotation:
"Warland/HerrmannlWiliits (1975) and Singh (1990) have performed research on

dissalisfied consumers taking external action. In Warland's study, Ihe action group
proved 10 be younger, more educaled and more upscale demographically, perhaps
reflecling Ihe consumerism movemenl al Ihe li me of the sludy. Singh faclor-analyzed
responses 10 dissalisfaclion and found Ihree faclors which he labeled voice (com-
plaining), private (word of moulh) and third party (public complaining 10 Ihird party
118 Jan A. Schulp
institutions). A cluster analysis of consumers' scores on these dimensions yielded

four groups of consumers, described as (1) passives, having low scores on all dis-
satisfaction dimensions; (2) voicers, scoring high on the complaining dimensions
and low on the others; (3) irates, scoring very high on ward of mouth; and (4) activ-
ists, scoring high on complaining and particularly high on third party actions.
A demographie and attitudinal analysis of these clusters showed that activists were
younger, were alienated and had positive atlitudes toward complaining and the likely
consequences. [... ] 80th voicers and irates were older, but differed on alienation.
Voicers were less alienated and had a positive evaluation of complaining and its
consequences. Irates were more alienated, saw greater gain in negative word of
mouth, and less gain in complaining when compared to voicers".
In general, literature on complaint management stresses the value of the voicer for
the business (De Ruyter 1994, Decker/Meissner 1997, pp. 3-4; Stauss/Seidel 1996,
eh. 3; Krips 1990). Voicers provide valuable suggestions for improvement of busi-
ness practices, and a timely recovery can maintain or even enhance their loyalty.
Activists on the other hand are a potential source of trouble, because they say noth-
ing and then recur to third parties. The companies that caused the dissatisfaction
only perceive it when it is too late: when they are approached by television, the Con-
sumentenbond etc. and when it becomes inevitable that they will be expased due to
their bad performance.
The present research focuses very much on activists or even service terrorists in the
sense of HesketUHartlSasser (1997, pp. 105-109); due to the nature of the source
material the passives and the irates are completely out of the picture. The research
will provide some insight in the process of the birth of an activist. This is fairly often:
the birth of an activist out of a voicer.
1.3 Perceived justice: considering the feelings of

customers
BlodgettlGranboislWalters (1993) have published a model on the connection be-
tween perceived justice and engaging in ward of mouth plus repurchasing intentions
(see figure 3).
stabilityl repatronage
controlability intentions
Fig. 3: The perceived justice model
Note: Only positive correlations are shown. Correlation coefficients have been omitted.
Source: Redrawn and simplified after BlodgettiGranboislWalters (1993)
Perceived justice is strongly influenced by the elements stability/controllability, prod-

uct importance, attitude towards complaining and likelihood of success. The cus-
tomer's perception of justice, of fairness plays an important role in the analysis of
complaining behavior. Seiders (1997) has published work on consumer-perceived
fairness, based upon in-depth individual and focus group interviews. She distin-
guishes four factars: account, campensation, price positioning and scenario. Her
initial hypothesis was that a low price positioning would diminish the customer ex-
120 Jan A. Schulp
pectations of fairness. Actually, she had to reject this hypothesis. Fairness is to the
customer non-negotiable, not even in exchange for a low price.
Cumby/Barnes (1997) presented a conceptual model in which they consider the

customer's feelings that are caused by the interaction with the service provider. They
give many examples of customers wounded on the affective level: customers feit let
down, hurt, treated dishonestly, humiliated etc. Cumby/Barnes founded part of their
research upon a paper by Price/ArnouldlTierney (1995) in which duration, affective
content and proxemics were postulated as important dimensions of the service en-
counter. Both positive and negative affective customer responses were important
factors for customer satisfaction. Price/ArnouldlTierny coined the term EAI en-
counters (Extended, Affective, Intimate) and their research was about relations of
this kind, so their methodology cannot be transmitted unaltered to the present study
but on the dimensions of affective content and proxemics, their study is important.
The affects of the customers are influenced by the employee's behavior. Bit-
ner/Booms/Mohr (1994) studied the reciprocal satisfaction/dissatisfaction between
customers and employees in critical service encounters in different service indus-
tries. They used the attribution theory in this study: that an unsuccessful service en-
counter is blamed on the customer by the employee It will be seen to what extent
this theory plays a role in the employee behavior in the present study.
Arecent study by TaxlBrown/Chandrashekaran (1998) established a relationship

between distributive, procedural and interactional justice in complaint procedures
and trust and commitment on the part of the customer. Feelings, positive and nega-
tive, played an important role in the perception of justice.
The importance of emotions in the customerlservice provider relationship is also

highlighted in the switching behavior study of Roos (1997). She made critical path
studies, based upon interviews with customers who had switched of supermarket
recently. By questioning them about the whole switching process, she could associ-
ate the feelings of the dissatisfied customers with the chain of incidents leading to
switching. The emotions played a very important role. Similar results have been
found by Keaveney (1995) who studied isolated critical incidents in connection with
switching behavior of customers with a broad spectrum of service providers. Refer-
Critieallneidents in Duteh Consumer Press 121
ences to feelings are also to be found in the CIT-studies of Olsen (1992), Ed-
vardsson/Luukkonen (1996) and Edvardsson (1997).
Although perceived justice seems the more important explanation for complaining
with the Consumentenbond, the role of the confirmation/disconfirmation paradigm in
perceived service quality cannot be ruled out. Basically, a customer expects to be
treated according to the rules of justice.
1.4 The role of ethics in dealing with customer dis-

satisfaction
In section 1.3 it was shown that fairness and the feelings of customers were very
important causes of customer satisfaction or dissatisfaction. If the feeling for what is
right is so important, ethics should be included when studying this subject. The im-
portance of an ethical analysis, for example according to a technique presented by
Khan/McCleary (1995), can be illustrated by the problem of the activists.
In section 1.2 it was stated that activists are a serious source of trouble because
they say nothing and then recur to third parties. They are a kind of time bombs that
are not even ticking but that explode without warning If however, promotions, or
services in general are designed with ethics inside, considering the ethical accept-
ability while designing the service, this kind of dissatisfaction can be prevented alto-
gether. Ethics should be built in, much like quality in TOM.
2 Methodology
2.1 Description of the data
A considerable number of complaints to a consumers' organization have been pub-
lished in a uniform format in the Consumentengids of the Consumentenbond in the
monthly column Stekeligheden. The Dutch name for this column will be used
throughout the paper. Every month 6-9 incidents are published. This amounts to
some 4.000 incidents in the history of the Consumentenbond. The Stekeligheden are
printed on the back page of the Consumentengids and are by far the most popular
122 Jan A. Schulp
column: for 97% of the readers it is the first item to read. The Stekeligheden are se-
lected by the responsible editor; often, but not exclusively, they are suggested by the
Consumentenbond's legal department from the approximately 1.000 complaints and
problems it receives every month. An incident qualifies for Stekeligheden if the
problem is caused by the service provider intentionally, or if the service provider is
negligent or unwilling to solve the problems. The problem should also be interesting
for the readers. All incidents are checked by the editor with the service providers
against whom complaints have been lodged. The editor doesn't use a standard for-
mat but because he has done this column for many years, a certain consistency has
grown. All incidents are checked for reliability: the complaint is sent to the company
concerned. They have at least five workdays after confirmation of receipt to give a
reaction. If they don't react, it is supposed that the complaint against the company is
true. If they do react and offer a form of explanation, recovery or apology, the Steke-
ligheid can be cancelled, or be published with a postscript. Approximately 8% of the
cases are published with a postscript. All Stekeligheden are checked by a legal ex-
pert of the Consumentenbond, read by fellow editors and finally approved by the
managing editor (personal communication Mr. Schleiffert, Stekeligheden editor,
Consumentenbond).
Typically, a Stekeligheid is between 80 and 120 words, comparable to the critical

incidents as given by Olsen (1992) on retail banking services, or by Ed-
vardsson/Luukkonen (1996) on public transport. Although the incidents reported are
summarized in this very short format, the information content is high. The editor suc-
ceeded in giving all relevant information in this small format. In each Stekeligheid is
stated
• the company causing the complaint (name, place),

• the article or the service concerned,
• gender, age group indication of the dissatisfied customer and
• adescription of the dissatisfiers.
Dependent upon the precise nature of the complaint, very often clues can be found
concerning
• amount of money involved,

• 1055 of time and eftor!,
• 1055 of utility for the customers or for members of her family,
• precise actions of the customer,

• employee reactions,
• involvement of management,
• feelings of the customers and
• redress or explanation from service provider.
For the present research, the Stekeligheden from January 1990 through February
1998 have been used. The total number of cases was 750. A more profound analy-
sis has been made of the cases on brand promotions and retail promotions. These
categories contained 52 and 137 items respectively.
For the purpose of the present research, all 750 Stekeligheden have been numbered
in chronological order.
A typical Stekeligheid is given in translation in figure 4 (Prices are given in Dutch

guilders, symbol: 'n.
#222 January 1992
"On the promotion wrapping of Edet Velvet toilet paper, action stamps can be
found. If you send two stamps to Edet's Roermond branch, you get a voucher
t
for a 4,- discount at Fleurop flower shops.
t
You also can obtain action stamps by sending a 2,95 check in an envelope
t
with a 0,80 stamp to Edet's Amstelveen branch. The two stamps you receive
t
from Amstelveen should be sent to Roermond, again for a 0,80 stamp. So
t t
you pay 4,55 for a 4,00 voucher.
The toilet paper producer keeps you in flowers at your own cost".
Fig. 4: A ·stekeligheid" in English translation

124 Jan A. Schulp
2.2 Stekeligheden are Critical Incidents

The incidents from Stekeligheden can be approached and analyzed by the critical
incident technique (CIT). CIT has been widely applied in service quality studies (Bit-
ner/BoomsfTetreault 1990; Edvardsson/Luukkonen 1996; Kelley/Hoffman/Davis
1994; Olsen 1992; Olsen/Bertilsson 1991; Stauss 1993).
The Stekeligheden do conform to the definition of critical incidents given by Olsen

(1993): "A critical incident can be considered as a moment in the service process, a
moment of truth that initially is perceived by the customer as not normal".
In figure 5, the Stekeligheden are compared with critical incidents in general, as de-
fined by Stauss (1993).
Ouestions to be asked in applying CIT What happens in

(according to Stauss 1993) Stekeligheden?
Exactly wh at happened (action)? same
Exactly who did what (actor)? same
Who or what was the subject of the incident same

(object)?
Where did the incident take place (place)? same
When did the incident take place (time)? same
How do you evaluate the incident (evaluation)? without exception: negative
Wh at exactly did you make feel the situation practically always: I complained,
was dissatisfying/satisfying (cause of evalua- but I got no adequate reaction
tion)?
How did you, or how do you intend to invoking help from the
respond to the incident (consequence)? Consumentenbond
Fig. 5: Critical Incidents compared with Stekeligheden
It can be conciuded, then, that Stekeligheden can be considered as critical incidents

and can be interpreted accordingly.
Limitations of the elT have been discussed in depth by Olsen (1992, 1993) and by
Edvardsson/Luukkonen (1996). The Stekeligheden are subject to the same limita-
tions. Also, the Stekeligheden are secondary sources and this requires a closer ex-
amination.
2.3 Stekeligheden are a secondary source:

strengths and limitations
It is clear that Stekeligheden are a secondary source. They have not been collected
by the present author, neither have they been intended as sources for research; they
have been selected for instruction and amusement of the readers of the Consumen-
tengids but the author agrees with Edvardsson/Luukkonen (1996) that"a source can
be defined in several ways. More exactly, it can be expressed as material put down
in writing that becomes a source as soon as somebody starts using it as a source"
(after Holme/Solvang 1983). Its value depends upon relevance, completeness and
reliability. It his been shown that "Stekeligheden" qualify very weil in this respect.
There is a tradition - or a prejudice - in social sciences that secondary sources are
second-rate; the most authoritative research is based upon primary data, mainly
questionnaires. Of late, criticism upon surveys is increasing, on elements like non-
~esponse and question construction, which cause a serious concern ab out the rep-
resentativeness of the sampies and the validity of the answers (Westendorp 1997). If
in the present research the sampie may have a bias towards high er educated, more
assertive consumers, then surveys have a bias towards people with time to spare
and too liltle assertiveness to say no to the poll-taker. Methods should be used
where possible in combination with each other, because a combination of method-
ologies gives arieher and more eomplete pieture.
Let us compare the normal elT with the use of secondary material like "Stekelighe-
den". Usually, in the CIT, interviews are held by the researcher. All researchers
quoted above have applied this approach, with the exception of Ed-
vardsson/Luukkonen (1996). They used wrilten complaints from the archives of a
public transport company. They give arguments in favor of working with data that
have not been collected by themselves: artifacts caused by the interview structure or
126 Jan A. Schulp
by the interviewer can be ruled out. Especially, all information ccnsidered important
by the complaining customer is present in the source material, in his/her own words.
Edvardsson/Karlsson/Lindström (1995) have stressed that customer needs and de-
sires when needed for working on service quality, are best expressed in the custom-
ers' own words without intervention or pre-structuring by an interviewer.
The present research go es even further than Edvardsson/Luukkonen (1996). The

source material has not been found with the company against which the complaints
were directed, but with a third instance: the Consumentenbond. As in the research of
Edvardsson/Luukkonen, the critical incidents were not told at the request of the re-
searcher; writing down the incident, on the contrary, was a spontaneous action of the
dissatisfied customer. For Edvardsson/Luukkonen, the use of data not collected by
themselves was a fortunate accessory aspect of their research. In the present re-
search, working with these secondary data is essential, because the main research
question is: "What is so special in the incidents reported that they have caused this
strong form of discontent?". For the discontent indeed must have been very strong,
because the customer has taken the trouble to write a letter to the Consumenten-
bond, even to spend money on a stamp for the letter.
In figure 6 the strengths and weaknesses of the different apprcaches to consumer

dissatisfaction have been summarized.
2.4 Reliability of the classification

The author made the classifications on his own. At first, they could not be screened
against inter-subjectivity, due to a lack of man power. The classifications by industry
and primary cause of dissatisfaction have been screened against intra-subjectivity
effects by repeating the classification after a certain time. After one year, only 1 out
of the 46 original cases in brand promotions had to be re-classified and out of the
original 141 cases on retailing promotions, one had to be added and 9 to be re-
moved. Between February 1997 and February 1998, 74 new cases could be added
to the collection. They all have been classified by industry and primary cause of
complaint. After one month this classification was repeated and no change at all was
necessary. Afterwards, acheck against inter-subjectivity could be performed for
these elassification. Three students were provided with the original criteria and did
the elassification. For the eomparison of the four elassifieations, Cohen's kappa was
caleulated as 0.73, 0.74 and 0.67 respeetively for the three pairs of students. These
data support the reliability of the elassifications by industry and primary cause of dis-
satisfaetion. Other elassifications eould not be sereened against inter-subjectivity
whieh is a limitation of this study. Classifieation methods and eriteria will be given
with the results.
Questionnaire Solicited CIT Stekeligheden
Customer's state- Researcher Researeher Customer

ment triggered by
Customer's own No Yes, apart from Yes, apart from

words interviewer's editor's influence
influenee
Importanee to Variable High Very high

eustomer
Completeness in Low High Very high

relevant detail
Interviewer's Very high Moderate Absent

artifaets
Verifieation of inei- Not relevant Poor Good

dents
Influenee of editor Not relevant Not relevant Moderate
Possibilities for Very good Moderate Moderate

quantitative
analysis
Fig.6: Comparison of the strengths and weaknesses of questionnaire method, nonnal CIT and
the use of secondary sources in studying customer dissatisfaction
2.5 Outline of the Research Process

After reeognition of the potential importanee of the Stekeligheden as a souree for
understanding eonsumer eomplaining behavior, the eritieal ineidents were read sev-
eral times. In order to identify the major problem fields, a elassifieation of the inei-
128 Jan A. Schulp
dents was attempted by industry and then by primary cause of the dissatisfaction. It
was concluded that promotions constituted the biggest primary source of discontent.
Initially, one minor category: promotions of manufacturing industries, mostly in fast
moving consumer goods (henceforth to be designated as brand promotions) was
analyzed with respect to promotion attributes only. This study was presented at the
EIASM Workshop on Service Quality in 1997 (Schulp 1997). Subsequent discus-
sions with colleagues and further analysis, both of brand promotions and retail pro-
motions led to a broader scope of the research in which also the customer and the
employee process in the critical incidents received attention. This study was pre-
sented at the EIASM Workshop on Service Quality in 1998 (Schulp 1998). In June
1998 the check of some classifications against intersubjectivity effects could be
performed. The present paper aims at an integration of these studies.
3 Results
In the analysis of Stekeligheden, frequent reference is made to the appendix in
which shortened description of incidents are given. The incidents are quoted by their
number (e.g. #074).
3.1 Classification of the critical incidents by

industry
The 750 critical incidents available have been classified first by industries. Separate
categories have been made whenever the category contained 1% or more of all
data. This classification is shown in figure 7.
Retail is by far the biggest category. Shops, mailorder companies and gas stations
have been placed in this category.
A considerable number of incidents is about manufacturing industry which uses

the retail as its distribution channel. Whenever a brand or a producer is mentioned
as the primary source of the problem the incident has been placed in this category.
In the category publishers, incidents have been placed in which no retailers were
involved. Invariably, subscriptions for journals or serial books were the source of
problems. In the words of Lovelock (1996, pp. 39-41) a membership type of relation-
ship between publisher and customer was concerned.
In the category leisure, tourism, travel, recreation and culture have been put to-
gether.
In the category hospitality, incidents with hotels and restaurants have been placed.
Retail banking proved a major source of consumer problems, justifying aseparate

category. This is in line with findings of Olsen (1992).
In the category public transport not only rail and bus and taxi are included, but also
airlines, unless the flight was part of a package tour. In that case, the incident has
been included in leisure.
Car dealers are in aseparate category because the relation with the customer is
more complex than with ordinary retailers. There should be a long-standing relation-
ship for maintenance and repair, in which a great amount of trust and money on the
part of the customer is involved.
All incidents with insurance companies have been placed in the category insurance.
The category telecommunications and postal services are mainly occupied by the
former state monopolist KPN. The monopolies of KPN have been abolished only re-
cently. Competitors simply did not yet have time enough to cause problems. Inci-
dents on cable television providers are also in this category.
Others includes a wide variety of commercial services.
The dominance of retailing and manufacturing industry is not amazing, considering

the great number of contacts between consumers and retailers, in which numerous
material items change owner. In comparison, all other industries involve less inten-
sive and/or less frequent contact; so the likelihood for problems to arise is less.
130 Jan A. Schulp
Category (n= 676) Number(n) Percentage of n
Retail 370 49.2
Manufacturing Industry 89 11.9
Publishers 49 6.5
Leisure 37 4.9
Hospitality 33 44
Retail banking 32 4.3
Public transport 23 3.1
Car dealers 14 1.9
Telecommunications 17 2.3
Insurance 13 1.7
Postal services 8 1.1
Others 65 8.7
Total 750 100
Fig. 7: Stekeligheden classified by industry
3.2 Classification by Primary Causes of the

Incidents
The classification according to industry has been supplemented by a broad classifi-
cation of the primary causes of the incidents. When studying the incidents, initially
the categories Promotion, Procedures, Product, Price and Personnel emerged.
Gradually, the number of incidents specifically on repair and on guarantees grew so
large that separate classes were made.
The classification has a kind of hierarchy. If the customer's dissatisfaction is caused

by a promotion, there are always aggravating factors: the procedures, Le. a stupid
promotion design (#351), a poor product quality (#074), problems with the special
promotion price (#076) or unacceptable behavior of the personnel (#442). The faulty
promotion, however, is considered as the primary cause of dissatisfaction.
Promotions are all activities involving a special offer in order to induce the customer
to buy or to try the product or service. They show a great variety (section 3.4.1).
Procedures are elements of the service design, either planned or unplanned. When
for instance a customer is told: "Weil, this is the way we do it here, so stop grousing"
or "Sorry, I cannot help it but our computer (or: our administration) cannot cape with
it", incidents belong in this category. Invariably, procedures are intangible parts of
the service delivery process; they are about how the service is delivered (Grönroos
1992, pp. 37-39). Products are in the first place the tangible part of the service de-
livery. Also, if the problem is about ''what intangible service is delivered?" rather
than "how is the service delivered?", the incident is classified under product. An ex-
ample: incident #605 (not in appendix) about funeral insurance: the package cov-
ered by the insurance is changed; this incident has been classified in the product
category.
Price has only been classified as such if no connections with promotion or proce-
dure could be established.
The category personnel involves only behavior of personnel that has nothing to do
with the service design itself. In most cases it is misbehavior that probably would not
be approved by the employer. A striking example is the behavior of an employee in a
clothing discount shop, who discovered a very good T-shirt in an otherwise low-
quality container. Immediately she snatched away the T-shirt before the flabber-
gasted customer, saying: "tough luck for you, but this is mine" (#433, not in the ap-
pendix).
Whenever problems occurred with a product shortly after purchase, and replace-
ment, money-back or repair caused difficulties, the incident has been classified with
guarantees, irrespective wh ether a formal guarantee had been given or not. The
new Dutch Civil Code since 1994 anyway has considerably extended the responsi-
bility of the supplier of goods or services.
Repair includes obvious problems with a product long after purchase. The most fre-
quent kind of problem is wrong advice on the possibility to have the object repaired.
The general pattern is as folIows: advice on repair is asked; the product is declared
132 Jan A. Schulp
either beyond repair, or a very high price is stated. Subsequently, another firm or a
good DlY-er repairs the item for a trifling amount.
The classification by primary cause is in figure 8. Promotion is the biggest class.

Hence the present paper will be devoted to an analysis of brand and retail promo-
tions. It should be kept in mind however, that the great number of promotion prob-
lems is also connected with the industries with the greatest number of incidents.
Some other industries have other causes of incidents. For each industry, the main
cause of problems has been printed in bold type. For many industries the procedure
(service design) is the most important cause of serious problems.
Industry n Pro- Proce- Product Price Per- Guaran- Repair

motion dures sonnel tees
Retailing 370 137 78 56 27 12 25 35
Manufacturing 89 52 2 23 3 0 2 7
Publishers 49 20 17 6 6 0 0 0
Leisure 37 7 21 6 3 0 0 0
Hospitality 33 3 17 9 2 1 1 0
Retail banking 32 7 19 2 3 1 0 0
Public transport 23 5 13 4 1 0 0 0
Car dealers 14 7 1 0 1 1 3 1
Telecommuni- 17 0 7 1 0 0 1 8
cations
Insurance 13 4 4 5 0 0 0 0
Postal selVices 8 0 7 0 1 0 0 0
Others 65 15 37 6 6 0 1 0
Total 750 257 223 118 53 15 33 51
Fig. 8: Stekeligheden classified by the primary cause of the problem
Discontent about the tangible aspects (product) is only a minor cause of complaints
with the Consumentenbond. The intangible aspects are far more important, and
promotions, in total 257 out of 750 are absolutely the main cause of irritation. Even if
in most critical incidents analyzed material products are involved, the core of the in-
eidents is not about the tangible product but about the process. Therefore, a service
management approach of these incidents is justified.
3.3 Understanding the process leading to a

complaint
Figure 9 visualizes the process leading to a complaint with the Consumentenbond
about a brand or retail promotion.
Dissatisfaction,
Promotion with
Leading to
a Flaw in its
Escalation Complaint with
Design
Consumentenbond
Customer
Fig. 9: The process leading to a complaint about a promotion
A manufacturer or a retailer create a promotion. Often, but not always, the promotion
contains certain dissatisfiers. These dissatisfiers are attributes of the promotion. In
figure 10, these are called dissatisfiers in promotion design. The customer per-
ceives the promotion. Possibly she discovers the dissatisfiers at once or after an
analysis, without contact with the service provider. If she does not yet see the dis-
satisfier, she gets in touch with the service provider, either by engaging in a discus-
sion or by accepting the offer. At this point, additional dissatisfiers can turn up that
are not caused by the personal contact but that arise back office, for example from
an out of stock situation or from mistakes in advertising. These are dissatisfiers in
promotion realization. Dissatisfaction can also be aroused, or, more often, made
worse by actions of the contact personnel. Here two layers can be distinguished
134 Jan A. Schulp
• A behavior that can be described in objective terms, e.g. a refusal to fulfill

what is promised in the advertisement or turning the meaning of the promo-
tion in a very particular way. It is the kind of behavior for success;
• A behavior that could less easily described in objective terms. This kind of
behavior has affected the feelings of the dissatisfied customers and may
have played an important role in the negative evaluation of the incident. Yet
it is nearly impossible to take legal action against this type of behavior.
All negative attributes of the promotions and negative behaviors of the contact per-
sonne I escalate into a dissatisfaction leading to a letter to the Consumentenbond.
3.4 Classification of the complaint triggers

The incidents have been classified first according to three criteria:
• Attributes of the promotion. This classification is based upon dissatisfiers in

promotion design and in promotion realization, both back office and front of-
fice.
• Employee and management behavior. Classification of the employee be-
havior was easy at the level of objective behavior. In contrast, it was far from
easy to classify the employee and management behavior as affecting the
feelings of the customer. The classification was performed by describing all
incidents and characterizing them in very colloquial Dutch, even in slang.
Only after describing more than a hund red incidents in this way, the classifi-
cation emerged.
• Customer behavior. After analyzing much of the customer behavior in all in-
cidents, a pattern of eight categories emerged that proved remarkably con-
sistent and needed no further additions as more incidents had to be classi-
fied. It has been drawn as a cladogram in figure 12.
3.4.1 Attributes of the promotions

In figure 10 the dissatisfiers have been listed with their importance in the incidents
studied. A short explanation on each dissatisfier is given. The numbers after the ex-
planation refer to the incidents given in the appendix.
• Money back actions involve arefund of (part of) the purchase price, mostly
after sending in purchase tokens. It is a promotion technique more often ap-
plied by brand manufacturers than by retailers (example: #033).
• Nature of premium. Dissatisfaction can be caused by the nature of the
premium, when it does not meet the expectations or when specific require-
ments e.g. size are not considered (examples: #002, 074).
• Conditions can cause trouble when they are not clear or overtly unreason-
able, for example when using the premium causes an unreasonable amount
of extra cast (examples #374,667).
• Expiry date of action can cause trouble in several ways. Announcement
may be unclear; items announcing the promotion may be present in shops
after expiry date, or refund/premium quotum may have been reached before
expiry date (examples #033).
• Discount is no discount. This problem can occur in money back actions or
with premiums when the cast for postage and handling to the customer ex-
ceeds the value of the money back or the premium. Also cases of announc-
ing samething as a sale item when actually the price had not been changed
is inciuded in this category (examples: #076, 093, 438).
• Third party involved. Difficulties can arise when premiums of firm Aare
supplied to the customer by firm B. Probably the arrangements between A
and Bare unciear in these cases. It cannot be ruled out that A and Bare
conspiring against the consumer (examples: #374, 399).
• Different prices. Sometimes in mailings to different target groups different
prices for the same product or service are quoted. If this is perceived by po-
tential customers, it can cause much irritation (example: #332).
136 Jan A. Schulp
• Loss leader. An expensive article is offered with an enormous discount.

Clearly no profit is made on it, or more likely, a loss. If however a customer
comes to the shop "the article is out of stock", or "we sold the last item just
an hour ago". "But we have many other [... ] camcorders, TV sets, videos that
are also good, or even better. .." (example: #556).
• High/low pricing. Unrealistic original prices are advertised in order to en-
hance the perceived discount. Kaufmann et al. (1994) have devoted a paper
to this practice (example: #190).
• Intrusion into privacy. Incidents with customer cards, e.g. discount only
available, if the customer is willing to give information about herself, family,
consumption pattern etc. (example: #624).
• Retailers or personnel not informed. In brand promotions sometimes the
retailer doesn't know about the promotion; in retailing promotions sometimes
the personnel is not informed (example: #351).
• Faulty mailing: aimed at the wrong target group (examples: # 420,730).
• Faulty advertisement. Includes mistakes and incompleteness in (free local)
newspaper advertising and in advertising pamphlets, posters, bills, stickers
etc. inside or near the shop (examples: #368,490).
• Out of stock. When advertised articles cannot be supplied. In some cases
the out of stock situation can be intentional, when used in support of a loss
leader (examples: #109, 281).
It can be concluded that brand promotions and retailing promotions each have their
own pattern of dissatisfiers. In brand promotions, conditions and expiry dates are the
main problem areas, often connected with money-back actions and premiums. In
retailing promotions, conditions are important as weil. It looks as if service providers
want to attract customers with a promotion, and then create for them maximum diffi-
culties, to minimize the using of the promotion. The number of loss leaders and high-
low pricing are not very high; they are limited to the retailing promotions.
Brand promotions Ratail promotions

n 52= =
n 137
Category of
# of
category dissatisfier Number I Parcent·
aga
Number I Percant·
aga
Dissatisfiers in promotion design
1. Money back actions 11 21,1 2 1,5
2. Nature of premium 9 17,3 11 8,0
3. Conditions 15 28,8 27 19,7
4. Expiry date of action 13 25,0 13 9,5
5. Discount is no discount 8 15,4 22 16,1
6. Third party involved 4 7,7 3 2,2
7. Different prices 2 3,9 1 0,7
8. Loss leader 0 0 12 8,8
9. High/low pricing 0 0 5 3,7
10. Intrusion into privacy 0 0 1 0,7
Dissatisfiers in promotion realization 'back office', not due to personal contact
11. Retailers or person ne I not informed 4 7,7 4 2,9
12. Faulty mailing 2 3,9 2 1,5
13. Faulty advertisement 0 0 44 32,1
14. Out of stock 0 0 9 6,6
Dissatisfiers in personal contact only
15. No faulty attributes in promotion de- 2 3,9 16,8

sign or realization 'back office'
Fig. 10: Dissatisfiers In promotion design and realization
Note: Percentages work out higher than 100% because in most incidents more than one dissat-
isfier is involved.
Faulty advertisements are very common dissatisfiers in retailing promotions. This

can be explained in part by a flexible behavior with in-store promotions; e.g. an an-
nounced low price can be changed in aminute. It also reflects the less professional
level of advertising in free local papers or in pamphlets made by the retailer himself.
138 Jan A. Schulp
3.4.2 Employee and management behavior

In section 3.3 it was stated that the employee behavior was composed of
• a behavior that can be described in objective terms, affecting the perception

of service fairness; and
• a behavior that could less easily described in objective terms, affecting the
feelings and the self esteem of the customers.
In the first group, three types of behavior can be distinguished. Each behavior is ex-
emplified in one or more cases in the appendix:
• Refusal to fulfill: In advertisement or in the shop, the price and other condi-
tions are indicated in clear words. Vet the service provider refuses to per-
form in accordance with the offering. This refusal is a violation of the Dutch
Civil Code, unless a very obvious error is involved (example: #076).
• Not keeping promises: in connection with a service offering, spoken or
written promises are made which are not kept afterwards (example: #621).
• Turning the meaning of the promotion in a very particular way: related to
"refusal to fulfill", but the employee starts a complex reasoning to justify his
refusal. Often combined with "cheeseparing" behavior (example: #442).
In the second group, the following types of behavior could be identified:
• Trying to help/showing some empathy: The employee clearly perceives

the customer's dissatisfaction; shows understanding; cannot do anything
about it, but gives an explanation without success. The helplessness of the
employee is often connected with lack of empowerment (example: #109).
• Playing ignorant: There is absolutely no reaction upon the customer's dis-
satisfaction. Insofar as 'communication' continues, the service provider
brings up different subjects, not what's important according to the customer
(example: #351).
• Fobbing off: "Weil, that's the way we do it. Nothing against it, though?" (ex-
ample: #175).
• Bloody excuse: Employees try at least to give a kind of an explanation. If

they make it too complicated, it borders on the "Iying and cheating" category
(example: #93).
• Cheeseparing: A behavior in which step by step the promotion benefits are
withdrawn. It is mostly combined with refusal to fulfill or turning the meaning.
The unspoken message to the customer is: "I shall see how I squeeze as
much money out of you as I can, and give you back as little as possible (ex-
ample: #208).
• Overt dishonesty: Openly admitting dishonest practices in making a pro-
motion. Not easily distinguished from fobbing off (example: #490).
• Lying and cheating: Piling one lie upon the next excuse, often in a contra-
dictory way. Never admitting that one's wrong. The complex form of "bloody
excuse" (example: #368).
• Making a fool of the customer: If something goes wrang, it is the cus-
tomer's fault, or even, the customer is fraudulent (Oliver 1996, pp. 369-370).
It exemplifies the attribution theory of Bitner/Booms/Mohr (1994, pp. 96-97)
(example: #454).
The results (see figure 11) clearly reflect the difference in personal contact in brand
and in retail promotions. In 39 out of 52 cases of brand promotions, no personal
contact took place as contrasted with 34 out of 137 cases of retail promotions.
Even if the employees behave dishonestly, they show in most cases a self-
justification behavior. Instead of a blunt refusal (30 cases), in most of the cases they
promise to do samething about the situation, or they seek an explanation for the
non-fulfillment of the promotion offer (5 plus 41 cases). One also sees inept attempts
to spare the customer's feelings somewhat, whilst keeping an eye on the ultimate
goal: giving the customer as little as possible ("cheeseparing"). Overt wickedness:
dishonesty and making a fool of the customer is not very common.
In only 4 cases, intervention by management or customer service department is

mentioned. In all these cases, the management conforms to the judgment of the em-
ployee. From the data it cannot be made clear whether management intervention is
140 Jan A. Schulp
very rare, or if management intervention is generally successful and thus prevents

complaints to the Consumentenbond.
Brand Retail
Class # Description promotions promotions
n = 52 n =137
0 No service provider behavior stated 39 34
Objective behavior
1.1 Refusal to fulfill 1 30
1.2 Not keeping promises 5 5
1.3 Tuming the meaning of the promotion 1 41
Subjective behavior (customer feelings)
2.1 Trying to help/empathy 0 6
2.2 Playing ignorant 2 6
2.3 Fobbing off 3 17
2.4 Bloody excuse 0 26
2.5 Cheeseparing 1 31
2.6 Overt dishonesty 1 11
2.7 Lying & cheating 0 15
2.6 Customer made 1001 of 2 9
Fig. 11: Classification of the employee behavior
Note: The lolal number of items is grealer than 52 or 137 respectively, because in many inci-
denls more Ihan one kind of employee response could be idenlified.
3.4.3 Customer Behavior

The consumer process between perception of the promotion and the writing of a
complaint is complex and varied. It can consist of a single step, but incidents involv-
ing multiple steps are more common. Figure 12 gives Ihe classification of Ihe cus-
lomer behavior.
ReJects < Immediately
after analysis 18
1A
/
Rejeetlon L d t Complaint to
olofter 2A ea s O l Consumenten- 28
by eustomer satlsfaetion bond only
Customer
pereelt~es --
t
En a es in
di:eu~sion --+
i Leads to
Purehase - - + dissatisfaction Solved
promo Ion 1 \ ( satislaetorily
Rejection Complaint
01 custo~er 20 to retailer
by service
provider Not solved
<
satislactorily
Satisfied Complaint to
2C
P h ( <consumenten- 3A
ure ase bond only
. . Solved
Dissatisfied satisfaetorily
Complaint to
retailer Not solved
satisfactorily 38
Fig. 12: A cladogram 01 customer behavior in reaetion to promotions
The classes can be described as folIows; examples are in the appendix.
Class 1: The customer perceives the promotion and rejects it. She is so indignant
about the negative traits of it that she does not keep this knowledge for herself, but
writes a letter to the Consumentenbond. Two cases can be distinguished:
• 1A: The customer distinguishes immediately the negative traits of the pro-
motion (examples: #420, 730).
• 18: The customer is attracted by the promotion initially and starts analyzing
the promotion. After some calculations she perceives that something is
wrong (examples: #399, 438).
Class 2: The customer is attracted by the promotion and contacts the service pro-
vider about it. An interaction between service provider and customer ensues. This
can take the character of asking for information, but also of a discussion, sometimes
142 Jan A. Schulp
even of a hostile nature. After the discussion, four different outcomes can be distin-
guished:
• 2A: Rejeetion by the customer. The customer perceives after the discussion
that something is wrong with this promotion. She is so indignant about this
that she does not keep this knowledge for herself, but writes a letter to the
Consumentenbond (examples: #093,368).
• 28: The customer decides to purehase the article for which a promotion is
made and to fulfill the rest of the conditions. If the outcome is unsatisfactory,
the customer makes direet complaints with the Consumentenbond (exam-
pies: #076, 332).
• 2C: The customer decides to purehase the article for which a promotion is
made and to fulfill the rest of the conditions. If the outcome is unsatisfaetory, the
customer first complains with the service provider and after an unsatisfaetory
outcome, still recurs to the Consumentenbond (examples: #109, 193).
• 20: After a customer has engaged in a discussion with the service provider,
it can be the service provider who is unable or unwilling to serve the cus-
tomer. Sometimes this is involuntary, e.g. in case of an out of stock situation.
Sometimes, however, the service provider refuses to work further for the
customer. This situation resembles what Stauss (1997, p. 10) calls "inten-
tionally pushed away customers· (examples: #624, 667).
Class 3: The customer perceives the promotion, accepts the offer, makes a pur-
chase or starts to collect stamps, bar codes etc. When it comes to receiving the
money back, the premium, etc. trouble starts.
• 3A: The customer makes direct complaints with the Consumentenbond (ex-
amples: #007,374).
• 38: The customer first complains with the service provider, and after an un-
satisfactory outcome, still recurs to the Consumentenbond (example: #490).
From a managerial viewpoint, the classes 1, 28 and 3A have a very important char-
acteristic in common: the dissatisfied potential (1) or aetual (28, 3) customer never
teils the service provider that she is dissatisfied and why; she writes immediately to
the Consumentenbond. The service provider doesn't even know that something is
wrong with the promotion until it is too late: when he receives a letter from the Con-
sumentenbond. He still can react and make amends. If not, he will see his name
printed in the Stekeligheden. This is disagreeable.
In classes 2A, 2C, 20 and 3B the service provider should have perceived that
something was wrong. After one or more contacts the customer feels that she has
been treated in a sloppy, indifferent or outright brutal way by the service provider,
and she seeks redress with the Consumentenbond.
The patterns of customer behavior are divided between the categories as shown in
figure 13.
Figure 13 also clearly reflects the difference in proxemics (contact) between custom-
ers and brand manufacturers and retailers respectively. For brand promotions, cases
of class 2 are practically absent; the very few cases are about contact by telephone
between interested customers and the manufacturer. The biggest categories of cus-
tomer behaviors are 1Band 3A. This means that a brand manufacturer sei dom re-
ceives complaints that can warn him for flaws in his promotion. The vast majority of
complaints are made directly to the Consumentenbond without any previous com-
plaint to or contact with the manufacturer. This stresses the importance of a promo-
tion 'with ethics built in' as stated in section 1.4.
In retailing promotions, all classes of consumer behavior are represented but the
classes 2A and 3B are the most prominent. The retailers concerned are warned in
time, by customer complaints, that something is wrong with their promotions. For
class 2A, a certain sensitivity for the irritations of non-buying customers is needed in
order to prevent complaints with third parties. For classes 2C and 3B, normal good
practices of complaint management could have prevented complaints with the Con-
sumentenbond.
144 Jan A. Schulp
Brand promotions Retail promotions

n = 52 n = 137
# of Class
class description Number Percent- Number Percent-
age age
1A Promotion rejected immediately after 1 1,9 9 6,6

perception
18 Promotion rejected after analysis, 18 34,6 8 5,9

without contact with service provider
2A Olfer rejected after discussion with 1 1,9 30 21,9

service provider
28 Olfer accepted after discussion; leads 0 0 11 8,0

to dissatisfaction; complaint to C8
only
2C Olfer accepted after discussion; leads 0 0 15 10,9

to dissatisfaction; complaint to service
provider first; then to C8
20 Customer eng ag es in discussion with 2 3,8 15 10,9

service provider; customer refused,
insulted or otherwise turned away
3A Olfer accepted without discussion; 24 46,2 10 7,3

leads to dissatisfaction; complaint to
CB only
38 Olfer accepted without discussion; 6 11,6 39 28,5

leads to dissatisfaction; complaint to
service provider first; then to C8.
Total 52 100 137 100
Fig. 13: Customer reactions in brand and retail promotions
3.5 The Costs and the Benefits

The cost of the problems with laulty promotions can be, according to Oliver's Eco-
nomic Complaint Model (Oliver 1996, p. 362): 105501 money, 105501 utility, waste of
time and effort; all of this connected to the importance to the customer. The direct
costs of making a complaint with the Consumentenbond are: time, exertion, the cost
of a letter and stamps. The benefits of complaining with the Consumentenbond can
be: money back, replacement, extra reimbursement, correcting the problem, apol-
ogy, and catharsis. From the Stekeligheden it is never clear what the complainants
consider as success. Is it money back, replacement or only the pleasure to see a
firm that treated her in a disgusting way exposed? For the greater part, it cannot be
decided. Vet, the Stekeligheden give some clues for partial conclusions. The first is
about the monetary 1055. In figure 14 the amount of money involved in the incidents
is presented. In most cases it is insignificant. Often, the costs of complaining with the
Consumentenbond must have been almost as high, or higher than the problem was
worth. In retaH promotions a few cases are concerned with high cost items: furniture,
electronics etc. Vet, half of the cases in which the monetary value is stated or rele-
vant, very low or even ti ny amounts of money are involved.
Brand promotions Retail Promotions

Amount involved n =52 n = 137
(inn
Number Percentage Number Percentage
< / 2,50 6 11,6 18 13,2
/2,50 - /10 13 25,0 31 22,6
/10-/25 9 17,3 8 5,8
/25-/100 1 1,9 16 11,7
/100-/500 0 0 8 5,8
/500- /1000 0 0 4 2,9
> / 1 000 0 0 2 1,5
Amount not stated or not relevant 23 44,2 50 36,5
Total 52 100 137 100
Fig. 14: Amount 01 money involved in cases on retail and brand promotions
It may be concluded, therefore, that financial reasons for complaining with the Con-
sumentenbond can playa minor role but are in most cases not foremost in the per-
ceived losses or benefits. In certain cases we see considerable 1055 of time and ef-
fort; in other cases 1055 of the utility of the goods or services.
As stated in section 2.1, the editor checks all the Stekeligheden for truth. If the serv-
ice provider gives a relevant explanation, or some kind of redress to the customer,
the incident is not published at all, or published with a postscript. The editor could
not give an indication of the number of successful, unpublished cases. Stekelighe-
146 Jan A. Schulp
den with a postscript were the only data available. Figure 15 is the result of a count
of Stekeligheden with a postscript, distinguishing between explanation and recovery.
Benefits in terms of recovery or apology do occur but by they are not very common.
Catharsis may be a strong motivator for complaining with the Consumentenbond.
The data do not permit a definitive conclusion.
Effect of complaint with RetaH promotions Brand promotions

Consumentenbond n 137= n 52 =
No reaction to Consumentenbond 106 42
Explanation 13 4
Recovery 18 6
Fig. 15: Service provider reactions published in "Stekeligheden" after intervention of the Con-
sumentenbond
4 Discussion
4.1 Disconfirmation of Expectations
The Gränroos-Gummesson model of consumer perceived service quality can be
used in interpreting the process of complaining with the Consumentenbond. In this
model, the difference between expected service and perceived service is decisive for
the service quality. Most of the promotions are for commodities; only a few are for
preference, specialty or shopping goods. The classification of goods is from van der
Kind (1995, p. 28). Therefore, the expectations about these goods are modest; an
acceptable quality for a low or acceptable price is all the consumer requires. When a
promotion is started, expectation rises: to the modest benefits of the product, so me-
thing attractive is added. Extremely attractive "1055 leaders" arouse even great ex-
pectations. If the implicit and explicit promises of the service provider are kept, the
customer is satisfied with the product and the promotion; this can lead to brand or
shop loyalty. Problems arise when expectations are raised by a promotion, and the
promotion ends in a disappointment to the customer. However, disconfirmation of
expectations never can provide the complete answer on the research question. If
promotions trigger the 1A and 1B or 2A type of customer behavior, the expectations
are disconfirmed almost as soon as they arise; they cannot be considered as ex-
pectations at all. In these cases, only the feeling remains that the service provider
has tried to cheat the customer. And the customer perceives the service quality as
so negative that she takes a very strong form of action: seeking publicity with the
Consumentenbond. She is an activist after all.
The disconfirmation paradigm can be refined further using the gap-model

(Parasuraman/Berry/Zeithaml 1985). Two gaps are important:
• Gap 1: does my customer like what I think s/he likes?

• Gap 4: how does marketing communication affect the customer expecta-
tions?
In the Stekeligheden are some very clear cases of not thinking the customer's way:
some market research could have revealed the most popular sizes of dog's collars
(example: #002). For a well-to-do industrialist or advertising person, perhaps
spending f 500,- on meals in a hotel is no problem if the hotel rooms are for free.
For.someone on a minimum wage it is a big problem and therefore an unattractive
promotion (example: #374).
4.2 Unresponsiveness
Unresponsiveness of the service provider is another big dissatisfier, both in brand
and in retail promotions. The customer behavior classes 2A, 2C, 2D and 3B start
with a discussion or a complainl. The service provider doesn't give an adequate re-
action. These classes make up 72,2% of all cases. Complaints to the Consumenten-
bond could have been prevented by a beUer responsiveness to the customer's com-
plaint. It should be kept in mind thatthe sampie of "respondents" consists completely
of 'activists' (Singh 1990). It is not everybody's approach to utter her dissatisfaction.
This is a limitation of this study. Yet, because of the high damage potential of these
activists, this style of response deserves attention.
148 Jan A. Schulp
4.3 Perceived justice: stability and controllability

The work of BlodgetUGranboislWalters (1993) offers a different approach for ex-
plaining the complaining behavior of customers. BlodgetUGranboislWalters made a
model on negative word-of-mouth and repatronage intentions as functions of per-
ceived justice (figure 3). Perceived justice is correlated with the factors: stabil-
ity/controllability, attitude toward complaining, likelihood of success and product im-
portance, which in their turn have both direct correlations with negative word of
mouth and repatronage intentions, and a correlation via perceived justice. All the
dissatisfied customers who provided the "Stekeligheden" have a positive attitude to
making complaints - they know that complaining often helps. Only in this exceptional
case about which they contacted the Consumentenbond, it did not work: the likeli-
hood of success had changed into certainty of failure, and this had triggered a form
of action even stronger than negative word of mouth: it made them seek publicity.
The criticism of the suppliers of Stekeligheden is about stability/controllability, prod-
uct importance and perceived justice.
Considering stability/controllability, in most of the cases a little more care, attention

and common sense on the part of the service providers could have prevented the
unpleasant incidents. As it can be seen in the appendix, out of stock situations, lack
of information to personnel or retailers, announcing promotions long after expiry
dates, unclearness ab out exact meaning of the promotion can all be described as
lack of controllability/stability. In the model of BlodgetUGranboislWalters lack of sta-
bility/controllability in itself leads only to negative word-of-mouth via perceived injus-
tice. The contributors to Stekeligheden feit they were treated unjustly by service pro-
viders who conducted a promotion in a careless, sloppy way. Lack of stabil-
ity/controllability is part of the explanation for seeking publicity.
4.4 Perceived justice: clear dishonesty and

wounded feelings
In the model of BlodgetUGranboislWalters, perceived injustice in itself is the most
important factor for engaging in negative word-of-mouth. In the present research,
most of the cases verge to dishonesty both in promotion attributes and service pro-
vider behavior. The attribute categories 3 (conditions), 5 (discount is no discount), 7

(different prices), 8 (Ioss leader) and 9 (highllow pricing) are clearly dishonest. Ex-
amples of brand promotion attributes with elements of dishonesty are:
• discouraging the customer to read in time, i.e. in the supermarket, the full
information (examples: #007,243, not in appendix, 399);
• giving incomplete information about financial consequences of accepting the
promotion (examples: #374,399);
• putting unannounced limitations to a promotion (example: # 033).
Dishonesty in service provider behavior is very common in problems with retail pro-
motions (figure 11). The categories 2.3 to 2.7, from fobbing off till Iying and cheating
are large.
When considering not only the attributes of the promotions but also the employee
behavior, then dishonesty becomes even clearer. This view can be supported mainly
from section 3.4.2 and figure 11. In more than half the retailing promotions about
which customers complained, dishonesty was involved to a degree that could in
principle be open to legal action. Often, the dishonesty takes the form of turning the
meaning of the promotion conditions. This is in accordance with the view of Bit-
ner/Booms/Mohr (1994) that employees have a bias towards blaming an unsuc-
cessful service encounter on the system and the customer. Also, reactions like "you
cannot expect something good for this price" (example: #074) or "be glad with the
discount you did get" (example: #478) confirm the findings of Seiders (1997) that
service fairness never can be bartered for a low price. Overtly dishonest actions of
employees do occur but make up only a minority of the cases. Also, accusing the
customer of dishonesty (Oliver 1996, pp. 369-370) accounts in some cases for the
dissatisfaction of the customer (example: #544, not in appendix).
In retailing promotions, 117 behaviors could be scored in 137 cases that caused
negative affects with the customers: the feeling that they were considered as greedy,
dishonest, stupid, as a nuisance by people who were paid to provide service to
them! The worst cases are those when a salesperson refuses to serve a customer
150 Jan A. Schulp
any further because of her offen ce that she stands for her rights (example: #f567). In
short: the customers are insulted.
The reaction of the dissatisfied customer can have been: "If you cheat me and rob
me this way, be it only for 80 cents, 1'11 tell it in so many hundred thousands of copies
of the Consumentengids".
4.5 The answers to the research question

From the qualitative analysis of only 192 critical incidents from brand and retail pro-
motions, an answer upon the main research question can be formulated. The ques-
tion ran: "What is so special in the incidents reported that they have caused this
strong form of discontent, i.e. a discontent making the customer seek advice and
publicity with the Consumentenbond?" The answers found, are:
• A promotion raises customer expectations considerably. Not to fulfill these

expectations makes a customer very discontented (disconfirmation).
• If the producer of the goods to be promoted is unresponsive to the custom-
ers' problems and if the customer has a positive attitude to complaining, the
customer chooses to go public (unresponsiveness).
• The customer's discontent is enhanced because she perceives that with
some care and common sense the stupid mistakes in the promotion could
have been prevented (lack of stability/controllability).
• The customer takes this strong kind of action because she feels that she has
been fobbed off, cheated (perceived injustice).
• The customer takes this strong kind of action because she feels insulted and
wounded by the service provider behavior (wounded feelings).
5 Managerial implications
Be generous or else refrain from promotions altogether
Only start a promotion if you can afford it. If your customer has to pay postage or
telephone costs, you cannot afford the promotion you have in mind or if you limit the
Critica//ncidents in Dutch Consumer Press 151
number of customers who get money-back or a premium, you cannot aftord the pro-
motion, it is proposed to think of something simpler and cheaper. Oare to be gener-
ous when realizing a promotion. Better no promotion at all than a promotion from
which you keep paring away more and more profit for the customer until practically
nothing remains. Avoid the use of 1055 leaders and unrealistic high/low pricing. They
give short term profit, but a long term bad reputation. Highllow pricing will be ex-
posed anyway since the introduction in Flanders and The Netherlands of Lowest
Price Agencies such as Miep Kniep - the Cost Cutting Company.
Try to think the customer's way
Sit down, calculate the costs to the customer in terms of money, trouble and time,
and estimate the customer benefits. If you were a customer, would you think your
promotion would be worth the money, the trouble and the time? 00 not suppose that
what is attractive to you, is attractive to the customer as weil.
Keep an eye to your expiry date
Announce your expiry date clearly. Also, think the customer's way and ask yourself
if the duration of your promotion long enough. Can the amount of stamps or tokens
be collected by a normal user of your article within the duration period of your pro-
motion? If not, you make an ungenerous promotion. Make sure that after the expiry
date alltraces of your promotion have been removed from the shops.
Keep everyone informed
Give the customer a fair chance to inform herself in detail about your promotion in
the shop. Check and double check your advertisements, both in-store and in pam-
phlets and papers. If you make amistake, take it like a man. According to Outch law
you are obliged anyway to deliver for the price advertised - unless the error is clear,
e.g. a TV set for f 50,-. Make sure that there is no difterence in interpretation of the
promotion between the customers, all members of your staft and yourself. Empower
your staft to solve unexpected problems in the favor of the customer.
152 Jan A. Schulp
Keep your logistics under control
Are all elements where they should be? Are the stamps upon the articles? Can the
customers easily obtain business books? Have you got enough promotion articles?
Or can you order additional items at a short notice? 00 all you can to avoid out of
stock situations with a promotion. Even if you announce "op = limited supply", give
the customers a realistic chance to profit.
Three last recommendations
• Keep It Stupid Simple (KISS).

• Be aware of the risks of a brand promotion with low personal contact. Check
and double check it before you get unexpected troubles with an angry cus-
tomer or a consumers' organization.
• Make an ethical audit with each promotion.
References
Bearden, W.O./Oliver, R.L. (1985): The Role of Public and Private Complaining in
Satisfaction with Problem Resolution, in: Journal of Consumer Affairs, Vol. 19,
No. 2, pp. 222-240.
Bitner, M.J./Booms, B.H./Tetreault, M.S. (1990): The Service Encounter: Diag-

nosing Favorable and Unfavorable Incidents, in: Journal of Marketing, Vol. 54,
No. 1, pp. 71-84.
Bitner, M.J./Booms, B.H.lMohr, L.A. (1994): Critical Service Encounters: The Em-
ployee's Viewpoint, in: Journal of Marketing, Vol. 58, No. 4, pp. 95-106.
Blodgett, J.G./Granbois, D.H./Walters, R. (1993): The Effects of Perceived Justice

on Complainants' Negative Word of Mouth Behavior and Repatronage Inten-
tions, in: Journal of Retailing, Vol. 69, No. 4, pp. 399-428.
Clark, G.L./Kaminski, P.F./Rink, D.R. (1992): Customer Complaints: Advice on how

Companies Should Respond Based on an Empirical Study, in: Journal of Con-
sumer Marketing, Vol. 6, No. 1, pp. 5-14.
Cumby, J.A./Bames, J.G. (1997): How We Make Them Feel: A Discussion of the
Reactions of Custorners to Affective Dimensions of the Service Encounter, in:
New and Evolving Paradigms, AMA Special Conferences, Dublin, Ireland, on
12 -15 June 1997, pp. 140-152.
Decker, A./Meissner, H. (1997): The Sequential Incident Technique for Innovations

- An Instrument for Generating Improvements and Innovations, in: Quality
Management in Services VII, Papers for the EIASM workshop at Agder College,
Kristiansand, Norway, on 28 - 29 April 1997, Book 2, Kristiansand: Senter for
service, organisasjonsutvikling og laering ved Hogskolen i Agder.
Edvardsson, B. (1997): Kundmissnöje och klagomälshantering - studier av kollek-

tivtrafik med kritisk-händelse-metoden, Forskningsrapport 97:3, Centrum för
tjänsteforskning (CTF), Karlstad, Sweden.
Edvardsson, B./Karlsson, P.lLindström, A. (1995): Kvalitetskartor, EU säU aU ar-

beta med kvalitetsutveckling i Norrköpings kommun, Centrum för tjänstefor-
skning (CTF), Karlstad, Sweden.
Edvardsson, B./Luukkonen, M. (1996): Kritiska händelser, en studie inom Göte-

borgsregionens Lokaltrafik AB, Forskningsrapport 96:1, Centrum för tjänstefor-
skning (CTF), Karlstad, Sweden.
Grönroos, C.G. (1992): Service Management and Marketing, Lexington.
Hart, C.W.L.lHeskett, J.L./Sasser, W.E. (1990): The Profitable Art of Service Re-
covery, in: Harvard Business Review, Vol. 68, No. 4, pp.150-156.
154 Jan A. Schulp
Heskett, J.L.lSasser, W.E.lSchlesinger, L. (1997): The Service Profit Chain, New

York.
Holme, I.M.lSolvang, B.K. (1983): Forskningsmetodik. Om kvalitativa och kvantita-

tiva metoder, Lund.
Kaufmann, P.J.lSmith, N.C.lOrtmeyer, G.K. (1994): Deception in Retailer High-

Low Pricing: A "Rule of Reason" Approach, in: Journal of Retailing, Vol. 70,
No. 2, pp. 15-138.
Keaveney, S.M. (1995): Customer Switching Behavior in Service Industries: An Ex-

ploratory Study, in: Journal of Marketing, Vol. 59, No. 2, pp. 71-82.
Kelley, S.w.lHoffman, K.D.lDavis, M.A. (1993): A Typology of Retail Failures and

Recoveries, in: Journal of Retailing, Vol. 69, No. 4, pp. 429-452.
Khan, M.M.lMcCleary, K.M. (1995): A Proposed Model for Teaching Ethics in Hos-
pitality, in: Hospitality & Tourism Educator, Vol. 8, No. 4, pp. 7-12.
Kind, R.P. van der (1995): Retail Marketing, Deventer.
Krips, F.J. (1990): Vijftig manieren om een klant kwijt te raken, Utrecht.
Oliver, R.L. (1996): Satisfaction, A Behavioral Perspective on the Consumer, New

York.
Olsen, M.J.S. (1992): Kvalitet i banktjänster - Privatkunders upplevda problem med

banktjänster - en studie med hjälp av kritisk-händelse-metoden, Centrum för
Olsen, M.J.S. (1993): At! utveckla och styra tjänsteverksamhet. Tjänsteprocessen i

fokus. Forskningsrapport CTF, Karlstad, Sweden.
Olsen, M.J.S.lBertiisson, T. (1991): Studies in Service Quality with the Aid of Criti-
cal Incidents and Phenomenography, Research Report 91: 13, Centrum för
Parasuraman, A.lBerry, L.L./Zeithaml, VA (1985): A Conceptual Model of Service

Quality and its Implications for Future Research, in: Journal of Marketing,
Vol. 49, No. 3, pp. 41-50.
Price, L.L.lAmould, E.J.lTiemey, P. (1995): Going to Extremes Managing Service

Encounters and Assessing Provider Performance, in: Journal of Marketing,
Vol. 59, No. 2, pp. 83-97.
Richins, M.L. (1983): Negative Word of Mouth by Dissatisfied Customers: A Pilot

Study, in: Journal of Marketing, Vol. 47, No. 4, pp. 68-78.
Roos, I. (1997): Customer Switching Behavior in Retailing, Licentiate Thesis,

Swedish School of Economics and Business Administration, Helsinki, Finland.
Ruyter, K. de (1994): Het (v)oordeel van de ontevreden klant, Utrecht.
Schulp, J.A. (1997): Critical Incidents in Dutch Consumer Press I: Methodology,

Research Program and an Analysis of Manufacturing Industry Promotions, in:
Quality Management in Services VII, Papers for the EIASM workshop at Agder
College, Kristiansand, Norway, on 28 - 29 April 1997, Book 1, Kristiansand:
Senter for service, organisasjonsutvikling og laering ved Hogskolen i Agder.
Schulp, J.A. (1998): Critical Incidents in Dutch Consumer Press 11: Dissatisfiers in
Retail and Brand Promotion Incidents. in: Quality Management in Services VIII,
Papers for the EIASM workshop at the Department of Business Administration
Ingolstadt, Catholic University Eichstätt, Germany, on 20 - 21 April 1998,
Vol. 2, Ingolstadt, pp. 675-698.
Seiders, K. (1997): How Consumers Evaluate Service Fairness, in: New and Evolv-
ing Paradigms, AMA Special Conferences, Dublin, Ireland, on 12 - 15 June
1997, pp. 153-154.
Singh, J. (1990): A Typology of Consumer Dissatisfaction Response Styles, in:

Journal of Retailing, Vol. 66, No. 1, pp. 57-99.
Stau55, B. (1993): Using the Critical Incident Technique in Measuring and Manag-
ing Service Quality, in: Scheuing, E.E.lChristopher, W. (eds.): The Service
Quality Handbook, New York, pp. 408-427.
Stau55, B. (1997): Regaining Service Customers - Costs and Benefits of Regain

Management, in: Quality Management in Services VII, Papers for the EIASM
workshop at Agder College, Kristiansand, Norway, on 28 - 29 April 1997, Book
1, Kristiansand: Senter for service, organisasjonsutvikling og laering ved Hog-
skolen i Agder.
Stau55, B.lSeidel, W. (1996): Beschwerdenmanagement - Fehler vermeiden -

Leistung verbessern - Kunden binden, München.
Tax, S.T./Brown, S.W./Chandrashekaran, M. (1998): Customer Evaluations of

Service Complaint Experiences: Implications for Relationship Marketing, in:
Warland, R.H./Herrmann, R.O./Willits, J. (1975): Dissatisfied Customers: Who

Gets Upset and Who Takes Action, in: Journal of Consumer Affairs, Vol. 9,
No. 4, pp. 148-163.
Westendorp, P.H. van (1997): Ondervraging en andere kwaliteitsproblematiek, in:

Tijdschrift voor Marketing, Vol. 31, No. 3, pp. 26-27.
156 Jan A. Schulp
Appendix
On the following pages, a nu mb er of "Stekeligheden" cases is given in a shortened

translation in English. Classification of promotion attributes, customer behavior and
employee behavior are given. The appendix contains all cases mentioned in the text
by their number.
Incident Description Class of Class of Service

# promotion customar provider
attribute behavior behavior
002 Frolic dog food afters a free dog collar for stamps. 9 3A 0
Customer orders a collar and writes she has an
lrish Setter. Receives the standard collar of 40 cm,
only fitting for the smallest dogs.
007 Sporex hand cream bought for money back action. 1 3A 0

At horne, folder opened; action only with purchase
of two Sporex products.
033 Money back action with a Cote d'Or chocolate 1 3A 1.1,2.5

product: for two bar codes f 2,50 back. Customer
sends in his bar codes in time; still he gets no
money back because money back action turns out
to be limited to the first 30,000 customers.
074 Opening of Praxis DIY: special offer of same tools 2 2C 2.3,2.6

for very low price. The tools break and bend very
soon. Back to supplier. Reaction: "You should
understand it can't be any good for that price".
076 Discount announced on Coca-Cola. Refused at the 5 2B 1.1,2.5

checkout of Supermarket Sanders Haaksbergen
"because our cola - price is so low already". No
customer reaction stated.
093 Free photo roll for testing as "ofter of the month". 5,12,13,10 2A 2.4
Within a few days a customer wants to try the
after. He doesn't get it for free "because it is the
ofter of the week" - the same week that isn't yet
over. And on top of that: all the free films are gone
- or so they say. No excuses.
109 In the Week of the Children's Book, the book pre- 2 2C 2.1
sent is out of stock. No acceptable alternative
available. Books are bought in spite of this, but
father and children are disappointed.
175 Discount on cellulose diapers: out of stock one day 12,4 1C 2.3
after advertising. Action started too early and
customer had to be disappointed.
190 Discount on a book: original price f 125,-; now for 9 3A 0

139,-. Original price was actually f 69,-.
Incident Description Class of Class of Service

# promotion customer provider
193 Discount on a pair of trousers, for f 69,50. Under- 5 2C 2.6

neath the discount price tag is the original one .
for f 69,50. Explanation: "That's the way we da it
always'.
208 HUBO gives business stamps: one stamp with a f 11 3A 1.1,2.5

10.- purchase, and f 25.- back for a complete
book. But a customer only receives 8 stamps with
a f 194,- purchase, not 19 or 20. "Otherwise
you've got a complete book far tao soon".
281 Supermarket Weltevreden (= very satisfied) or- 13 2C 0

ganizes a coloring contest for children; every par-
ticipant is promised a button but before end of
action, buttons are out of stock. Children give the
solution themselves: why not buy as many buttons
as coloring plates?
332 Yves Rocher offers articles for only f 5,-. Custom- 7 2B 1.3
ers go es to the shop; selects articles; only then
she is told that she has to spend at least f 15,-;
otherwise she does not qualify for the discount.
351 Supermarket De Boer Emmen advertises cheese 6 2B 1.1,2.2

at a discount. Personnel not informed. Only after
rescuing advertisement from dustbin, cheese is -
reluctantly- provided at advertised price.
368 Color monitor advertised by ZYZ Division, Am- 10,12 2A 1.3,2.7

sterdam ZO, at low price. An interested customer
first checks by telephone if monitor is in stock.
After a confirmation he travels to the shop. Then
the monitor turns out to be only available WITH a
computer. A more expensive monitor is then pro-
posed to him.
374 Pally tea biscuits offers three free hotel nights for 3,6 3A 0
five bar codes. After receiving the vouchers the
customer finds out that she has to spend at least f
250,- for breakfast and dinner.
399 Ajax cleaner offers a free fiower bulb package. 3 1B 0

From the booklet she perceives that she has to
send a coupon to bulb firm for a free catalogue.
Then she has to spend at least f 15,- with that firm
before she qualifies for a "free" bulb package.
420 Philishave mailing for a man sent to a woman. 12 1A 0
438 Dommelsch beer offers a set of beer glasses worth 5 1B 0

f 17,50 for 150 saving points. This involves drink-
ing a lot of beer. Furthermore he has to pay f
14,95. And p&p is f 1,60; bank costs f 0,60. A lot
of work for f 0,35.
158 Jan A. Schulp
Ineident Description Class of Class of Service

442 At Waayer Poultry, with Dirk van der Breek, 15 2D 1.5,2.5,

Bilderdijkkade, Amsterdam: Discussion on priee ot 2.8
chicken: salesperson retuses to seilless than 500
g tor a discount price. Long discussion; customer
was retused the sm aller quantity tor the discount
price and was sent away: "It you don't want it tor
the ordinary price please tuck off". After CB inter-
vention, management admitted salesperson was
wrang.
454 Combi-Foto Esser Amsterdam ofters a two-pack ot 15 2D 1.1,1.3,

films tor f 9,96. This advertised price is interpreted 2.8
by salespeople as f 9,96 tor one film. "You don't
read what's written! f 9,96 tor \wo films would be
tar too cheap!" Films at advertised price retused.
478 Super in Bovenkarspel. Park chops tor a discount 15 3B 1.1,2.4,

price. Customer buys 2 kg. Discount only applied 2.5
to the first kilogram. Customer complains. Reply:
'Be glad with the discount you got".
490 Prenatal ofters a quilt cover with a Mickey Mouse 12 3B 2.3

pattern tor only f 25,-. Turns out to be sold only
with a pillow case, total price : f 75,-. Customer
calls customer service. 'Sorry, I can't do anything
about it; this kind ot things happens so often".
556 Speet, Alphen ad Rijn ofters camcorder with f 8 2A 1.2,2.6,

900,- oft. Couple trom Gouda makes a eall: are 2.7
the camcorders on stock? Yes!! A camcorder can
be reserved tor them. Within one hour they are in
Alphen. No camcorder. A call to the Leiden
branch. No camcorder with this discount. Weil, no
problem, Speet has any kind ot camcorders tor
them .....
621 Reddy trying oil promises a recipe booklet tor a none 4 1.2
bar code. Bar codes se nt. No reaelion. Reminder:
no eftec\. Booklet only sent after action ot Con-
sumentenbond.
624 A Super branch. Customer receives "Rocks" sav- 10 2D 1.1,2.3

ing stamps tor months. On a certain day these are
retused to hirn unless had consents to apply tor a
customer pass which involves providing much
personal intormation.
667 Joosten Paint & Wall paper: with customers' pass 3,11,15 2D 1.3,2.7,
20% oft on items trom stock; 10% oft on items that 2.8
have to be ordered. 9 rolls ot wallpaper are sup-
plied to a eustomer tram stock; 1 additional roll
has to be ordered; on all the 9 rolls only 10% dis-
count is given. When the customer protests the
pass is destroyed and customer is torbidden ever
to buy at Joosten's.
Ineident Deseription Class of Class of Service

730 A BMW driver receives a mailing, suggesting con- 10,12 1A 0

nection with his brand of car. Mailing is actually for
erotic underwear and pornography.
Appendix: Critical incidents in Dutch consumer press as mentioned in the text

Achieving Service Quality through the Application of
Importance-Perforrnance Analysis
Robert Govers and Frank M. Go
Contents
Abstract 163
Problem definition 163
2 Literature 164
3 Objectives 165
4 Methodology 166
5 Attributes 169
6 Importance 170
7 Survey 173
8 Performance 177
9 Importance-performance analysis 179
10 Comparative Analysis 181
11 Summary and conclusion 182
References
Importance-Performance Analysis 163
Abstract
A sUNey was conducted amongst different types of consumers to measure the mar-
keting effectiveness of the Dutch domestic tourism organisation, The Stichting Toer-
isme en Recreatie AVN. Specifically the purpose of the present study is to investigate
how the AVN might improve its ability to manage consumer expectations and deter-
mine which attributes the organisation should devote its attention and resources to, in
order to improve, meet, or exceed consumer expectations. First, the importance of
attributes is being determined through a management survey. Next, the performance
is measured through multiple consumer surveys. Finally, the differences between
importance and performance are analysed in an effort to identify the organisations
strengths and weaknesses.
1 Problem definition
The present article is based on an empirical study which is being conducted for the
Dutch domestic tourism organisation (DTO), the Dutch Foundation on Tourism and
Recreation AVN (in short AVN). The foundation is a co-operation between the Dutch
Automobile Association (ANWB), the Netherlands Bureau for Tourism (NBT) which
is the Dutch national tourism organisation, and the Tourist Information Offices
(VVV's). The AVN is concerned with the sustainable development of domestic tour-
ism and recreation in the Netherlands. It performs many activities, such as trade
fairs, consumer shows, direct mailings of brochures, travel guides, familiarisation
trips, and media exposure. In order to practice marketing in a responsible manner,
AVN decided to measure, on a regular basis, the effectiveness of its marketing ef-
forts. A marketing effectiveness monitor offers two important benefits:
1. It provides stakeholders a reasonable degree of 'transparency' in !he mar-

keting affairs of !he AVN and therefore may pre-empt many time-consuming
enquiries from its stakeholders;
2. It provides a 'success indicator', which enables AVN to measure, to what
extent their efforts gain results.
164 Robert Govers / Frank M. Go
The bottom line is that domestic tourism organisations (DTO's) should generate
sales for the tourist and recreation industry. Private, but also public organisations
participate in activities, only with the prospect of influencing consumer behaviour in a
way that is beneficial to the organisation. In general such effectiveness studies, also
known as accountability studies, are done on an ad hoc basis where separate activi-
ties are evaluated on their specific objective, such as the amount of distributed mate-
rial on a consumer show (Pizam 1994, pp. 573-581). This type of research however,
only generates quantitative information which offers few, if any indications in terms
of changes in attitude, perceptions or behaviour. In summary, the AVN expressed a
desire for a standard instrument, a monitor which is capable to periodically measure
the overall performance, transcending the activities of individual actors. Therefore,
an instrument should be developed which measures the overall performance of the
AVN on their structural marketing objectives.
2 Literature
The purpose of DTO's is to market a tourist destination. In this way, AVN is respon-
sible for the marketing of domeslic tourism in the Netherlands. When considering the
effectiveness of the marketing activities being deployed, one should first decide on
the distinguishing attributes of the product which the organisation is marketing. What
are the most important characteristics of the service, which are communicated to the
consumer, in the ho pe of raising their awareness of the product and eventually to
increase their interest, which might lead to a purehase? Marketing effectiveness of
an organisation might be measured as the ratio between the extent to which mar-
keting communication efforts, based on these attributes, contribute to the change in
awareness, perceptions and behaviour of consumers (Perdue et al. 1994,
pp. 565-571). The travel and tourism literature refers to destination image research
(Echtner et al. 1993, pp. 3-13), which is concerned, amongst others, with the con-
sumer deciding on a specific holiday destination and distinguishing those service
characteristics that are important to ensure customer satisfaction. The latter will be
achieved when organisations within a particular destination perform in a manner,
which either meets or exceeds consumer expectations.
Therefore, the present study draws on the service quality literature such as Zei-
thaml/Parasuraman/Berry (1990), who state that "the key to good service quality is
meeting or exceeding wh at customers expect from the service" and LeBlanc (1992,
pp. 10-16), who supports the notion of destination image research being a form of
service quality measurement, because both quality and satisfaction are concerned
with the difference between expectations and perceptions. The latter study suggests
a significant relationship between perceived quality and image. At the same time all
aspects on which customers base their perceptions and form expectations are be-
lieved to be part of the image. Conceptually, image has been referred to as a set of
attitudes based upon evaluation of characteristics that are deemed important by
customers (LeBlanc 1992, pp. 10-16). Furthermore, the present study builds on the
importance-performance analysis methodology, which Go/Zang (in press) applied 10
measure the marketing effectiveness of Beijing as an international conference desti-
nation.
3 Objectives
Through the present study, the researchers will attempt to develop a service quality
monitor by creating an instrument designed to measure on a regular basis the DTO's
marketing performance on aspects deemed important to the Dutch domestic holi-
daymaker Ennew/Reed/Binks (1993, pp. 59-70) state that "it is of particular value to
a service provider to be able to produce a consistent measure [... ] (of importance
versus performance) [... ] across service attributes in order to indicate areas of
strength and weakness". This is exactly what DTO's should do, to be able to indicate
just how weil their overall performance, their overall service quality, is perceived by
consumers and where there is room for improvement. Carman (1990, pp. 33-50) de-
fines "overall quality (0), as an attitude, a multidimensional construct composed of
differences between perceptions (P) and expectations (E) [... ] (which we will call
performance) [ ... ]. Written in a linear compensatory, expectancy value formulation,
this definition becomes";
Q = ~ li (Pi - Ei)
where: I is the importance of service attribute i, the sum is over the K service attrib-
utes (eight in this study, as later on will be explained).
Based on this model, the following research objectives were formulated:
• first, those attributes that are relevant to the consumer when deciding on the
Netherlands as their holiday destination, should be identified;
• second, the importance of attributes has to be determined;
• third, while developing the monitor, performance, which is the difference
between expectations and perceptions, should be measured for all important
attributes.
The project described in this paper should yield a monitor, which could be used to do
aseries of performance measures, which enable the DTO to periodically analyse
variances between importance and performance of attributes, in order to evaluate
marketing efforts. Therefore, it was hypothesised that:
• by identifying and emphasising image attributes, that are important to con-

sumers, the overall perception of service quality, and therefore performance
can be improved, which increases marketing effectiveness;
• consumer expectations and perceptions depend on product 'familiarity',
which implies that the former should be analysed on a market segment-by-
market segment basis.
4 Methodology
The measurement of the expectation-perception differences creates some practical
problems. First of all, expectations should be measured before and perceptions after
the service has been delivered. In case of domestic tourism and recreation this is
rat her a difficult challenge in that most consumers will regularly experience or at
least come in contact with the service providers, so that expectations and percep-
tions may be constantly changing and no accurate 'before and after' measure could
be taken. It is not like a service rendered at a hair dresser, where we could stop
Imporlance-Performance Analysis 167
people at the door to complete the expectations questionnaire, and ask them again
to fill in the perceptions battery while leaving (and then, there is still the question of
willingness of consumers to do so). Second, the inclusion of the importance measure
for all the attributes, would result in such an extensive amount of questions, that as-
pects of fatigue and boredom would be likely to become a factor in the interviewing
process.
Carman (1990, pp. 33-55) suggests that expectations change with the extent to
which consumers are familiar with the service rendered, therefore the AVN's total
target group, which is the national public, is divided into the following four groups,
based on the AI DA-model (creating Awareness, Interest, Desire and Action) (Ster
et al. 1987):
Group 1: Those consumers that have no experience with, and therefore are not fa-
miliar with the Netherlands as a holiday destination;
Group 2: Those who are aware, but show no interest in the Netherlands as a holi-
day destination;
Group 3: Those who show interest, but do not take part in any activities;
Group 4: Those who do take part in activities in the Netherlands.
At the same time, AVN's market segments and marketing goals are formulated in the
same way. Therefore, the rationale for such an approach was straightforward. Obvi-
ously expectations will differ between these four groups, as would the attributes' im-
portance. Due to the conceptual and practical consistency of the segmentation, one
would expect that in the absence of a major change in the service delivery, the
measures of expectations within groups would be quiet stable (Carman 1990,
pp. 33-55). Therefore, no separate measures of expectancy were included in the
questionnaires, instead, the perception-expectation difference is directly included in
the formulation of the questions. As Carman (1990) states: "this is likely to be useful
in situations where norms for expectations are weil formulated in the respondents
mind from past experience with similar services". In this case it seems unlikely that
many people would argue that there are consumers who have no opinion about
tourism and recreation at all. Each item in the perceptions battery would be an-
swered on a seven-point Likert scale with the reference to expectations included in
the questioning in the form of the words 'sufficient' or 'enough'. For example, a query
might be posed as folIows: "The Netherlands has enough historical sites and places
of interest". That leaves us with the measurement of the importance of the various
attributes.
As AVN's management has gained a significant amount of experience with marketing

the four target groups, one would expect that they are likely to be knowledgeable
about the types of aspects that are important to the various groups. This could also
be based on prior research. Currently the AVN is making a concerted effort to target
the four segments by closing the gap between expectations and perceptions for
those aspects which are deemed most important for the specific segment (LeBlanc
1992, pp. 10-16). Therefore, discussing the current marketing policy and the way in
which service attributes are marketed should give us valuable insight into the im-
portance of attributes to consumers in the four segments.
In a management session destination image attributes and their relative importance

for the four segments were identified. At the start of the session a Bernstein (1984)
discussion was conducted. AVN participants were asked to mention those attributes
of the "tourism and recreation in the Netherlands" -product, that play an important
role in determining the organisation's promotional activities. In a subsequent discus-
sion, the total number of attributes was reduced to eight essential va lues which all
participants consider important and believe to contribute to the identity of the Neth-
erlands as a holiday destination for the various target segments ..
In the second part of the session the Analytic Hierarchy Process (AHP), a method
developed in the seventies by Saaty (1988), was applied to determine the impor-
tance of the various attributes for the four segments. With the use of a PC and LCD-
panel a hierarchy of goals and alternative actions was composed and relationships
between attributes and target segments interactively determined and prioritised. The
pinnacle of the hierarchy might be viewed as the superordinate goal, in this case:
"the effective promotion of domestic tourism and recreation". Positioned underneath
are the four target segments where this effectiveness should be achieved. The base
of the model consists of the alternatives, in this case the attributes that are possibly
being used for marketing the four segments. Next, a prioritisation within the hierar-
chy will take place, through pair-wise comparison of the alternatives on a specific
level in relation to the above positioned goal. Therefore, the four segments were first
pair-wise compared in relation to the amount in which they are likely to be of interest
in effectively promoting domestic tourism and recreation. Based upon the obtained
scores, a rating of the significance of the target segments was calculated. Subse-
quently, for all four target segments, the attributes are pair-wise compared on their
contribution to the effectiveness of the marketing efforts aimed at the specific seg-
ment. In this way, a prioritisation of attributes was obtained simultaneously. Finally,
the overall importance of attributes was determined by weighing the ratings of attrib-
utes over the four target segments.
5 Attributes
As previously mentioned, identifying those attributes that are important to consumers
when deciding on a holiday destination is done through destination image research.
figure 1 shows the results of three of those studies. The far right column lists the at-
tributes that participants identified during the Bernstein discussion. All attributes that
resulted from the management session seem to correspond with aspects mentioned
in past research. Just the 'opportunities in all seasons' attribute seems to be an odd
one. The attribute refers to the Dutch climate and unfavourable weather conditions,
which makes it unlikely to be mentioned by consumers as a motive. Nevertheless,
AVN management wants to consider it as an important service attribute, because
they argue it could be a strength. Because of the fact that Holland has an unpredict-
able climate, which can be quite unpleasant, there are a lot of bad weather facilities
and things to do when the skies are not blue. That is why there are always possibili-
ties for holiday making, even in fall and winter.
Destination image Destination choice The Netherlands' Image attributes iden-

attributes attributes, based on strengths, based tified through this
(Echtner et al. tourists' perceptions on NBT's destina- study's Bernstein-
1993, pp. 3-13) of European countries tion image re- discussion
(Haahti et al., 1983, search (Stichting
pp. 34-42) AVN, 1995, pp. 14)
accessibility elose by good accessibility
interest/adventure entertainment and night variation in a small variation in activities

life area
facilities for sports bikinglwater sports
natural state a peaceful and quiet water/duneslpolder variation in scenery

holiday
inexpensiveness good value for money inexpensive possibilities for any

budget range
tourist facilitation easy to organise weil organised
resort atmospherel opportunities in all sea-

climate sons
cultural distance cultural experience familiar familiar
comfortlsecurity friendly and hospitable easy to get informa- high service standard
people tion
lack of language
barrier
Fig. 1: Destination image attributes
6 Importance
Based on the previous discussion, the AHP is constructed and represented in fig-
ure 2. It also shows the overall importance of the image attributes and target groups
and each group's contribution to the attributes' overall importance scores. It clearly
shows that AVN management considers group 3 and 4 to be equally important, but
far more important than groups 1 and 2. Group 2 however, is twice as important as
group 1. The AVN obviously concentrates its marketing efforts on those consumers
that are active or at least interested in the Netherlands as a holiday destination. Ap-
parently the AVN pursuits a market cultivation in stead of a market differentiation
strategy.
Fig. 2: The AHP and the overall importance 01 target groups and image attributes and each
groups' contribution to the attributes' importance scores
When the overall importance of the attributes for the whole population is considered ,
'weil organised' is perceived to be the most important attribute according to the AVN.
Less of an issue when marketing the Netherlands as a holiday destination seems to
be 'variation in scenery', followed by 'variation in activities' and 'possibilities for any
budget range'. A high service level and familiarity of the destination are perceived to
be the least important aspects to communicate.
Figure 3 shows the importance of the attributes for the four target groups and the
market as a whole through the use of lines and the right y-axis. The bars combined
with the lett y-axis reflect the relative importance of the four groups. For group 1 all
attributes are depicted as being equally importanl. The reason for this is that during
the management session it was agreed not to go through the pair-wise comparisons
for this group. Group 1 is composed of those consumers that have no experience
with, and are therefore not familiar with the Netherlands as a holiday destination.
AVN management believes that those unaware consumers are mostly people that,
for a variety of reasons, do not go on holiday at all. Because of the diversity within
this group and the fact that those consumers are unaware, it is quite difficult to de-
eide on the relative importance of the attributes. Also due to the relative insignifi-
cance of this group it was decided to equalise the importance of the attributes for
group 1 at 12.5% (100% divided by 8 attributes).
100% 35%
90%
30%
80%
70% 25%
- V.nation in scenery
60% 20% - High ..1Yi<o $land> ..
50% - GoodIOC:.SSlbifty
15% ....... F8mi1i1r

40%
30% 10% •••• Oppoft1JniblS In •• SNSOnJ
20%
5%
10%
Group 1 Group 2 Group 3 Group 4 Overall
Fig . 3: The importance ot destination image attributes tor the tour target groups and the total
population
It appears that there are large differences between the four groups in the perceived
importance of the attributes 'variation in scenery', 'variation in activities' and 'weil
organised'. The first two aspects should be emphasised in the communication to-
wards group 2, as opposed to group 3. Then again group 4 also needs to be told
about the variation in scenery and activities in the Netherlands, which puts those
attributes in an overall second and fourth position. 'Weil organised' is judged as the
major communication aspect to address group 3, but also group 4, whereas it is of
minor importance to group 2. It is argued that the consumers in group 3 are already
aware of the existing advantages of the Netherlands as a holiday destination, such
as the variation in activities and scenery, after all, they are interested. So there
should be some pragmatic reason that keeps them from participating. Communicat-
Ing that it is relatively easy to organise a holiday in the Netherlands, without having
to pay too much, might give them that final prod. The same considerations might
persuade consumers in group 4 to take that extra vacation in their home country.
Due to this and the importance of groups 3 and 4, 'weil organised' takes the lead in
the overall ranking. Obviously, 'possibilities for any budget range' and 'good acces-
sibility' are also important to communicate to group 3. Finally, it is also perceived to
be useful to emphasise to group 4 that there are opportunities for holiday making in
the Netherlands in all seasons of the year.
The importance measures that were obtained through the AHP are now considered
as stable factors in the determination of service quality. In the process of developing
a sustainable effectiveness monitor, a first measurement of expectations-perceptions
differences has been completed. From there the monitor will be refined and new pe-
riodic measures taken in order to track AVN's marketing effectiveness. Repeatedly,
occurring variances between attributes' importance-performance will be analysed.
The aim is to achieve importance scores, which are stable, and the performance
scores, which migrate in a positive direction. In this way strengths and weaknesses
can be identified and determined where progress is being made. Results from the
first monitor sampie will be discussed in the subsequent pages.
7 Survey
A questionnaire containing 73 questions was mailed to 1170 addresses selected
from the AVN customer database. Respondents were assured confidentiality and
asked to complete the questionnaire and to return it in a self-addressed, postage-
paid envelope within three weeks. Up to 500 questionnaires were returned, which
works out to a response rate of 42.7%. 476 valid questionnaires were retained in the
sampie. The first part of the questionnaire contained questions about holiday be-
haviour. The responses to these questions were used to segment the respondents
into the four target groups. For instance, questions were asked with respect to the
number cf times the respondent went on holiday during the last three years, how
many of those were main holidays in the Netherlands and abroad (1 per year, maxi-
mum 3), how many secondary holidays took place in the Netherlands and what the
respondents' plans are for the present year. Based on this, 15 (3.2%) respondents
were classified in group 1, 68 (14.3%) in group 2, 39 (8.2%) in group 3 and 354
(74.4%) in group 4. The findings revealed that most of the respondents are active
tourists in their home country. This is not surprising, because they were drawn from
the AVN customer database. This database consists of people that either requested
information about tourism and recreation in the Netherlands or took part in one of the
activities that are organised by the AVN to increase participation. This sampie was
used to do apre-test of the monitor, because of the relatively low costs. In later
monitor surveys sampies will be drawn from the total Dutch population, which should
give a more representative distribution.
The survey also includes a query concerning with whom the respondents travel. 19
(4.0%) of the respondents travel alone, 104 (21.8%) travel with their partner and is
younger than 55 years old, 116 (24.4%) travel with a partner and is over 55, 117
(24.6%) travel with their family including children under 18, 14 (2.9%) travel with
their adult children, 26 (5.5%) of the respondents travel with friends or relatives in-
cluding children under 18 and 77 (16.2%) travel with friends or relatives who are all
adults. With respect to the type of accommodation which the respondent frequents
most of the time, 164 (34.5%) preferred a hotel or motel, 165 (34.7%) an apartment,
bungalow or room, 46 (9.7%) private accommodation on a permanent location whilst
99 (20.8%) of the respondents preferred to travel with their own tent, caravan or
camper to a temporary location.
The last part of the questionnaire contains statistical questions like area code, age,
gender (52.3% female), education and income. Also included were questions about
the respondents' awareness of the AVN. In order to be able to credit the AVN for any
positive performance in marketing the Netherlands as a holiday destination it has to
be known how familiar the respondent is with the AVN. Again the AIDA model was
used to segment respondents into four groups. 23 (4.8%) respondents indicated that
they had never heard of the AVN, 38 (8.0%) were aware, but not interested, 93
(19.5) indicated that they were interested, but had never made use of any of the
AVN's products and 305 (64.1 %) respondents indicate that they have at least on ce
taken part in an event organised by the AVN or made use of one of its products. It
should be noted that group 4 is relatively overpopulated due to the sampling frame
which was applied, as discussed earlier.
The main body of the questionnaire contains 40 seven-point Likert scales with 5
questions corresponding to each attribute. 3 items contained negative wording in
order to reduce 'yea-'saying tendencies, as Churchill (1979, pp. 64-73) suggests.
But, as Carman (1990, pp. 33-55) states, "in a long questionnaire, many respon-
dents find this change in wording difficult to comprehend, and thus they misread the
item [... ] so that [ ... ] the disadvantages of including reversed items probably outweigh
the advantages". Such was clearly the case in this study where the reversed items
did not correlate with any of the other items, and therefore had to be excluded from
further analyses. One item was eliminated because it contained a double-barrelled
statement (Churchill 1991).
All other items were factor analysed as listed in figure 1. Ta assess reliability Cron-
bach's alpha was calculated for each theoretical factor, and coefficients greater than
.75 were considered satisfactory (Peter 1979, pp. 6-17). Based on this criterion and
factor analyses on all 36 items, the attributes 'good accessibility' and 'weil organ-
ised', together composed of ni ne items, were excluded from further analyses. The
results of the principal components analysis with varimax rotation on the remaining
27 items are listed in figure 4.
Items Factor Factar Eigen- %01 eron-

Loadlng Name value vilr.exp. bach',
(cumulatlve) Alpha
Nt- MI enough places ofinlerest aod historieal sites .71

NL afters enough touristic and recreational attractions and events .54 variation .75
There are sufticient possibilities 10 90 out in Nt .49 in actMties
NL off"'" plonty opportunities 'or blkin9, hlkin9 or water sports .67 10.1 37.6'16 on bIoc.66
NL h•• enough beauliful countryslde .66 (37.6'16)
NL has enough nk:e beaches, lakes and coastline .62
NL has sutricient accommodation in a variety of nice environments .55 and scenery .60
NL offers enough urban beauty .66
NL provid.. suffIClent variety. both countryside a. wella. eille. .72
NL as a holiday spot offers sufflClent opportunnies 'or any budget .82 possibilities
range
In NL the..... hoIiday opportun nies In adequate price braekets .60 tar any 2.4 9.0'16 .90
In NL thn is suffielent entertainment to sun everybodys purse .66 budget (46.6'16)
Also with litUe money you can have a reasonable pleasant holiday in .79 range
NL
Service and quality in the Dutch tourislic seetor is adequate .55
Annually, NL has plenty new opponunities tor vacation and recreation .46 high
HoUday and recreatkm facilities in NL are of a adequate service sian- .72 service 1.3 4.9% .79
dard
The Outch tourist Industry takes the consumers' neecls into consid- .63 standard (51.5'16)
eration
The Dutch tourtsm sector Is coneemed abou1 ns customers .54
There ia enaugh ta da in NL in any season .75
Despe ns cllmate NL offers sulficlent hollday opportunrtles .49 apportunities 1.2 4.3% .77
NL os a ho/iday spot offers adequote bad weather facllities .49 in all seasons (55.8%)
Also in fall or Yt'inter NL has enough to offer as a holiday destination .63
As a holtday destination NL is safe enaugh .32
As a holiday destination NL js familiar enough .38
Gaing on holiday in NL is niee and eosy enough .63 familiar 1.1 3.9'16 .77
The hollday atmosphere In NL Is adequate .63 (59.7'16)
To spend my days off in my home eountry is vacation enough for me .76
Touristic itineraries are generally adequate in NL
The public transport in NL durtng holldays is generally sulficlent good .56
NL haI adequote slgnposling accessibility
Accessibility of touristic destinations in NL is satisfactory
Nl as a holiday spot is weil organised
rhere sre sufflCient touristic facilities as far instance accommodation
Ta get infonnatlon on holidays in NL is easy enough weil organ- .70
lsed
Ta organise a holiday in NL is easy enough
Organisations as the V'N and ANWB offer sufficient help and info
• The Nether1ands is abbreviated by Nl
Fig. 4: Results of factor analysis on 36 items measuring service performance
Factors whose eigenvalue was greater then 1 were retained in the factor solution.
This approach resulted in 5 factors explaining almost 60% of the variance, with the
attributes 'variation in scenery' and 'variation in activities' being joined together.
However, because these attributes' individual Cronbach's alphas are also greater
than .75 they will be maintained as separate attributes in the monitor. The results for
the remaining four attributes, 'possibilities for any budget range', 'high service stan-
dard', 'opportunities in all seasons' and 'familiar', were satisfactory. As for the attrib-
utes 'good accessibility' and 'weil organised', it was decided to do aredesign of the
items involved before purifying the measure through any subsequent surveys.
For each respondent the perceived performance on the attributes is represented by

the geometrie mean of the items relating to the specific attribute (figure 1). An ex-
ception to the latter is the attribute 'good accessibility' which, for purposes of this
paper, is represented by the geometrie mean of the items 'touristic itineraries are
generally adequate in the Netherlands', 'the Netherlands has adequate signposting'
and 'accessibility of touristic destinations in the Netherlands is satisfactory'. Lastly,
Ihe overall performance as perceived by the respondent is represented by Ihe geo-
metrie mean of all attributes.
8 Performance
First it was hypothesised that the overall perceived performance would increase with
the extenl to which Ihe respondenls were familiar with the Netherlands as a holiday
destination. This appeared 10 be true, wilh Ihe following overall performance scores
resulting from the analysis of variance (F = 10.7, P = .000) (1 = lowest performance,
7 =highest performance):
Group 1: 5.59
Group 2: 5.18
Group 3: 5.33
Group 4: 5.74
Not surprisingly, those consumers that are active holidaymakers in the Netherlands
tend to have a more favourable opinion about its service quality compared to non-
participanls. However, Group 1 also gives a relatively high overall performance
score. As discussed earlier, AVN management believes that those consumers who
were unaware tend to be mostly people who, for a variety of reasons, do not go on
178 Robelt Govers / Frank M. Go
holiday at all. This could mean that they might have a positive attitude towards their
home country as a holiday spot, but do not participate tor so me other reason beyond
their control (tor instance invalids, elderly or people with a private business). When
analysing the separate attributes' performance differences between the tour seg-
ments, this general effect had to be extracted trom the attribute scores. Theretore,
individual attribute scores were divided by the overall performance scores. Figure 5
shows these standardised attribute performance scores tor the tour segments. When
an attribute score in this bar chart equals 1 it means that the performance of that
attribute is on par with the overall performance as perceived by that group.
. GroYp I (N. ' 5) • GfOUp 3 (N-.39) o Gtoup '" (N-354)
~~ ~l
0;-
~~E
FO~
a;4i ~..5.
10
Se ~ ::..
~~ i~~
~
~~& U.!!
i'o.-gi Q.~
~=~ :§tl:
i
2'&t.I
~ ~
~. ~
", -
~r
~".
~r
lIl:l5i!c
,,~e
", .. e
~8.fII
"," .
~>
~~~
~~ .
:::I!
... .
' !i~
~~~"'~
1:.0
Fig . 5: Standardised attribute performance scores tor the tour target groups
The most notable observation in figure 5 is the low performance score tor group 2
with respect to the attribute 'familiar'. Because this group consists of consumers that
prefer to go on holiday abroad, it was hypothesised that they might find their home
country to be too familiar to be an attractive holiday destination. A control quest ion
reading "As a holiday destination the Netherlands is too familiar", was added to the
questionnaire and indeed group 2 scored significantly higher than the other seg-
ments in this respect (F =5.34, p =.001). Lastly, price seems to be the main issue
for improving performance, although this is not as much the case for group 4. This
finding supports the nation that AVN should emphasise in its communication to
group 3 that there are holiday possibilities for any budget range in the Netherlands.
Such approach might motivate those consumers in group 3 to participate.
9 Importance-performance analysis
Importance-performance analysis is a useful method in examining the strengths and
weaknesses while assessing service quality. A particular strength of the importance-
performance analysis is its use of a graphical technique in the form of a matrix or
action grid, which is shown in figure 6. It is a potentially valuable tool to present mar-
ket research results in an easy to understand visual display, which provides a basis
for formulating market strategies (Crompton et al. 1985, pp. 69-80).
As explained by Go/Govers (1997, pp. 57-65), the action grid is first divided into four
quadrants through the subjective hands of the researcher. A commonly used ap-
proach is to calculate the overall me an performance and importance, in this case 5.6
and .125 (if all attributes are equally important, 100% is divided by 8, wh ich equals
12.5%). Those figures are then used to split both dimensions and make sure that
attributes are equally divided into more and less important attributes and spread
over the four quadrants. The top right quadrant 11 hauses the major strengths, that is
those attributes which contribute to above average performance from the consumers'
viewpoint. Martilla/James (1977, pp. 77-79) refer to this quadrant as the one that
suggests to "keep up the good work". In quadrant I performance ranks high on at-
tributes that are not very important to the consumer. This might suggest possible
'overkill' in the implementation of strategy. Resources that are committed to these
attributes could perhaps be more effectively allocated elsewhere. The lower left
quadrant 111 consists of attributes that are also less important, but are performing
below average as weil. Marketers should monitor these attributes, but not necessar-
ily allocate more marketing resources to counter such performance. Instead, to in-
crease perceived service quality, they should concentrate the major marketing ef-
forts on the attributes in quadrant IV, which are the major weaknesses. These attrib-
utes are considered to be very important, but are perceived by the consumer as ar-
eas where performance is below average.
Total population
.,
I n
. gOOCS1eQtUibilitt
lIC
p
• • •
. w.' organl. .d
,. Mgh • . , . standatd
X pouibiiih. foranybud".tran".
'"
... x v.n.1on In ,cHI. .
•. . ,...onln.~,..,.,
. . oppor1Uniliea in.1 '.II'on.

x
m IV
,125 0.25
Importanc l
Fig. 6: Importance-performance matrix tor the total population
Figure 6 clearly shows that the major weakness in marketing the Netherlands as a
holiday destination is price. Although 'possibilities for any budget range' is an im-
portant attribute, it received the lowest performance score. Apparently the Nether-
lands is not perceived as adestination where anybody, with any purse, could have a
holiday. Another attribute which seems to perform less favourable is 'opportunities in
all seasons', which suggests that the bad weather facilities and off season attrac-
tions and events may not significantly increase the perceived service quality of the
Netherlands as a holiday destination. On the other hand, the variation in scenery
and activities offered does seem to increase customer satisfaction. Areal strength,
which appears at the same time to be the most important attribute, is the way in
which the tourist industry in the Netherlands is organised. Although this service at-
tribute does underpin quality, there is room for improvement in which the AVN could
make a reasonable contribution.
10 Comparative Analysis
As a comparative analysis an importance-performance matrix is constructed for all
four segments as represented in Figure 7. This gives more detailed information on
what aspects need most attention when marketing the Netherlands as a holiday des-
tination. Clearly price is a major issue especially for groups 3 and 2. Of course
group 4, representing those already participating in domestic tourism, appears to be
satisfied about all attributes with the exception of the 'opportunities in all seasons'
attribute. Marketing this aspect as weil as price could increase participation.
Group1 Gro~2
I· . . ". . ., ,
IN-15) IN-MI
~
••
t7
.ii 5 IA
• welorglnlsed
hlghseMCIS1lndBrd ~ 5
ii
•
E • pOSllOllitlesforany E
budgelrwtg'
~
Ir. t7
•
V_bonll'C1M~1I5
varlaUonll'lscenety
..~ •
o land ..
e
. 0 0125
ImpOftance
E:I opportunrtllSnd
...sons
•0 O.12i!1
Importance
Group3 Gro~4
tN- 31) IN-l54)
t7
· ".. "....."" ,I •
t7
••
• • • • welorgltllsed
~ 56 • hlgtlseMCIllandll'd,
I
5.' a"
E • possitlilrbnlorany
budgellWlgl I e
.~ •
Oe
'0'
•
venllllOtlll'lI.;tl/ltl151
..,."IIIOftmSCen-v I •
I
10 farn". !
je I
.
qlpOltunl,esln'"
...san •
I
0 Q.125 0'"
Impartance Irnportance
Fig.7: Importance-performance matrices per segment
11 Summary and conclusion

The findings of this study support the notion that consumer expectations and per-
ceptions depend on product familiarity. As hypothesised, performance increases with
the extent to which consumers are familiar with the service. This is obvious because
one would expect somebody that uses a service to be more positive than somebody
who does not. Therefore, to monitor the effectiveness of marketing the Netherlands
as a holiday destination, performance is periodically measured on a segment-by-
segment basis. The need to do so is best illustrated by means of the results con-
cerning the image attribute 'a familiar destination'. Performance scores on this at-
tribute are significantly lower for group 2 than for the other groups. It is assumed that
this group consists of consumers that are aware of the Netherlands as a holiday
destination, but show no interest because they prefer going on holiday abroad. AI-
though once in a while they might spend their days off, maybe as a secondary holi-
day in the fatherland, they are not likely to become 'heavy users'. The hypothesis
that this group might find their home country to be too familiar to spend a holiday,
appeared to be true. Therefore, marketing the Netherlands as a safe, nice and cosy,
familiar destination might be effective when addressing those consumers that are
al ready participating in tourist and recreational activities in the Netherlands. How-
ever, using this same attribute to address group 2, would have the reversed effect.
This implies that to increase marketing effectiveness, those attributes that are im-
portant to a specific segment should be emphasised in the communication towards
that segment. Organisations could improve their perceived service quality and
therefore their overall performance, by paying special attention to the attributes that
are of significant importance to each of the target groups.
With the use of the attributes' importance scores, that were determined through a
management survey, strengths and weaknesses were identified using importance-
performance analysis. It appears that the Netherlands major weakness is its price
level. Especially by groups 2 and 3, it is not perceived as adestination that suits
everybody's purse. However, this negative opinion concerning price, is less sup-
ported by group 4 and, because there is no proof that this group comes from a
higher social class, this implies that, once consumers have more experience with
holiday making in the Netherlands, this aspect appears not to be as bad as it seems.
Therefore, by keeping prices stable and communicating this to groups 2 and 3 it
might motivate some of those consumers to go on holiday in the Netherlands. An-
other aspect, which could improve perceived service quality when marketed effec-
tively, is 'opportunities in all seasons'. Emphasising that there are possibilities for
holiday making in the Netherlands all year round could especially increase participa-
tion by consumers in group 4. Finally, the Netherlands major strength is the way in
which its tourism industry is organised. Emphasising this to group 3 could also ex-
pa nd domestic tourism.
If performance and therefore perceived service quality can indeed be improved by

identifying and emphasising service attributes that are important to consumers, will
have to be determined through future monitor sampies. This first survey, which has
been of an exploratory nature, gives an indication about the present situation and
the justification of the methodology used. The study indicates that, in situations
where most consumers have had experience with the specific or a similar service,
performance could be measured directly, on a segment-by-segment basis. However,
one of the disadvantages of the methodology applied here is that, the differences in
performance scores between groups, although statistically significant, are small as
compared to the differences in importance, which are determined through the use of
the AHP. Therefore the comparative analysis using importance-performance analy-
sis does not produce evident differences in the interpretation of the matrices for the
various segments. The divergent appearances are mainly created by the differences
in importance scores, not the variation in perceived performance. This is probably
caused by the use of self-stated performance evaluation using Likert scales. The
results might be more evident when some sort of paired comparison would be used
to evaluate performance as weil. This could be an area of attention in future re-
search.
184 Robelt Govers / Frank M. Go
References
Bernstein, D. (1984): Company Image and Reality, Eastbourne.
Carman, J.M. (1990): Consumer Perceptions of Service Quality: An Assessment of the

SERVQUAL Dimensions, in: Journal of Retailing, Vol. 66, No. 1, pp. 33-55.
Churchill, GA Jr. (1979): A Paradigm for Developing Better Measures of Marketing

Constructs, in: Journal of Marketing Research, Vol. 16, No. 1, pp. 64-73.
Churchill, GA Jr. (1991): Marketing Research: Methodological Foundations, 5th edi-

tion, Orlando.
Crompton, J.L.lDuray, NA (1985): An Investigation of the Relative Efficacy of Four

Alternative Approaches to Importance-Performance Analysis, in: Journal of the
Academy of Marketing Science, Vol. 13, No. 4, pp. 69-80.
Echtner, C.M.lRitchie, J.RB. (1993): The Measurement of Destination Image: An Em-

pirical Assessment, in: Journal ofTravel Research, Spring, pp. 3-13.
Ennew, C.T.lReed, G.V.lBinks, M.R (1993): Importance-Performance Analysis and the

Measurement of Service Quality, in: European Journal of Marketing, Vol. 27, No. 2,
pp. 59-70.
Go, F.M.lGovers, R (1997): The Asian Perspective: Which International Conference

Destinations in Asia are the Most Competitive?, in: CEMS Business Review, Vol. 2,
NO.1, pp. 57-65.
Go, F.M./Zang, W. (in press): Applying Importance-Performance Analysis to Beijing as

an International Meeting Destination, in: Journal ofTravel Research.
Haahti, A.Navas, U. (1983): Tourists' Perceptions of Finland and Selected European

Countries as Travel Destination, in: European Journal of Marketing, Vol. 17, No. 2,
pp. 34-42.
LeBlanc, G. (1992): Factors Affecting Customer Evaluation of Service Quality in Travel

Agencies: An Investigation of Customer Perceptions, in: Journal of Travel Re-
search, Spring, pp. 10-16.
Martilla, J.A.lJames, J.C. (1977): Importance-Performance Analysis, in: Journal of Mar-

keting, Vol. 41, No. 1, pp. 77-79.
Perdue, RR/Pitegoff, B.E. (1994): Methods of Accountability Research for Destination

Marketing, in: Brent, R.J.R.lGoeldner, C.R. (eds.): Travel, Tourism, and Hospitality
Research: a handbook for managers and researchers, 2e<1 edition, New York,
pp. 565-571.
Peter, J.P. (1979): Reliability: A Review of Psychometrie Basics and Recent Marketing
Practices, in: Journal of Marketing Research, Vol. 16, No. 1, pp. 6-17.
Pizam, A. (1994): Evaluating the Effectiveness of Travel Trade Shows and Other Tour-
ism Sales-Promotion Techniques, in: Srent, R.J.R.lGoeldner, C.R. (eds.): Travel,
Tourism, and Hospitality Research: a handbook for managers and researchers,
2nd edition, New YOrk, pp. 573-581.
Saaty, T.L. (1988): Decision Making for Leaders: the analytical hierarchy process for
decisions in a complex world, Pittsburgh.
Ster, Van der & Van Wissen (1987): Marketing & Detailhandel, 4th edition, Groningen.
Stiehting Toerisme & Reereatie AVN (1995): Middellange Termijn Plan 1996-1998,
Voorschoten, pp. 14.
Zeitharni, V.AJParasuraman AJBerry, L.L. (1990): Delivering Quality Service: Salane-

ing Customer Perceptions and Expectations, New York.
The Sequentiallncident Technique for Innovations (SITI) -
A Tool for Generating Improvements and Ideas in Service Processes
Alexander Decker and Harald Meissner
Contents
Abstract 189
Introduction 189
2 Requirements of the Research Tool from a Theoretical Point of

~w 1~
3 The Sequentiallncident Technique for Innovations (SITI) 195
3.1 From SITto SITI - the Enhancement of a Research Technique 196
3.2 Data Collection 197

3.2.1 Focus Group Interviews 197
3.2.2 Individual Interviews 198
3.2.2.1 Individual Interviews with Service Customers 198
3.2.2.2 Individual Interviews with Members of the

Organization 199
3.3 Data Analysis by Classification and Interpretation 200
4 Generating Innovations by Applying SITI in the Car Service

Industry 201
4.1 Research Design 201
4.1.1 Focus Group Interviews 202
4.1.2 Individual Interviews 203
4.1.2.1 Individual Interviews with Service Customers 203
4.1.2.2 Individual Interviews with Members of the

Organization 203
188 Alexander Decker / Harald Meissner
4.2 Data Analysis and General Findings 204

4.2.1 Customer Paths from Focus Group Interviews 204
4.2.2 Reported Incidents 205
4.3 Innovations 208
4.3.1 Analysis and Systematization of Proposed Innovations 208
4.3.2 Development of Concepts for Service Innovations 211
5 Implications, Limitations and Hypotheses 213

5.1 Implications for Future Use and Research 213
5.2 Limitations of the Tool and the Study 216
5.3 Hypotheses 217
6 Summary 219
References
Appendix
The Sequentiallncident Technique for Innovations (SI TI) 189
Abstract
Service innovations are crucial for service marketers to differentiate from competitors
and to meet the expanding needs of the focused market segments. Nevertheless,
service innovations are usually generated in a haphazard way. There is a lack of ap-
p/ication of market research techniques in this area as weil. Therefore, the purpose
of the paper is to show the development and app/ication of a research tool, which
should help generating service innovations systematically.
In order to fulfill theoretically deduced requirements regarding the special character-

istics of services and service qua/ity, the Sequentiallncident Technique was modified
into the Sequential Incident Technique for Innovations (SITI). This paper describes
the method in detail and gives deeper insights into the results of its first app/ication.
Study findings suggest that especially negative, but also positive incidents can trigger
innovative suggestions from the customers' side. The development of coherent
service innovation concepts from these suggestions, however, remains left to the
research team. Imp/ications for the future use of the sm as a research tool as weil
as /imitations of the study are discussed.
1 Introduction
There is consensus in scientific and practical discussions that service quality can be
considered as a key success factor for service marketers. As intense competition,
new technologies and dynamic changes are occurring in the markelplace for most
service induslries, there is a need to imprave and re-design services or even de-
velop service innovations to fulfill the various and expanding needs of the focused
market-segments in order to differentiate from competitors.
Unfortunately, it seems to be the case that new services "just come about" (Mar-
tin/Horne 1993, p. 49) or "happen" (Rathmell 1974, p. 14), instead of being devel-
oped with highly sophisticated management tools (Martin/Horne 1993, p.49). For
that, a new service failure rate of 80 per cent, reported in the financial service indus-
try for instance, is not very surprising (Clancy/Shulman 1991, quoted fram Flynn-
Reinecke/Goldsmith, 1993, p. 97). But as Scheuing/Johnson (1989, p. 28) stated,
"successful new services rarely emerge by mere happenstance". That is why service
literature points out the importance of a systematic development of service improve-
ments and service innovations (see e.g. Scheuing/Johnson 1989).
Theories of and research in innovation are mainly based upon new products
(Gadrey/GalloujlWeinstein 1994, p.4). Nevertheless, some of the conceptual in-
sights, that have emerged from studies of new product development also apply to
new services (de Brentani 1989, p. 240). But as services have specific characteris-
tics and differ from physical products, a transfer or adoption of this knowledge into
the service context is often less adequate. As a consequence, a different type of so-
lution is required.
Therefore, the area of new service development has gained more attention (Ed-
vardsson/Haglund/Mattsson 1996, p. 1). One major effort of these research activi-
ties' can be seen in the proposed models of service development (see e.g. Bowers
1989, rendered in fig. 1, or Scheuing/Johnson 1989), which help service marketers
to develop new and successful services systematically (Bowers 1989, p. 16).
Whereas the process as a whole pro pos es a systematic approach, this cannot be
confirmed for the stage of idea generation (see figure 1, stage 3), which is the focus
of this paper. Following de Brentani (1989, p. 239 w.f.r.), ideas for new services are
often generated in a haphazard fashion Usually, service customers are plainly
asked wh at more could a service company do or how could they do it better. There-
fore, answers are mostly limited in number as weil as in the value of their conten!.
One major reason is that in this context marketing research techniques find only lim-
ited use (Scheuing/Johnson 1989, p. 33).
An oveNiew 01 previous research in this area can be found in Wilhelmsson 1994.

r Development of a Business StrategyI

+
I Development of a New Service Strategy I
+ Object
I Idea Generation ·t.~ I~ ofthe
.. paper
I Concept Development and Evaluation
..
I
I Business Analysis
..
I
r Service Development and Evaluation
.. I
I Market Testing
..
I
I Commercialization I
Fig. 1: A normative model of new service development
Source: BowelS (1989, p. 18)
Consequently, the topic of this paper is to make a contribution to this research field
by presenting a tool, which helps service providers to improve service quality and
generate service innovations systematically at an early stage of the service devel-
opment process. As mentioned above, such a tool must take the specific character-
istics of services and service quality into account. The following section (chapter 2)
gives a brief review of these aspects and shows the required components from a
theoretical point of view. Subsequently, the Sequential Incident Technique for Inno-
vations (SITI) will be presented, first at the current state of methodological discus-
sion (chapter 3), then on the basis of an exploratory study conducted in the car
service industry (chapter 4). Eventually, implications and limitations for service pro-
viders and researchers will be discussed (chapter 5).
2 Requirements of the research tool from a

theoretical point of view
In the following we will use the term "service innovation" in a wide sense. Besides
new service ideas, service innovations are understood to encompass also substan-
tial improvements of basically existing services. Service innovations can refer to
specific service transactions or even to the service process as a whole. Moreover,
service innovations have to focus on superior service quality, with service quality
being understood from the customers' perspective as the congruence or difference
of service consumers' perceptions of the service and their expectations towards it
(Zeithaml/Berry/Parasuraman 1988, p. 36).
As stated above, the specific characteristics of services and service quality have to
be considered while developing a research tool for service innovations. Among
these are
• intangibility, as one of the !wo most important constitutive elements of serv-

ices,
• the other being customer integration, (de Brentani 1989, p.240 w.f.r.;
StausslWeinlich 1996, p. 49). From the Ialter, which stands in the fore-
ground of the subsequent considerations, !wo more characteristics of serv-
ices can be derived:
• the process character of the service production and
• the episodic nature of the customer's service experience (Stauss/Seidel
1995, p. 187) as distinctive features of service quality.
In the following, these components will be briefly introduced from a theoretical point
of view. Based on this theoretical background, requirements for a tool, which is to
help service providers to generate service innovations in a more systematic and
therefore better way, are developed. Finally, the consequence for the development
of the research tool will be derived.
The Sequentiallncident Technique for Innovations (SITI) 193
Theoretical background
One major constitutive element of services is Intangibility. That means that custom-
ers are not able to see, taste, touch, smell, or hear the service in the same manner
in which goods can be sensed (Zeithaml 1991, p. 41).
The other of the two most important constitutive elements of services is customer
integration, which means that the customer participates in the service process, or
goods or information from a customer become apart of the service process, respec-
tively (see e.g. MeyerlWesterbarkey 1995). Customer integration in the service pro-
duction process implies that services are delivered and consumed in interaction
between customer and service provider. With a rising number of interactions, the
process character of service production, as one distinctive feature of service
quality, becomes more obvious. Consequently, all interactions within one service
transaction, as a cuslomer perceives Ihem, comprise Ihe so-called customer process
or service process (Stauss/SeideI1995, p. 185; StausslWeinlich 1997, p. 38), wh ich
normally includes several episodes (LiljanderlSlrandvik 1995). Thai is why services
are perceived as processes and are also stored, to a large extent, as episodes, too
(Stauss 1993, p. 409). This implies another distinctive feature of service qualily, the
episodic nature of the customer's service experience.
According to the goal of the paper, a research 1001 for the idea generation in serv-
ices has to consider these specific characteristics, so substantial requiremenls can
be derived from them. The next part of this chapter deals wilh these matters.
Requirements
Since the service innovations sought are meant to lead 10 superior qualily, and since
service quality always has to be seen in terms of customer-perceived quality (Stauss
1993, p. 408), there is a need to focus on customer feedback. Usually, customer in-
formation is collected in satisfaction and service quality surveys, or through com-
plaint analysis. For the purpose of idea generation in services, this is not sufficient.
Through customer integration, customers become apart of the service process.
They gather special knowledge based mostlyon their actual perceplions. Therefore,
they should be integrated into the idea generation process as weil, so that their po-
tential can be better utilized in new service development. Additionally, it is possible
to gain deeper insights into the customers' perception, which is expressed in the
customer process.
As a consequence of both customer integration and the process character of the

service experience, the customer has not only to answer satisfaction questions, but
also has to express individual requirements, make contributions and bring in her or
his capabilities, as we know this from new product development (e.g. Baiietti/Litva
1995, p. 3). "More and better market inputs is the number one factor in new service
success" (de Brentani/Cooper 1992, p. 239). Customer-focused market studies early
in the new service development process pinpoint customer problems, needs, wants,
desires, requirements, and operating systems. Such studies provide the input to
service process design and help to ensure the service product I market fit at the out-
set of the project (ibid.).
The service encounter involves at least two people (external customer and frontline
employee as internal customer), and it is important to understand services from mul-
tiple perspectives (Bitner/Booms/Mohr 1994, p.95). Therefore, Harvey (1992,
quoted by Wilhelmsson 1994, p. 7) also directs attention to internal customers in
new service development. As the frontline employees have concrete knowledge of
the internal processes and the underlying problems on the one side, and are familiar
with the needs of consumers on the other side, this source of information can be an
asset to the process of service idea generation. However, there is a limitation for the
integration of this source. Surveys among frontline employees will not ren der the
process character of service experiences as required, because they are not able to
simulate the .voice of the customer" in terms of the customer process adequately.
As shown above, the episodic nature of the service experience must also be
taken into consideration, tao. For this aim, incident-based methods such as the Criti-
cal Incident Technique (see e.g. Bitner/BoomslTetreault 1990) seem to be more
suitable than the traditional, standardized, and attribute-based quality and satisfac-
ti on surveys (Stauss 1993, p. 409). Attribute-based methods can record only to what
extent the episodes are translated into attitude or satisfaction scores, not the stored
The Sequential Incident Technique 'or Innovations (SITI) 195
episodes themselves. The service quality items are necessarily formulated in an ab-
stract manner. As a consequence, survey results are not particularly concrete (ibid.;
Stauss/Hentschel 1992, p. 62). Since the goal of the research is to develop an tool
to help generate service ideas at an early stage of a service development process,
there is no need for highly standardized data supplied through attribute-based
measurement methods.
As a result of intangibility, services have to be visualized to help consumers ex-

press their needs and reduce complexity in order to generate ideas. A research
method for the generation of service ideas therefore should encompass a tool to
support transparency, such as the service blueprinting (Shostack 1984, 1987) for an
inside-out perspective or the service mapping (Kingman-Brundage 1992; Gummes-
son/Kingman-Brundage 1992) for an outside-in perspective as required through
customer integration. One of the major purposes of service mapping is the system-
atic analysis of processes within a service by recording the course of a typical serv-
ice process from the customer's point of view (StausslWeinlich 1997, p. 39). In addi-
tion, service mapping can be seen as a helpful instrument for a systematic research
procedure, too.
3 The Sequentiallncident Technique for

Innovations (SITI)
An existing research instrument, which combines most of the mentioned require-
ments, is the Sequential Incident Technique (see e.g. StausslWeinlich 1997). AI-
though the results of such an incident-based tool to measure service quality provide
service marketers with useful and concrete information to improve services, there is
no evidence for customer-induced service innovations as required through customer
integration. For that reason, the method had to be modified into the Sequential Inci-
dent Technique for Innovations (SITI).
The SIT will be briefly introduced in the following chapter before the steps of modifi-
cation into the SITI, according to the developed requirements, will be described. The
SITI will then be presented at the current state of methodological discussion.
3.1 From SIT to SITI . the enhancement of a

research technique
The fundamental purpose of the SIT is to record all of the incidents customers per-
ceive in a specific service transaction sequentially in the course of the consumption
process (StausslWeinlich 1997, p. 37). The SIT combines the elements of two qual-
ity measurement and planning methods frequently applied in the service sector:
• Service Mapping and

• the Criticallncident Technique (Stauss 1992, p. 46).
Usually, data collection by means of the Sequential Incident Technique encom-

passes four steps (see StausslWeinlich 1996, p. 50): First, the steps of the customer
process ("customer path") are identified; second, following this customer path, serv-
ice customers are asked to report incidents. Third, the reported incidents are as-
sessed, and finally, superordinate levels of satisfaction (e.g. overall satisfaction) with
the service are assessed.
This paper proposes omitting step four and enhancing step three in order to gather
more of the available information about customers' innovative ideas. This can be
done by simply asking how their service experience of each particular phase of the
service process might have been improved (step 4 of the SITI).
So, the SITI follows the following steps:
1) Identification of all process sleps of the service as gone through by Ihe cus-
lomer ("customer path"),
2) collection of (critical and usual) incidents on the basis of the identified path,
3) assessment of the reported incidents, and
4) generation and collection of suggestions and ideas relating to the reported
incidenls, how service quality can be improved from the customer's point of
view.
This procedure is based on the presumption that customers gain a higher awareness
of innovation potentials through the sequential approach and the narration of the
(positive or negative) experiences they had and, as a consequence, are able to de-
velop ideas from their point of view more easily.
3.2 Data collection

Data can be collected by group interviews or in individual interviews with respon-
dents of different provenience.
3.2.1 Focus group interviews

Since the customer path is aprerequisite for the subsequent steps, it is desirable to
set it up within a limited share of the total running time of the study, while still en-
compassing a multitude of service experiences and views of the service process. So,
especially for this first step of the SITI, focus group interviews are an appropriate
technique, but can also be applied for the following steps. They are discussion fo-
rums of about seven to ten discussants, led by a moderator, which allow to obtain
insights into the views of special interest groups on a certain subject within limited
time and with low interviewer bias. The dynamics of the group situation help to raise
a broad variety of aspects which would rather not be mentioned in individual inter-
views (Krueger 1994, pp. 16 ff.; Morgan 1994, pp. 15 ff.; Churchill1995, pp. 153 ff.).
Dur goal was to obtain a broad variety of aspects, not representativity. Conse-
quently, a deviation from the usual suggestions towards homogeneity of group mem-
bers (ChurchiIl1995, p. 155) seems appropriate. Instead, the participants of a focus
group interview ought to represent a heterogeneous mixture of the service com-
pany's customers. Heterogeneity applies here to demographie features such as age,
gender or profession as weil as to characteristics directly related to the service or
the service company, such as the duration of the customer relationship or the num-
ber of service processes experienced.
First, after an introduction and explanation of the targets of the study, customers first
were asked one by one to think of their latest service experience with the respective
service company and to name the subsequent steps of the service process as they
had experienced it (as proposed through step 1 of the SITI). The term of ,service
process" here is to be applied in its widest sense: Narrators should be encouraged

not only to tell about the period between their first contact with a service employee
(e.g. arriving at the service desk) to the last (e.g. leaving) but to think of earlier and
later steps, which, in their perception, are related to their own service experience.
Afterwards, the multitude of narrated processes ought to be consolidated to one or
more typical customer paths. These customer paths form the basis of the following
survey in individual interviews with customers and employees.
Second, based on the customer path as set up in the session, customers are asked
to report a coherent service experience they had recently and to tell extensively what
happened (corresponding to step 2 of the modified method). The intention here is
not so much to collect narrations of single incidents, but rather, on the one hand, to
check the path for appropriateness, completeness, and consistency and, on the
other hand, to make the participants discuss clear, vivid examples of good or bad
service in its different stages.
Finally, after these incidents have been discussed in the group, process innovations
are generated by means of brainstorming. This technique builds not only on the im-
mediate articulation of ideas, irrespective of their feasibility, it also comprises seizing
other group members' ideas and developing them by modification or addition (Os-
born 1963). This represents step 4 of the SITI. We decided to avoid an interruption
of the dynamic interaction among the participants and did not run through step 3 with
the focus groups.
In order to coHect a greater number of innovations, subsequent to the focus group

interviews, we suggest to conduct individual interviews with service customers and
employees.
3.2.2 Individual interviews

3.2.2.1 Individual interviews with service customers
Aided by the presentation of a copy of the customer path generated in step 1 of the
method through the focus groups, individual customers are asked to recaH the sepa-
rate phases of a service experience and tell extensively wh at exactly happened
(step 2). In case an incident is reported 10 have led 10 cuslomer salisfaclion or dis-
satisfaclion, Ihe interviewer ought to trace it back to a specific element of the inci-
dent by further non-suggestive questions. It is a helpful, even if not a vital part of the
quest for innovations, to ask for a quantitative assessment of the degree of
(dis-)satisfaction on a rating scale (step 3) (see fig. 2 as an example).
A = Anger H = Happiness
l1li111111
MI
Fig. 2: Rating scale for the quantitative assessment of the degree of (dis-)satisfaction
Subsequent to each reported incident, respondents were asked to go through step 4

of the SITI and therefore tell,
• what could have been done beUer in cases when they had experienced
service failures or poor service quality,
• how their service satisfaction might have been increased even in cases they
had assessed as positive, and/or
• which improvements or innovations to the service process they would sug-
gest.
3.2.2.2 Individual interviews with members of the organization
As shown above, a search for innovative ideas would be incomplete, if no use was
made of the potential that lies in the internal process knowledge of the members of
the organization, e.g. the service staft. Similar interviews as with customers can be
conducted with them, especially with frontline employees, who regularly have con-
tact with customers in various stages of the service process.
Still, it does not seem appropriate to let frontline employees go through all the steps
of the SITI, as mentioned above. On the basis of the customer path generated in the
focus groups, each service worker first has to identify the phases of the process, in
which he or she has direct contact with customers. The following step 2 (collection of
(critical and usual) incidents) concentrates on these phases only. OmiUing step 3,
employees are asked to express their ideas according to the respective phase and in
addition to analyze unpleasant customer contacts and the causes of problems and to
develop how the overall service process could be modified (step 4).
Apart from additional input about strengths and weaknesses of the service process,
their active and early involvement in the innovation process may augment their ac-
ceptance of process modifications due to the findings of the study.
3.3 Data analysis by classification and

interpretation
The collected data comprises, on the one hand, a multitude of incident stories and,
on the other hand, a broad range of innovations differing in the problems addressed,
in extent and in the organizational consequences that their implementation would
entail. Data analysis can be carried out by applying content analysis, which is espe-
cially appropriate for qualitative data.
For the purpose of our paper, incidents were first separated from innovations to
analyze them separately. For the qualitative analysis of the incidents, several ap-
proaches of classification, as a major part of content analysis, can be thought of A
deductive classification according to the phases in which they occurred seems ap-
propriate to systematize the collected data in a first step. Another deductive ap-
proach of assigning customer perceptions of a service process to categories is to
make use of the five-dimensional scheme developed by Parasura-
man/ZeithamliBerry (1988). These approaches may be replaced or completed by
other deductive or inductive approaches to be derived from the actual problem field
and research focus.
Whatever approach for classifying the data is chosen, the analysis of the qualitative
data must not be limited to a quantitative analysis of the formed classes or catego-
ries. The most valuable information lies in the qualitative data itself; categorization is
only a necessary and essential step for identifying incidents that are relevant for an-
swering some defined research questions. As the focus of this paper is on innova-
tions, we must refer the reader to Kassarjian (1977) or Viney (1983 w.f.r.) for details
of the content analysis of reported incidents.
For the systematization of the generated innovations, the same approach es as men-
tioned above can be thought of. A deductive approach may be based e.g. on the
phase at which the innovation will be implemented. Other possibilities are to refer to
the superordinate problems or dimensions of service quality addressed or to the in-
struments used. Still, also an inductive way of proceeding as in usual creativity tech-
niques may seem appropriate.
4 Generating Innovations by applying SITI in

the Car Service Industry
4.1 Research design
During the summer term of 1996, the SITI was developed and applied within a re-
search project for a major German car manufacturer. Seeking for innovations in the
maintenance and repair services of its widespread network of dealers, who usually
also run arepair garage, the research objective was to find answers to the following
questions:
• Haw da es the custamer perceive the service process at her/his car dealer?
• Which incidents (positive/negative) do occur to the customer during this pro-
cess?
• Which innovative suggestions for improvement can be collected from cus-
tomers in order to augment customer satisfaction?
Since the research field was widely unstructured, an explorative research design
was chosen. Three car dealers in southern Bavaria were selected by the regional
distribution center of the mentioned car manufacturer. This selection was based on
the following guidelines: Each of them was to have al least 30 repairs a day and at
least 5 fronlline employees; moreover, one of them was to be situated in a city, one
in a middle-sized town and one in the country.
As proposed in 3.2, data were collected in focus groups and individual interviews
with service customers and employees.
4.1.1 Focus group interviews

In a first step, 25 customers were invited 10 lake part in three focus groups (8/8/9 at
each dealer). Due 10 unforeseen cancellations, the required heterogeneity of the
available focus group members was limited. In one group, all of Ihe members turned
oul 10 have been regular customers for years or even decades. Anolher group had
been composed by making use of the dealers database, so that the group's diversity
not only regarding age and gender but also car category and relationship lenglh
could be assured.
The focus groups were basically conducted in the way described in 3.2.1. The mod-
erator was assisted by two recording assistants, who put down the keywords of a
described process on posterboard and displayed them on Ihe wall for visualization.
After a description was completed, reassuring questions made sure that these lerms
and Iheir order complied with the process steps as perceived by the respective cus-
lomer. A minor deviation from the melhod set out above was necessary when focus
group members confused rendering the abslract order of subsequent phases with
Ihe narration of their personal experience in one concrete service encounter in thai
phase. In this case, the reported incidenl was welcomed but afterwards assigned to
a more generally expressed phase of the service process.
The service processes were aggregaled 10 typical customer palhs for each dealer,
forming the basis for the next research sie ps.
The Sequentiallncident Technique 'or Innovations (SI TI) 203
4.1.2 Individual interviews

4.1.2.1 Individual interviews with service customers
At each of the three cer dealers, about 40 individual interviews with customers were
conducted. Due to time limitations and restricted availability of a customer database,
the selection of respondents had to be a convenience sampie. Customers ap-
proaching or leaving the service desk were asked for an interview. Introductory se-
lections made sure that interviewees had gone through the complete service process
at least once and used their car not only for business purposes. The interviews were
recorded on tape in order to facilitate full transcription prior to analysis.
During the interview, respondents were shown the prepared customer path for their
particular cer dealer, then were asked which of these phases applied to their recent
service experience and in which order. Moreover, they were explicitly asked to add
further steps which in their opinion were missing on the flowchart. Afterwards, the
reported incidents of the service process were recorded step by step, as described
in 3.2.2.1. A nine-point rating scale was used for the assessment of the single inci-
dents. The answering options offered four grades for each of the mostly emotional
items "happiness" and "anger" and one for neutrality.
After the narration of each incident, customers were asked for innovative ideas as
set out above in 3.2.2.1. In order to facilitate the finding of real innovations, respon-
dents were explicitly encouraged to think of other service industries, where similar
situations are handled in a beUer or different way or even by different processes.
Finally, some personal data were collected to characterize the structure of the sam-
pie of interviewees in terms of demography, customer group and relation towards the
car dealer.
4.1.2.2 Individual interviews with members ofthe organization

In a third step, the perception of five service workers of each car dealer was also
surveyed in individual interviews. The questioning basically followed the same pat-
tern as the individual customer interviews. First, the interviewees were asked to
identify the phases relevant for them, i.e. the phases of the service process, in which
they had contact with customers. Then, for each of these phases, incidents were
reported. Instead of an assessment of the reported incidents on ascale, service em-
ployees were asked to analyze unpleasant customer contacts and the causes of
problems and to develop suggestions what might have been done better regarding a
single incident or phase and how the overall service process could be modified to
ensure this.
4.2 Data analysis and general findings

4.2.1 Customer paths from focus group interviews
As expected, trom the foeus group interviews there did not evolve one single, stringent
eustomer path to be applied for all cf the dealers. First, the paths showed variants and
biturcations for phases cf the service process that are only gone through altematively
(see also Oe Ruyter et al. 1997, p. 239).2
Second, while showing several identical steps that had been mentioned in every fo-
cus group, e.g. making a date tor the repair or paying, the customer paths showed
remarkable differences regarding other steps. So, for instance, the customers of the
car dealer situated in a city unanimously were the only ones to mention that one of
the steps to take when taking the car to the dealer was finding a parking lot.
The explanation for this fact is as simple as it is intriguing: Since leaving the car on
the premises of their car dealers was no problem for the customers cf the other two
dealers, these did not perceive this phase at all or did not find it worth mentioning.
For the others, still, searching for a parking lot near their dealer was a distinguish-
able stage of their service experience. Other incongruities between the customer
paths can be explained by service components which are actually only offered at
one dealer, e.g. a test drive with the mechanic after arepair. For reasons of compa-
rability in data analysis and presentation, the three dealer-specific paths were
eventually combined to a consolidated customer path (see figure 3).
The Ihree palhs are rendered in Appendix A 10 C.

Phase 1 Recognition of the necessity of car service
Phase 2 Setting an Appointment
Phase 3 Driving to dealer
Phase 4 Reception and transfer of the car to the dealer
Phase 5 Leaving the premises
Phase 6 Repair and maintenance
Phase 7 Call back during service works

• by customer
• by dealer
Phase 8 Retrieving the car
Phase 9 Explanation of carried out works
Phase 10 Test drive
Phase 11 Settlement of the bill
Phase 12 Driving back
Phase 13 After-service feedback

• initiated by customer
• initiated by dealer
Fig. 3: The consolidated customer path
4.2.2 Reported Incidents

As proposed in 3.3, the reported incidents from focus groups and individual inter-
views were separated from the suggested innovations for analysis purposes. After
excluding some narrations collected among the incidents, which did not fulfill the
formal requirements to be accepted as a single and distinguishable incident (see
StausslWeinlich 1997, p. 40 for details), the remaining 543 incidents were classified
by two independent schemes:
According to the consolidated customer path, the incidents were assigned to phases
of this process. Most of the incidents had already been assigned correctly by the
respondents during the interviews so that only same deviations between the individ-
ual phases of the service process and the order of narration had to be corrected.
Figure 4 gives an overview of the distribution of the reported incidents on the phases
of the service process. It is obvious that each phase was not mentioned by each
customer. This even goes for phases that necessarily must have been undergone by
every respondent. A reasonable explanation lies in timely restrietions which may
have caused customers to concentrate on phases of the service process which they
could illustrate with especially remarkable incidents.
It is worth mentioning that more than a third of the incidents were assessed as highly
positive, scoring the top two ratings on a nine-point rating scale while only 5% were
given one of the warst two ratings; more than half of the events were attributed rat-
ings in the central five points of the scale (so-called usual incidents).
The insights from the qualitative analysis cannot be rendered here in detail. Instead,
the findings regarding the phases mentioned most frequently are outlined below:
• In arranging a date for the repair, customers positively acknowledged the

possibility to get an appointment at short notice.
• What was appreciated most in the reception and transfer of the car to the
dealer, were techniques offered to drop the car keys at the dealer's, inde-
pendent of opening hours. While the basic technique of putting the key into
an envelope and throwing it into a box was described positively, security
matters of this technique were criticized. Same customers criticized queuing
at peak hours and the time-consuming process of entering customer and car
data.
• The main factor of satisfaction with the retrieval of the car after the repair
seems to be, whether the repairs have been completed by the agreed time.
The Sequential Incident Technique for Innovations (SI TI) 207
• When seitling the bill, the amount payable seems to be of less importance
than the mode of payment. What obviously infuriated customers was, when
they had to pay cash while the customer next to them was allowed to pay by
Eurocheque card or credit card.
Recognition of the
necessity of car service
12
Setting an appointment
Driving to dealer
Receplion and transfer of
the car to the dealer
Leaving the premises
Repair and maintenance
CaU back during service

works
Retrieving the car
Explanation of canried out
works
Test drive
Settlement of the bill
Driving back (horne)
After-service feedback
.5
_ 1S
Fig. 4: Distribution of incidents by phases of the service process
The analysis of the qualitative data further led to the conclusion that some phases
that were part of the service process at a single dealer, like the test drive, were
highly appreciated there, while their colleagues (rather than competitors) did not of-
fer such an added value to their service.
Additionally, a deductive categorization was conducted using a slightly modified

scheme proposed by Hentschel (1992, pp. 245 ff.) which explicitly transposes the
five dimensions of service quality "tangibles", "reliability", "responsiveness", "assur-
ance", and "empathy" (Parasuraman/Zeithaml/Berry 1988) to the car service industry
and supplies a set of 30 car-service-related subcategories. As the focus of this paper

is on innovations, this categorization is not set out in detail.
4.3 Innovations
As the suggestions for improvement and the innovative ideas stemmed from different
respondents (customers and service personnei) and had been collected in different
interview settings (individual and focus group interviews), they did not only refer to
different phases and problems, but also greatly varied in the degree of elaboration,
concreteness and feasibility.
In order to cope with the large amount of valuable but unstructured qualitative data,
a team of four researchers was set up. Thus, the number of information interfaces,
which usually entail a loss of information, was limited, and highly intensive discus-
sions were granted. This team carried out the following steps:
• Assignment of those suggested innovations, which had not been mentioned

with reference to a specific phase of the customer path anyway, to phases of
the consolidated customer path;
• Analysis and inductive systematization of the suggested innovations;
• Condensation of selected partial innovations to a small number of comple-
mentary coherent concepts for the redesign of the service process at a car
dealer.
4.3.1 Analysis and systematization of proposed

innovations
From all interviews - focus group interviews and individual interviews with customers
and individual interviews with service staff - 262 innovations could be collected,
which, due to repetitions, made 72 different suggestions.
Related to the total number of respondents, the average number of innovations per
interviewee amounts to 1.7. The number of innovations per respondent was re-
markably higher in the heterogeneous focus groups where the average was between
3.1 and 3.8 innovations. An evident explanation could be thai group heterogeneity
provided a broader variety of perspeclives, resulting in more vivid discussions and

greater creativity. Still, creativity in the homogeneous group could also have been
lower, not due to its homogeneity, but because most of its members were pensioners
and had been regular customers of that dealer for a very long time.
Obviously, the distribution of innovations among the phases as rendered in figure 5

only faintly resembles the distribution of the incidents in figure 4.
.3
Recognition of the
necessity of car service
Setting an appointment
Driving to dealer _
_11 22
.8.11
Reception and transfer of 88
the car to the dealer
Leaving the premises _ 2S
Call back during service
worlts
Retrieving the car _
Explanation of carried out
works
Testdrive
.7
12
18
Driving back (horne)
.11
11
37
Not phase-related _ 24
Fig. 5: Distribution of innovations by phases of the service process
It is noticeable, that in the phases "Reception and transfer of the car to the dealer"
and "Settlement of the bill', which had the highest number of negative incidents, the
number of innovations generated was overproportional. This gives a first hint, that a
closer look at the relation between an innovation generated with respect to a specific
incident and the assessment of this incident may yield some valuable insights.
As only the ineidents eolleeted in the individual customer interviews were assessed
on a numerical seale, it is necessary to coneentrate on incidents and innovations
from that source. Interviews with 115 customers resulted in 543 reported incidents,
519 of which also had been validly rated on the nine-point scale mentioned in chap-
ter 3.2.2.1. The number of innovations generated here amounts to 66, which means
0.6 innovations per interview or that about 12% of all reported incidents were fol-
lowed by the suggestion of an innovation. More interesting than these average val-
ues is the distribution of innovations by the degree of (dis-)satisfaetion with the re-
spective ineident.
It could be found that, while incidents rated as neutral or positive were the most fre-
quent, negative ineidents seem to have a substantially higher potential to be the
starting point for innovations (figure 6). Whereas less than 13% of all incidents were
rated as negative, these entailed overproportionally many, i.e. 26%, of all innova-
tions.
Percentage foltowed by a
suggestion of an innovation
40% 37%
30%
Average: 12%
19%
20%
15%
___ ~tQ% ____~~ ~___ _____ _
A4 A3 A2 A1 o H1 H2 H3 H4
A = Anger H = Happiness
Rating of ineident
Fig. 6: Percentage of incidents followed by a suggestion of an innovation, by rating of incident

The Sequentiallncident Technique For Innovations (SI TI) 211
Surprisingly, no obvious relation could be found between the number of innovations

generated by a customer and the number of incidents reported by this person or
demographie or service-related background variables.
However, the explorative nature of the study and the quality of the data did not allow
testing hypotheses regarding the relations between number or assessment of the
incidents and the number of innovations.
While a listing of all innovations suggested would be of liltle use for the reader,
some still seem worth mentioning.
a) For the phase of transferring the car to the dealer, several customers pro-
posed ways of submitting car, keys and papers independent from the
dealer's opening hours. A locker system which they had encountered at a
car rental company was mentioned several times as a model.
b) Furthermore, chip cards storing customer data were suggested as a way to
replace the process of keying in.
c) For their waiting time in front of the reception desk, customers asked for in-
stance for a children's corner and media (incl. multimedia) entertainment but
some also were interested in an extension of the offer of accessories at an-
other counter.
d) The many innovations suggested for the phase of seitling the bill are also
worth mentioning due to their mere number. All of the car dealers show
strong resistance against other forms of payment than cash. Now, especially
among the many alternative proposals for non-cash payment, it is apparent
that not only do customers show a strong demand for greater flexibility but
also the service staff, who have to inform their customers about the rigid
payment regulation, feel rather embarrassed by this procedure and would
prefer to offer other ways of payment.
4.3.2 Development of concepts for service innovations

The innovations suggested did not represent a coherent concept for process modifi-
cation. Few of them referred to the service process as a whole; rather most sugges-
212 A/exander Decker / Hara/d Meissner
tions referred to singular problems or single steps of the customer path. So, the
systematization and the integration of selected ideas into an innovative concept was
up to the researchers. In order to iIIustrate this potential, selected innovative single-
aspects were chosen to exemplarily build three innovative service processes by the
research staft. Two of these concepts are outlined roughly, a third is presented in
more detail:
• Many innovation proposals reflected how the easy availability of certain bits
of information may promote customer satisfaction. From the whole of these
ideas it was possible to set up a concept for an improved database com-
prising selected information concerning the customer, his car(s) and her/his
contacts to car dealers. The focus of the concept was not on extensive data
collection but on applications useful to customers and dealers.
• An existing integrative concept for a customer club referring to the car
manufacturer could be modified and enriched by combining innovations to
consistent club programs directed towards certain target groups.
• The transfer of the car to the dealer and the reception had obviously not only
triggered many incidents worth reporting but also fanned the creativity of
customers and staft while searching for innovations. Therefore, a variety of
innovations were integrated into the concept of an automatie service coun-
ter which enables customers to take their car to the dealer independent from
any opening hours. A chip card is the key to the system: Based on the data
stored about the customer and the car, it allows the customer to park her/his
car on the premises of the dealer, assigns scarce parking lots, gives access
to a customer information and dialogue system and is used for payment like
a credit card. By means of a touch-screen-based information system, the
customer can either collect more information about arepair or maintenance
order she/he would like to place, enter this order or simply ask to be called
back by a service employee. Whenever material things need to be trans-
ferred between the customer and the company such as keys, papers, in-
voices or bills, a simple locker system can be applied.
5 Implications, limitations and hypotheses

5.1 Implications for future use and research
Successful new services - as stated above - ought to be developed systematically.
Consequently, a tool for searching for ideas as one stage in the development proc-
ess of new services should be established. In this sense, the developed and pre-
sented technique, the Sequential Incident Technique for Innovations (SITI), can be
seen as a useful research tool.
As proposed in 3.1, the SITI encompasses four sieps. Taking the experiences of the
study into consideration, another step has to be added. Since the innovations did not
represent a coherent concept for process modification, the systematization and the
integration of selected ideas into an innovative concept was up 10 the researchers.
Therefore, we propose to enclose this analysis as the fifth step of the SITI.
As a result, Ihe SITI encompasses the following five steps:

1) Identification of all process steps of the service as gone through by the cus-
tomer (.customer path"),
2) collection of (critical and usual) incidents on the basis of the identified path,
3) assessment of the reported incidents, and
4) generation and collection of suggestions and ideas relating to the reported
incidents, how service quality can be improved from the customer's point of
view,
5) systematization and integration of selected ideas into an innovative concept.
Data was collected in group or individual interviews with respondents of different

provenience. To gain innovations trom inside and outside the service organization,
the study integrated internal (the frontline employees) and external customers (the
consumers) into the idea generation process. Within these different procedures, the
data collection varied in the application of the proposed research steps of the SITI.
The reasons tor these deviations were specified in the respective chapters. Follow-
ing the experiences of the study, the same pattern can be suggested for future re-
search (see figure 7).
214 Alexander Decker I Harald Meissner
Phase Focus groups Individual interviews Individual interviews

with customers with customers with service staff
1 X
2 X X X
3 X
4 X X X
5 X
• time
Fig. 7: Proposed research steps of the SITI
Nevertheless, situations can occur, in which it may be reasonable to depart from this
suggestion. For example, in this study, foeus groups built the basis for the next re-
search steps. Under certain circumstances, another design can be used, if a com-
pany can rely on some experience in conducting surveys using SIT or SITI and
therefore can fall back on an existing eustomer path. Then it is possible to invert the
proposed procedure and to use individual customer interviews to select customers
with a high innovation potential and to invite them to focus group interviews. In this
case, customers who tell many (negative) incidents and generate many innovations
in the individual interviews, might be an asset to focus groups. This may be crucial,
as we found that customers had different abilities and willingness to generate inno-
vations from the reported incidents. It is noticeable, that especially regular custom-
ers, who were familiar with the service transactions, tended to hold onto existing
processes. The study in the car service industry showed, that these customers could
hardly integrate their knowledge of other service sectors into the situation with their
car dealers. Through the narration cf experiences, the respondents were too in-
volved into the specific subject, so they had difficulties in imagining different proc-
esses or procedures.
According to the study, interna I customer integration comprised only individual em-
ployee interviews. Although several remarkable contributions were made, it is as-
sumed, that the innovation potential was not exhausted. The special knowledge cf
frontline employees eoncerning internal proeesses and eustomers' needs also

seems to suggest the conduetion of internal eustomer foeus groups, tao. Sinee it is
assumed, that the innovation potential can be raised through the experienees made
while eonducting and taking part in the sessions, foeus groups with employees
should be conducted more often with the same staff. UsuaHy, this is not possible with
customers.
This is important, as the results showed, that the focus groups showed a much larger
and qualitatively higher innovation potential than the individual interviews. Customer
focus groups can be seen as popular means of discovering new ideas, especiaHy
when applying this kind of sequential approach. In this case, the advantage of the
dynamic interaction in focus groups can be combined with the pros of a systematic
research tool. For that and the time restrietions in individual interviews, this result
was not surprising.
Although the focus grau ps showed remarkable advantages in comparison to the in-
dividual interviews, there are also some limitations. First, as shown above and pro-
posed in the method, it is impossible to ascertain the customers' assessment of the
experiences in the sessions, because it would break the dynamic of the narration
process. Subsequently, the more important innovations cannot be evaluated from
the customers' viewpoint, but have to be determined by the researchers. Second, to
coHect a certain amount of customer-induced innovations from a quantitative point of
view, a vast number of focus groups have to be conducted. Taking the pros and
cons into account, service marketers must concentrate on the research goals to
choose one of the !wo alternatives or even both, as in the study described above.
As the results of the study showed, problems can be seen as a source of service
innovations. Through the application of the sequential approach, it was shown, that
the propositions made can form the basis for an internal benchmark, especiaHy
within franchise systems. Although the customers of the selected dealers were highly
satisfied and reported much more positive than negative incidents, many innovations
could be generated through the application of the SITI. Positive experiences also
contain an interesting potential for innovations, especiaHy, if these positive findings
are related to only one single outlet.
Furthermore, franchisees will have a higher tendency to implement innovations that

have been generated by their very customers and not by professional consultants or
members of the franchiser. Since the interviews are unaided, valid evaluations are to
be expected. It is assumed, that such perceptions have a higher relevance to cus-
tomers than evaluations on the basis of a given set of items.
It is remarkable that by the suggestion of innovations, shortcomings of the service

process were revealed that would hardly have been mentioned in an application of
the usual SIT. This was the case whenever improvements were suggested which
meant adding steps to the service process the lack of which is unlikely to be re-
ported when going through the stages of the established process.
5.2 Limitations of the tool and the study

As the SITI is a qualitative research tool, limitations refer to the problems generally
considered with these methods. As the design of the technique shows, the SITI en-
compasses five steps. The required information has to be collected in costly and
time-consuming interviews, especially concerning the collection of incidents and the
generation of innovations. Furthermore, these costs are increased due to complex
non-standardized forms for the processing and analysis of data (Stauss 1992, p. 45).
In addition, the quality of the results is highly dependent on the quality of the inter-
viewers - with regard to interviewer bias - and analysts. Due to time limitations dur-
ing the interviews, customers did not have enough opportunities to come back to
propositions and develop them further. Consequently, the integration of the innova-
tions to a concept had to be undertaken through non-standardized analysis by the
researchers.
As the goal of the survey was to develop a tool for the generation of customer-
induced innovations and to apply the method in an exploratory study, we did not
concentrate on the development of hypotheses. Since one of the general objectives
of exploratory research is to specify vague problem statements into more precise
sub-problem statements and hopefully into specific hypotheses (Churchill 1995,
p. 147), we summarize research questions in the following chapter.
The Sequentiallncident Technique tor Innovations (SI TI) 217
Apart from the mentioned limitations, the developed method can only be seen as one
stage in a new service development process. It has to be integrated in models of
new service development as proposed by Scheuing/Johnson (1989, p. 30) or Bow-
ers (1989, p. 18). For that, the generated innovations have to be first tested in pro-
totypes and in market finally.
5.3 Hypotheses
As service innovations are usually generated in a haphazard fashion, we assumed
that customers have a higher sensibility for improvements and innovations through a
systematical approach, which requires the narration of the experiences they made
on the basis of a customer path customers have generated through service mapping.
Thus they are able to develop ideas from their point of view more easily than if only
asked for new service ideas. Due to the limited number of participants, it was impos-
sible to prove the stated assumption. As a tendency, we found that sequentially
leading customers (and employees) through their past service process experiences
triggered a large number of innovations. Therefore we hypothesize:
Hypothesis 1
H1: Using the SITI, innovations generated through customers are higher in num-
ber and in value of their content than by plainly aSking customers what could
have been done belter.
As stated above, positive incidents can be seen as a helpful source for the genera-
tion of innovations and therefore can be used as an internal benchmark. Even if only
a limited number of incidents could be collected in individual interviews with custom-
ers there was a clear tendency to be found: The share of negative incidents from
which innovations were developed was noticeably higher than the respective share
of neutral or positive incidents. Therefore we hypothesize:
Hypothesis 2
H2: With the rising number of negative incidents, the number and the value of
the content of innovations will be higher than in comparison to innovations
generated on the basis of positive incidents.
As we could not prove the method in comparison to a usual approach, there is no

evidence, that creativity techniques such as brainstorming, brainwriting or synectics
are superior as weil in number of generated ideas as in their content to the devel-
oped method. Regarding creativity techniques, they concentrate on a process, which
first will lead participants away from a problem in order to raise the variety of ideas
and then to come back to a special problem. Although the methods can be seen as
useful tools in generating innovations, there are also some limitations. One major
problem can be seen in the requirement, that these techniques have to be learned
by using them several times.
In developing and using the SITI, we ignored the suggestion to take distance from a
certain problem or incidenl. We chose an phase-oriented approach in order to fulfill
the required elements of services and service quality and, consequently, to facilitate
the procedure for the participants in order to generate ideas.
In addition, we found that the service providers in our study had some problems in
selecting customers for the focus groups. Therefore we think, that it might be even
more difficult to find customers in the service industry, who are willing to participate
in several meetings in order to learn and then use the creativity techniques. Conse-
quently, the benefits of the creativity techniques cannot be used comprehensively.
Therefore we hypothesize:
Hypothesis 3
H3: Using the SITI in the service induslry, innovations generated through cus-
tomers are at least as high in number and in the value of their content as by
applying creativity techniques in the usual way.
6 Summary
service innovations are usually generated in a haphazard way. There is a lack of
application of market research techniques in this area as weil. Therefore, the pur-
pose of the paper was to develop and apply a research tool, which should help sys-
tematically generate service innovations.
Although idea generation in services and new product development have much in
common, the specific characteristics of services and service quality had to be taken
into account. According to the required elements, the Sequentiallncident Technique
for Innovations (SITI) was presented at the current state of discussion. An explora-
tory study in the service area of three car dealers was conducted to apply the
method and generate service innovations. 262 proposals could be generated by the
customers in the survey, which led to a total of 72 different innovations.
The results of this first exploratory survey suggest that the Sequentiallncident Tech-
nique for Innovations is a helpful tool. The reported incidents deliver a broad and
qualitative data basis to identify industry-specific strengths and weaknesses in serv-
ice processes. These episodes built the background for the generation of customer-
induced service innovations, since a tendency was shown, that respondents had a
high sensibility through the narration of their experiences and therefore could de-
velop ideas more easily. Because customers generated innovations from their point
of view, ideas can be seen as relevant for service quality perception. Although the
majority of innovations contained partial solutions of problems, the information
showed a large potential for the development of new services.
The study also showed some limitations. Therefore, in forthcoming surveys, the im-
plications have to be considered, especially with regard to the conduction of focus
groups. Three hypotheses were developed, which have to be proven in future re-
search. According to the current state of discussion, there is need to undertake con-
ceptual research as weil, in order to solve the problem of time limitations while gen-
erating innovations through the voice of the customer.
220 A/exander Decker / Hara/d Meissner
References
Bailetti, A.J.lLitva, P.F. (1995): Integrating Customer Requirements into Product

Designs, in: Journal of Innovation Management, Vol. 12, No. 1, pp. 3-12.
Bitner, M.J./Booms, B.H./Tetreault, M.S. (1990): The Service Encounter: Diag-

nosing Favorable and Unfavorable Incidents, in: Journal of Marketing, Vol. 54,
No. 1, pp. 71-84.
Bitner, M.J./Booms, B.H.lMohr, L.A. (1994): Critical Service Encounters: The Em-
ployee's Viewpoint, in: Journal of Marketing, Vol. 58, No. 3, pp. 95-106.
Bowers, M.R (1989): Developing New Services: Improving the Process Makes it
Setter, in: The Journal of Service Marketing, Vol. 3, No. 1, pp. 15-20.
Churchill, GA (1995): Marketing Research. Methodological Foundations, Fort

Worth et al.
Clancy, K.J./Shulman, R.S. (1991): The Marketing Revolution, New York.
Oe Brentani, U. (1989): Success and Failure in New Industrial Services, in: Journal
of Product Innovation Management, Vol. 6, No. 4, pp. 239-258.
De 8rentani, U.lCooper, RG. (1992): Developing Successful New Financial Serv-

ices for Susinesses, in: Industrial Marketing Management, Vol. 21, No. 3,
pp. 231-241.
Oe Ruyter, K./Wetzels, M.lLemmink, J.lMattsson, J. (1997): The Dynamics of the

Service Delivery Process: A Value-Sased Approach, in: International Journal of
Research in Marketing, Vol. 14, No. 3, pp. 231-243.
Edvardsson, B.lHaglund, l.IMattsson, J. (1996): Analysis, Planning, Improvisation

and Control in the Development of New Services - Some Swedish Experi-
ences, Research Report, Service Reserach Center, University of Karlstad,
Sweden.
Edvardsson, 8./0Isson, J. (1996): Key Concepts for New Service Development, in

The Service Industries Journal, Vol. 16, No. 2, pp. 140-164.
Flynn-Reinecke, l.IGoldsmith, RE. (1993): Identifying Innovators in Consumer

Service Markets, in: The Service Industries Journal, Vol. 13, No. 3, pp. 97-109.
Gadrey, J.lGallouj, F./Weinstein, O. (1994): New modes of innovation - How serv-

ices benefit industry, in: International Journal of Service Industry Management,
Vol. 6, No. 3, pp. 4-16.
Gummesson, E./Kingman-Brundage, J. (1992): Service Design and Quality: Ap-

plying Service Slueprinting and Service Mapping to Railroad Services, in:
Kunst, P.lLemmink, J. (eds.): Quality Management in Services, As-
sen/Maastricht, pp. 101-114.
Harvey, J. (1992): Technology and the creation of value in services: a conceptual

model, Paper for the Service Productivity and Quality Challenge Conference,
The Wharton School, University of Pennsylvania, October 23-24.
Hentschel, B. (1992): Dienstleistungsqualität aus Kundensicht - Vom merkmalsori-

entierten zum ereignisorientierten Ansatz, Wiesbaden: Gabler.
Kassarjian, H.H. (1977): Content Analysis in Consumer Research, in: Journal of

Consumer Research, Vol. 4, No. 3, pp. 8-18.
Kingman-Brundage, J. (1989): The ABC's of Service System Blueprinting, in: Bit-

ner, M.J.lCrosby, L.A. (eds.): Designing a Winning Strategy, AMA, Proceeding
Series, pp. 30-33.
Liljander, V./Strandvik, T. (1995): The nature of relationship quality, in: Swartz,

TA/Bowen, T.E.lBrown, S.w. (eds.): Advances in Services Marketing and
Management, London, pp. 141-168.
Krueger, RA (1994): Focus Groups. A Practical Guide for Applied Research,

2. ed., Thousand Oaks et al.
Martin, C.R. Jr.lHome, DA (1993): Services Innovation: Successful versus Unsuc-

cessful Firms, in: Internationals Journal of Service Industry Management,
Vol. 4, No. 1, pp. 49-65.
Meyer, A./Westerbarkey, P. (1995): Bedeutung der Kundenbeteiligung für die

Qualitätspolitik von Dienstleistungsunternehmen, in: Bruhn, M. / Stauss, B.
(eds.): Dienstleistungsqualität, 2nd edition, Wiesbaden, pp. 81-103.
Morgan, D.L. (1994): Focus Groups as Qualitative Research, Newbury Park.
Osbom, A.F. (1963): Applied Imagination; Principles and procedures of creative

problem-solving, 3rd ed., New York.
Parasuraman, A./Zeithaml, VA/Berry, L.L. (1988): SERVQUAL: A Multiple-Item

Scale for Measuring Customer Perceptions of Service Quality, in: Journal of
Retailing, Vol. 64, No. 2, pp. 12-40.
Rathmell, J.R. (1974): Marketing in the Service Sector, Cambridge.
Scheuing, E.E./Johnson, E.M. (1989): A Proposed Model For New Service Devel-
opment, in: The Journal of Service Marketing, Vol. 3, No. 2, pp. 25-34.
Shostack, G.L. (1984): Designing Services That Deliver, in: Harvard Business Re-
view, Vol. 62, No. 1, pp. 133-139.
Shostack, G.L. (1987): Service Positioning Through Structural Change, in: Journal
of Marketing, Vol. 51, No. 1, pp. 34-43.
Stauss, B. (1992): Service Problem Deployment: Transformation of Problem Infor-

mation into Problem Prevention Activities, in: Intemational Joumal of Service
Industry Management, Val. 4, No. 2, pp. 41-62.
Stauss, B. (1993): Using the Critical Incident Technique in Measuring and Manag-
ing Service Quality, in: Scheuing, E.E.lChristopher, W. (eds.): The Service
Quality Handbook With Contributions from 57 Intemational Experts,
pp. 408-427.
Stauss, B./Hentschel, B. (1992): Attribute-Based versus Incident-Based Measure-

ment of Service Quality: Results of an Empirical Study in the German Car
Service Industry, in: Kunst, P.lLemmink, J. (eds.): Quality Management in
Services, AssenlMaastricht, pp. 59-78.
Stauss, B./Seidel, W. (1995): Prozessuale Zufriedenheitsermittlung und Zufrieden-

heitsdynamik bei Dienstleistungen, in: Simon, H.lHomburg, C (eds.): Kunden-
zufriedenheit, Wiesbaden, pp. 179-203.
Stauss, B./Weinlich, B. (1996): Die Sequentielle Ereignismethode - ein Instrument

der prozeßorientierten Messung von Dienstleistungsqualität, in: der markt,
Val. 35, NO.1, pp. 49-58.
Stauss, B./Weinlich, B. (1997): Process-oriented measurement of service quality:

applying the sequential incident technique, in: European Joumal of Marketing,
Val. 31, No. 1, pp. 33-55.
Viney, L.L. (1983): The Assessment of Psychological States through Content Analy-
sis of Verbal Communications, in: Psychological Bulletin, Val. 94, No. 3,
pp. 542-563.
Wilhelmsson, M. (1994): Service Development - Some Empirical Results, Project

Report, Service Research Center, University of Karlstad, Sweden.
Zeithaml, V.A. (1991): How Consumer Evaluation Processes Differ between Goods
and Services, in: Lovelock, C.H. (ed.): Services Marketing, Englewood Cliffs,
New Jersey: Prentice Hall, pp. 39-47.
Zeithaml, V.A.lBerry, L.L.lParasuraman, A. (1988): Communication and Control

Processes in the Delivery of Service Quality, in: Joumal of Marketing, Vol. 52,
No. 4, pp. 35-48.
The Sequential Ineident Teehnique for Innovations (SI TI) 223
Appendix
A Customer path of the car dealer in the city
B Customer path of the car dealer in a middle-sized town
C Customer path of the car dealer in the country

AppendixA
Recognition of the necessity of car service
Setting an Appointment
Driving to dealer
Searching for a place to park the car
Parking the car outside the

Reception dealer's premises, leaving the
keys inside the car
Transfer of the car to the

dealer Reception
Call back during service works

• by customer
• by dealer
Retrieving the car
Driving back
Appendix A: Customer paths of the car dealer in the city

The Sequentiallncident Technique tor Innovations (SI TI) 225
Appendix B
Setting an Appointment Agreement on order
Driving to dealer Agreement on transfer
Reception and transfer of the car to

the dealer Driving to dealer
Reception and transfer of the car to

Leaving the premises the dealer before or after opening
hours
~ ~
'--1--------.1
. Repair and maintenance .

• by customer
• by dealer
Retrieving the car during opening Retrieving the car before or after
hours opening hours
Settlement of the bill Bill sent to customer
~r-------------------------~~
Appendix B: Customer palh of Ihe car dealer in a middle-sized lown

AppendixC
Setting an Appointment
Driving to dealer
Reception and transfer of the car to the dealer

• by customer
• by dealer
Retrieving the car
Test drive with a service employee
Explanation of works
Driving back
Appendix C: Customer path of the car dealer in the country

Martin Wetzeis and Ko de Ruyter
Contents
Abstract 229
Introduction 229
2 Perceived Support 231
3 Consequences of Perceived Support: A Proposed Model 233
4 Method 236
4.1 Sampie and Procedure 236
4.2 Measures 236
4.3 Modeling Approach and Results 237
4.3.1 Measurement Model 237
4.3.2 Structural Model 239
5 Conclusions and Implications 240
References
Appendix
Beyond the call of duty 229
Abstract
Call centers can be considered as rapidly growing communication channels that
companies use to be more customer-oriented. In many cases, temporary statt, em-
ployed by specialized temporary helping service agencies, is used within call centers.
These agencies playa pivotal role in motivating and establishing desirable behaviors
of the agents. In this paper we examine exchange relationships of temporary call
center agents with their agency as weil as the intermediary working for the temporary
employment agency as antecedents to job satisfaction and in- and extra-role behav-
ior. The results indicate that perceived intermediary support has a direct effect on in-
role behavior and that perceived organizational support has a direct relationship with
extra-role behavior as weil as an indirect impact through job satisfaction of temporary
call center agents.
For the relationship between intermediary support and in-role behavior it could be
surmised that working in call centers requires both technical skills and 'people' skiffs.
The temporary work agencies generaffy provide training focusing on information and
communication technology as weff as training focusing on customer orientation. The
intermediary plays a facilitating role in taking care of the training needs of the tempo-
rary staff. Furthermore, organizational support by the temporary help service agency
positively affects extra-role behavior by the temporary staff or 'to go the extra mile'.
For the relationship between the temporary help service agency and their clients, the
willingness of temporary staff to go beyond traditional job requirements may consti-
tute a crucial competitive advantage.
1 Introduction
Call centers are quickly emerging as central information hubs for diverse customer
access channels in the areas of sales and services in many companies (Anton
1996). By fostering a relationship-oriented fOcus, these organizational units which
have as their primary task the processing in- and outbound communications with
customers and prospects are rapidly becoming a key source of customer-specific
knowledge and a source of competitive advantage. Arecent Economist (1997) article
230 Marlin Wetze/s / Ko de Ruyter
estimates the annual growth of call centers in the United States at 50% and in West-
em Europe at 40%. This growth has been fueled by the development of sophisti-
cated new information and communication technology (Holland/Hunt 1997). How-
ever, it has been argued that competitive advantage based on technological innova-
tion may not be sufficient for long-term success, as this may be easily matched by
competitors. What is frequently overlooked is that the success of a call center de-
pends largely on the combination of technological sophistication and personalized
service (Cowles/Crosby 1990; Sparrow 1991). With respect to this, the attitude and
behavior of call center employees plays a pivotal role (Ubaldi 1997).
Due to the phenomenal growth of call centers, around-the-clock opening hours, pan-
european operations, corporate downsizing, technological push and customer pull,
most firms use temporary person ne I as call center agents. In the Netherlands, for
instance, the major temporary help service agencies (THSA's) have established
business units specializing in temporary call center employment. Hiring temporary
workers as call center agents offers flexibility, lower cost structures and the avail-
ability of specific expertise for client organizations. However, these benefits may be
elusive if temporary call center agents are dissatisfied and less committed to their
work or to the client organization and since these employees form an important in-
terface with the firm's customers the effect of a negative employee attitude and be-
havior can be detrimental (Von Hippel et al. 1997). Therefore, in addition to relation-
ships with client organizations, temporary help service agencies specializing in call
center agents face the challenge of developing and maintaining good relationships
with their employees. Hutt (1995) suggests that this forms the motivational basis for
positive employee attitudes and beneficial behavior, i.e., going (Iiterally) beyond the
call of duty. A relatively large effort of academic study has been devoted to the con-
cept of relational exchanges between buyers and seilers (Dwyer et al. 1987;
Morgan/Hunt 1994; Sheth/Parvatiyar 1995; Sheaves/Bames 1996) and intra-
organizational relationships (Hunt/Morgan 1994; Nevin 1995). However, in the
scarce but increasingly significant literature on temporary employment the relational
exchange perspective has not yet been taken. Therefore, such a perspective is
adopted in this paper. The rest of this paper unfolds as folIows. First, we examine two
levels cf relational exchange conceming the THSA-temporary call center agent relation-
ship. These concem the perceived organizational support by the THSA and the interme-
diary at the THSA. Secondly, we identify types of favorable employee attitudes and be-
havior and we discuss the relationships between the relational exchange concepts and
employee outcome variables. Next, we report on the results of an empirical study de-
signed to test aforementioned relationships. In conclusion, we address the theoretical
as weil as the managerial implications of our findings.
2 Perceived Support
As the number of temporary call center agents continues to increase and is extended
to higher job-levels (i.e., supervisor or management level), TSHA's specializing in
this business are facing the challenge of managing exchange relationships not only
with their client organizations but also with their employees. Temporary employees
essentially can be viewed as internal customers. It has since long been advocated
that exchange relationships may be established both on economic and social
grounds (Blau 1964). Relationship management by THSA's towards temporary call
center agents has been based mainly on economic principles, for instance, by
offering competitive salaries and the chance to gain valuable work experience as
weil as investing in skills training and instruction in areas such as workplace
orientation (Feldman et al. 1994; Von Hippel et al. 1997). However, it has been
argued and demonstrated extensively in the organizational behavior literature
(Eisenberger et al. 1986; Organ 1988; Rousseau 1989; Konovsky/Pugh 1994;
Settoon et al. 1996) that the motivation al basis behind employees commitment lies in
the concepts of social exchange. Many of the concepts used in the literature of
relationship marketing and management, such as trust, commitment and social
bonds (Wilson 1995) are based on social exchange theory. Relationships are social
in nature in that they are based on the belief that Ihe exchange of mutual support will
be beneficial to the partners involved in the relationship. Social exchange, in
addition to economic incentives, therefore, may be an important key to maintaining
long-term relationships with temporary call center agents.
In addition to the fact thaI TSHA's have 10 establish and maintain dual relationships,
Ihe relationship has 10 be managed al two levels; (1) the relalionship between Ihe
232 Martin WetzeIs / Ko de Ruyter
THSA and the call center temp at a global level and (2) the relationship between the
boundary spanning intermediary (i.e., the contact person at the TSHA) and the agent
at the more specific, dyadic level. With respect to the former, it can be argued that
temporary agents develop a perception regarding the way in which the THSA helps
them in attaining their values, goals and objectives. It has been argued that the
support by the TSHA may have a ceiling effect on the relationship between
temporary agent and client organization (Feldman et al. 1994). Alternatively, high-
quality exchange relationships between TSHA and call center agents may reduce
the perceived risk that client organizations face in developing a relationship with the
TSHA. Client organizations will be more likely to hire temporary agents that are
content to be working for the TSHA. In this sense, taking care of people means
taking care of business. Furthermore, it has been demonstrated empirically that
perceived organizational support can be regarded as an attitudinal and disposition
predictor of on-the-job performance, job satisfaction, organizational commitment and
organizational citizenship behavior (Eisenberger et al. 1986; ShorelWayne 1993;
Settoon et al. 1996). Alternatively, lack of organizational support will result in
temporary call center employee discontent, which in turn may lead, to theft,
sabotage, non-cooperation and other forms of employee resistance (Tucker 1993).
In addition to the TSHA-agent exchange relationship, temporary workers are also

engaged in a relation with the intermediary working for the TSHA, which acts as their
main contact person. In advertising campaigns by some agencies the intermediary is
positioned as 'the agent's personal account manager'. Indeed, many temporary
employees interact with their intermediary on a frequent basis (Feldman et al. 1994).
In the context of regular employment, it has been shown that due to perceived
support and consideration by their supervisor, employees are more satisfied about
their work and that they may express loyalty to the organization and exhibit
beneficial behavior that is neither formally required nor enforceable (Brown/Peterson
1993; Podsakoff et al. 1993; Wayne/Green 1993; Settoon et al. 1996).
Intermediaries are often viewed by temporary personnel as their manager. Thus,
perceived intermediary support mayaiso have a positive influence on temporary
workers' aUitudes and dispositions. In the next section we will discuss the potential
influence of perceived TSHA and intermediary support on attitudes and behaviors of

temporary workers.
3 Consequences of Perceived Support:

A Proposed Model
Based on a review of the literature, we propose three potential consequences of
perceived TSHA and intermediary support for the temporary marketing employment
setting: (1) job satisfaction; (2) in-role behavior and (3) citizenship behavior. These
variables have been shown to be important with respect to a variety of exchange
relationships (Graen et al. 1982; Eisenberger et al. 1986; Konovsky/Pugh 1994;
Settoon et al. 1996).
As a result of the central role that job satisfaction plays in minimizing employee
turnover (Brown/Peterson 1993; Lee/Mitchell 1994), organizational effectiveness
(Brooke et al. 1988) and attainment of organizational goals (Chatman 1991), we
decided to focus on job satisfaction as the attitudinal consequence of perceived
support by temporary call center temps. Job satisfaction is often conceptualized as a
subject's evaluation of the work experience (both the job itself and the work
environment) based on the basis of both cognitive and affective aspects (cf.
Matthieu/Zajac 1990; Matthieu/Farr 1991). As such, job satisfaction represents the
attitude or knowledge structure that encapsulates workers feelings and beliefs about
the nature of their job and the organization (George/Jones 1996). Research has
shown that no differences in overall job satisfaction levels exist between permanent
and temporary workers (Von Hippel et al. 1997). For our purposes, we operationalize
job satisfaction as the evaluative judgement pertaining to the temporary job
experience.
As perceived organizational support can be viewed as agents' confidence that the

organization will live up to its exchange promises by, for instance, a fair reward
allocation, it has been positively associated with commitment and job satisfaction
(Eisenberger et al. 1986; Organ 1988; Mackenzie et al. 1993; Livingstone et al.
1995). Therefore, we predict a positive relationship between perceived TSHA
support and job satisfaction. According to path-goal theorists, supervisory support
234 Martin Wetzeis / Ko de Ruyter
has a positive influence on job satisfaction, as support and consideration of

(marketing) employees enables them to reach their objectives (Brown/Peterson
1993; Churchill et al. 1993; Podsakoff et al. 1993; Netemeyer et al. 1997). Hence,
we hypothesize a positive relationship between intermediary support and job
satisfaction.
Frequently, temporary employment is associated with very specific expertise or job

responsibilities. This is often the raison-d'etre for hiring temporary marketing
personnel. In-role behavior can be defined as the fulfillment of job responsibilities as
pre-specified in the job description or contract. It reflects the behavior that is
designed as a requirement for the job and as such formally rewarded. Performing job
requirements with uncommon effort can be viewed as a manner of a favorable
response to perceived support by (temporary) employees (KatzlKahn 1978). In this
way, in role behavior is an instrument for social exchange. There is empirical
evidence that perceived organizational support has a direct positive influence on the
performance of job requirements (Eisenberger et al. 1986; 1990; Williams/Anderson
1991). Therefore we hypothesize a positive relationship between perceived TSHA
support and in-role behavior. Based on the fact that perceived intermediary support
also forms a beneficial action aimed at temporary personnei, which may lead to
reciprocation, we also hypothesize a direct positive relationship between perceived
intermediary support and in-role behavior.
In addition to the extra effort put in the fulfillment of job responsibilities, agents may
also reciprocate perceived support by exhibiting extra-role behavior, Le., behaviors
not pre-specified in role descriptions that nevertheless facilitate the attainment of
organizational goals (Katz 1964). One form of extra-role behavior that has frequently
been studied in the organizational behavior literature and to a lesser extent in the
marketing literature is organizational citizenship behavior (e.g., Van Dyne et al.
1994; Netemeyer et al. 1997). Organizational citizenship behavior 'represents indi-
vidual behavior that is discretionary, not directly explicitly recognized by the formal
reward system, and that in the aggregate promotes the effective functioning of the
organization' (Organ 1988, p. 4). Several forms of citizenship have been identified
(GeorgelBrief 1992). These range from helping co-workers and conscientiousness to
civic virtue or showing concern for the organization. Previous research has dem on-
strated a direct positive relationship between perceived organizational support and
citizenship behavior (ShorelWayne 1993). In addition a direct positive relationship
between leadership support and citizenship behavior has also been identified
(Wayne/Green 1993; Settoon et al. 1996).
Finally, while it remains unclear whether job satisfaction leads to behavior or vice
versa, most theoretical and empirical evidence has identified job satisfaction as an
antecedent of in- and extra role behavior (Niehoff/Moorman 1993; Organ/Ryan 1995;
Netemeyer et al. 1997). Therefore, we also hypothesize indirect effects of both in-
termediary and organizational support on the two types of role behavior by incorpo-
rating a link between job satisfaction and in-role behavior and citizenship behavior.
Based on the arguments presented above our proposed model is depicted in fig-
ure 1. All relationships in the model are expected to be positive.
Fig. 1: Conceptual Framework
In the next section we report on an empirical test of the model.

236 Marlin Wetzeis / Ko de Ruyter
4 Method
4.1 Sampie and Procedure
Questionnaires were handed out by the researchers to temporary call center agents
at locations of !wo THSA's in the Netherlands specializing in temporary call center
marketing services with instructions to return the completed questionnaire directly to
the researchers in the postage-paid envelope that had been provided. Respondents
were guaranteed complete anonymity. Questionnaires were handed out to 200 po-
tential respondents and 98 questionnaires were returned resulting in a response rate
of 49%. Because of too many missing data, 8 of the surveys returned could not be
used for further analysis. Finally, the results of the time-extrapolation test revealed
no statistically significant differences (a=0.05) on key variables. Of the respondents,
55% were female. The average age was 29 years. The large majority was employed
at agent level (88%), while a small number of respondents (12%) was employed at
the supervisor or analyst level. Most respondents (40%) have been employed tem-
porarily for longer than 1 year.
4.2 Measures
On the basis of extensive pre-testing measurement scales were kept as brief as
possible on the basis of a special request by the participating TSHA's. Perceived
TSHA support was measured with the adapted four item version of the scale devel-
oped by Eisenberger et al. (1986). In order to measure perceived intermediary sup-
port, a scale developed specifically for the purpose of this research was used. Four
items asked the call center temps to evaluate the support received from the interme-
diary at the THSA. The measure of job satisfaction used was an adapted four item
version of the scale developed by Churchill et al. (1974). In-role behavior, the extent
to which temporary agents feel they meet the requirements of their job responsibili-
ties was measured with a four item scale, which was adapted from Wil-
liams/Anderson (1991). Finally, citizenship behavior, or more specifically helping co-
workers, was also measured with an adapted version of a scale developed by Wil-
liams/Anderson (1991). Four items measured the degree 10 which temporary workers
feit that they engaged in behaviors beyond the call of duty. For each ilem of the
Beyond the cal/ of duty 237
questionnaire, a 7-point response format ranging fram 1 = completely disagree to 7 =

completely agree was used. The Appendix contains sam pie items of the measures
employed in this study.
We used self-reported measures only in our study. As Netemeyer et al. (1997) argue
self-reported rather than management or intermediary reported measures have the
advantage of providing a more complete picture of employee attitudes and behaviors
of which intermediaries or client company supervisors might not be fully aware. All
items were translated from English into Dutch via a pracedure of double-back trans-
lation by a qualified translator in the Netherlands (cf. Brislin 1980).
4.3 Modeling Approach and Results

As suggested by Anderson/Gerbing (1988) we employed a two-stage approach to
estimating the model in our study. In the first stage we estimated the measurement
model using a first-order confirmatory factor analysis. In the second stage we esti-
mated the structural model. This approach takes care of the potential problem of in-
terpretational confounding (Anderson/Gerbing 1988; SilvialMacCalium 1988; Wil-
liams/Hazer 1986). Interpretational confounding refers to the occurrence of distorted
structural parameters by simultaneously estimating measurement and structural
models.
4.3.1 Measurement Model

The results of the first stage of our modeling approach is exhibited in figure 2.
We initially obtained an unacceptable fit for the measurement model

(l(265)=577.69, [p<0.001]; GFI=0.80; AGFI=0.75; CFI=0.82; TU=0.80). However,
after an iterative process, in which we inspected the t-values, the patterns of stan-
dardized residuals and the modification indices (Anderson/Gerbing 1988; Stehen-
kamp/van Trijp 1991), ten items were deleted and we obtained an adequate fit:
l(80)=113.45 (p=0.008); GFI=0.93; AGFI=0.90; CFI=0.97; TU= 0.96 (BagozzilYi
1988; Bollen 1989; Steenkamp/van Trijp 1991). Inspecting the standardized residu-
als we found that none of these substantially exceeded 12.581. the cut-off value sug-
gested by Jöreskog/Sörbom (1989). Finally, we examined the Q-plot ot the stan-

dardized residuals. The Q-plot clearly showed a linear trend through the plotted val-
ues indicating a good fit (BagozzilYi 1988; Jöreskog/Sörbom 1989). For turther
analyses there remained three items tor organizational support, three items tor in-
termediary support, four items for job satisfaction, two items for in-role behavior and
three items for organizational citizenship behavior.
Goodness of Fit Measures
X2 113.45 (p=0.008)
Df 80
GFI 0.93
AGFI 0.90
CFI 0.97
TU 0.96
Construct Number of Items Composite Reliability
Organizational support 3 0.79
Intermediary support 3 0.78
Job satistaction 4 0.85
In-role behavior 2 0.80
Citizenship behavior 3 0.71
Fig. 2: Results of Measurement Model
Before we proceed with the structural model we first have to assess construct validity
of the measures (Steenkamp/van Trijp 1991). All of the constructs show sufficient
uni-dimensionality, since none of the standardized residuals substantially exceeds
12.581. Furthermore, none of the modification indices for the matrix l' substantially
exceeds 3.84. Within-method convergent validity was assessed testing the signifi-
cance and magnitude of the elements of the matrix l x • We found that all items
loaded higher than 0.5 on their respective constructs with a minimum t-value of 8.45.
Beyond the ca/l of duty 239
Discriminant validity was evaluated by testing wh ether pairs of constructs were cor-
related less than unity. We used a l'Ji with one degree of freedom to test for unity
between the ten pairs of constructs. All test were significant at a=0.05. Composite
reliability was calculated as suggested by Jöreskög (1971) All the constructs exhibit
adequate reliability as can be seen in Table 1. Dur analyses show that our measures
exhibit adequate uni-dimensionality, convergent and discriminant validity, and reli-
ability with regard to all constructs (Steenkamp/van Trijp 1991).
4.3.2 Structural Model

A variance-covariance matrix on the basis of listwise deletion of missing va lues was
calculated using PRELlS2 and was used as input to LlSREL8 (Jöreskog/Sörbom
1993). This program provides maximum likelihood estimates of the free parameters
in the structural. As the l-value is not independent of sampie characteristics
(Marsh/Hocevar 1985; Bollen 1989; Jöreskog/Sörbom 1989), a wide variety of fit in-
dexes have been developed that are supposedly independent of sampie character-
istics (Marsh et al. 1988; Gerbing/Anderson 1985; Hu/Bentler 1995; Marsh et al.
1996). Among these the Tucker-Lewis Index (TLI) (Tucker/Lewis 1973; Bent-
ler/Bonett 1980) and the Comparative Fit Index (Sentier 1990) seem to be relatively
unaffected by sampie characteristics (Marsh et al. 1988; Marsh et al. 1996).
We estimated the conceptual model as depicted in figure 1. It can be concluded that

the fit of this model is adequate: l(81 )=214.68, (p<0.001); GFI=0.88; AGFI=0.82;
TLI=0.90; CFI=0.91. Furthermore, the values of the l/df ratio also indicates a good
fit of the data to the hypothesized model. Inspection of the standardized residuals
showed that none of these substantially exceeded 12.581. Finally, the Q-plot clearly
showed a linear trend through the plotted values indicating a good fit. As our study is
exploratory, we followed the suggestions by Silvia/MacCalium (1988). We estimated
all parameters in the structural model and proceeded by fixing non-significant pa-
rameters one at a time. We found that three of the paths hypothesized in the con-
ceptual model were not significantly different form zero. Deleting these paths did not
significantly affect the fit (i1x\3)=7.52; p=0.06) and resulted in a more parsimonious
model (l(84)=22220, (p<0.001); GFI=0.88; AGFI=0.82; TLI=0.90; CFI=0.91). This
model is exhibited in figure 3. Finally, it should be noted that given the exploratory
nature of our study further research will be needed to assess the cross-validity of our
model (CudeckiBrowne 1983; MacCalium et al. 1992; Silvia/MacCalium 1988).
0.47
(5.7
0.32
(4.41)
0.30
(3.3
Fig. 3: Hypothesis Testing

Note: Standardized path coefficients with t-statistic between brackets.
As can be observed form figure 3 in-role behavior is determined by intermediary

support (standardized path coefficient=0.47; t-value=5.70) and job satisfaction
(standardized path coefficient=0.32; t-value=4.41); the R2 for in-role behavior equals
0.42. Citizenship behavior is affected by organizational support (standardized path
coefficient=0.34; t-value=3.66) and job satisfaction (standardized path coeffi-
cient=0.30; t-value=3.37) the R2 for citizenship behavior equals 0.28. Job satisfac-
ti on is influenced solely by organizational support (standardized path coeffi-
cient=0.40; t-value=5.05).
5 Conclusions and Implications

Relational exchange is a phenomenon that is receiving a great deal of interest from
both academics and practitioners in marketing and related areas. Frequently, tempo-
rary employees are not perceived as THSA customers. The majorily of marketing
activities is aimed at client organizations. However, it is increasingly acknowledged

that such a one-sided focus might constitute a severe limitation. In order to be com-
petitive in the growth market of temporary call center employment services, relation-
ship marketing efforts will have to be extended towards the agent. Therefore, we
took a relational exchange perspective in order to explain temporary call center
agents' attitudes and behaviors in this paper. Dur results show that the two ex-
change relationships that were discerned (i.e., support at the organizational and in-
termediary level) exert a different impact on employee job satisfaction, in- and extra
role behavior. More specifically, we find that a relatively strong direct effect of inter-
mediary support on in-role behavior. Individual support by the intermediary appar-
ently is not mediated by job satisfaction. The strength of the association between
intermediary support and in-role behavior is consistent with previous findings in the
organizational behavior literature (Podsakoff et al. 1990; SeUoon et al. 1996). A pos-
sible explanation for this may be that working in call centers requires specific exper-
tise in the areas of information and communication technology as weil as a sense of
customer orientation. The THSA's that participated in our study offer on-going train-
ing in these areas whereby the intermediary functions as a main contact person.
Therefore, the intermediary in particular plays a pivotal role in motivating temporary
employees to fulfil the required job responsibilities. It can be noted that intermediary
support does not have a direct influence on job satisfaction. This is in line with the
so-called 'substitutes for leadership' theory according to which characteristics of the
task counterbalance the effect of leadership-related variables on attitudinal variables
such job satisfaction and organizational commitment (Podsakoff et al. 1993).
The other exchange relationship taken into consideration in our study, organizational
support, exerts a direct influence on citizen behavior. A high level of perceived or-
ganizational support by the THSA apparently motivates temporary employees to 'go
the extra mile'. Perceived organizational support reflects the agents' belief about the
way in which the TSHA values their efforts for client organizations. The positive as-
sociation with citizenship behavior could be explained by the fact that many TSHA's
have been focusing on the creation of structural conditions that increase the com-
patibility of agendas and expectations of their workers. Many agencies are incorpo-
rating reward systems and employee benefits that recognize responsive behavior
and aim at a congruence of the individual employee with the strategie objectives and
values of the organization. The direct effect of perceived organizational support on
citizenship behavior is consistent with results reported in the organizational behavior
literature (Eisenberger et al. 1990; ShorelWayne 1993). However, in the temporary
call center employment setting exhibiting extra-role behavior benefits the dient or-
ganization and in turn is likely to enhance a high-quality TSHA-client organization
relationship. Moreover, in a temporary employment setting that is increasingly proj-
ect-focused, it seems more and more important for employees to go beyond the tra-
ditional job requirements for the benefit of the projec!.
In addition, we find that organizational support has an indirect effect on both types of
behavior thraugh its effect on job satisfaction. Previous studies have demonstrated
the positive impact on beneficial employee behaviors (Organ 1988; Niehof/Moorman
1993; Organ/Ryan 1995). A theoretical rationale for the indirect effect of organiza-
tional support on in- and extra-role behavior forms the 'norm of reciprocity' (Gouldner
1960; Levinson 1965). As support by the organization results in employees that are
satisfied with their job, they will exhibit behavior that is beneficial to the organization
in return for the benefit of organizational support. In this way a high-quality exchange
relationship will develop.
Part of the strength of a research praject lies in the recognition of its limitations. This
may suggest potential issues that merit future research. In the first place, caution
should be taken in reference to statements of causality. As both cross-sectional data
and a non-experimental design were used and despite the fact that advanced mod-
eling techniques were employed, any causal inferences based on the results of this
study are precluded. Furthermore, inferences about generalizability should be
treated with equal caution. It is unknown whether our findings are generalizable to
the other forms of temporary employment (e.g., praject management services) that
are currently developed in this marke!. Future research will have to reveal the gen-
eralizability potential of our results. A second limitation concerns the fact that all
variables (both endogenous and exogenous variables) were measured from one
source, as self-reported data. Despite the advantages cited earlier, the results ob-
tained may be influenced by consistency effects and common method variance. In
further research designs it would be advisable to have call center managers fram the
client organization rate in- and extra-role behavior from temporary agents. Thirdly,
all concepts were measured at one point in time, thus essentially from a static per-
spective. It may be worthwhile to study temporary employment exchange relation-
ships over time in order to be able to take into account the dynamics of relationships.
Finally, we suggest that in future research other types of exchange relationships
should be incorporated, as temporary employees in particular are engaged in multi-
ple exchange relationships (KrackhardtlHanson 1993). In addition to organizational
and intermediary support, exchange relationships with the management of the client
organization as weil as co-workers should be taken into consideration. Each of these
relationships may exert an influence on the attitude and behavior of temporary call
center agents.
Our results also suggest a number of managerial implications. First of all, the results
of this study suggest that the types of exchange relationships with temporary em-
ployees are differentially related to types of behaviors. As intermediaries play an
important part in stimulating in-role behavior, it seems important to ensure that in-
termediaries have an up-to-date knowledge of job requirements and working envi-
ronments of their agents .. Furthermore, the impact of intermediary support on in-role
behavior indicates the importance of leadership and relational competence on the
part of the intermediary in accomplishing this type of desirable behavior. Secondly,
given the impact of perceived support by the THSA, it appears to be important to pay
attention to the structural, social and individual factors that factors that facilitate the
TSHA's ability to be responsive to the needs, demands and expectations of call
center temps. This may vary from providing fair reward systems and training oppor-
tunities to personal coaching of temporary agents. This suggests that the temporary
employment organization should exhibit an active approach to interacting with their
temporary employees. High-quality exchange relationships with temporary employ-
ees and the development of human capital is a permanent job that contributes to the
development and maintenance of relationships with client organizations.
References
Anderson, J.C.lGerbing, D.W. (1988): Structural Equation Modeling in Practice: A

Review and Recommended Two-Step Approach, in: Psychological Bulletin,
Vol. 103, No. 3, pp. 411-423.
Anton, J. (1996): Call center management by the numbers, Santa Maria.
Bagozzi, R P.Ni, Y. (1988): On the evaluation of Structural Equation Models, in:

Journal of the Academy of Marketing Science, Vol. 16, No. 1, pp. 74-94.
Bentier, P.M. (1990): Comparative fit indexes in structural models, in: Psychological
Bulletin, Vol. 107, No. 2, pp. 238-246.
Bentier, P.M./Bonett, D.G. (1980): Significance tests and goodness-of-fit in the

analysis of covariance structures, in: Psychological Bulletin, Vol. 88, No. 3,
pp. 588-600.
Blau, P.M. (1964): Exchange and power in sociallife, New York.
Bollen, K.A. (1989): Structural equations with latent variables. New York: Wiley.
Brislin, R W. (1980): Translation and content analysis of oral and written materials,
in: Triandis, H.C.lBerry J.w. (eds.): Handbook of cross-cultural psychology:
Methodology, Vol. 2, Boston.
Brooke, P.P./Russell, D.W./Price J.L. (1988): Discriminant validation of measures

of job satisfaction, job involvement, and organizational commitment, in: Journal
of Applied Psychology, Vol. 73, No. 2, pp. 139-145.
Brown, S.P./Peterson, RA. (1993): Antecedents and consequences of salesperson

job satisfaction: meta-analysis and assessment of causal effects, in: Journal of
Marketing Research, Vol. 30, No. 1, pp. 63-77.
Carey, R. (1997): Help wanted - now, in: Sales and Marketing Management,
Vol. 149, No. 8, pp. 30-31.
Chatman, J. (1991): Matching people and organizations: selection and socialization

in public accounting firms, in: Administrative Science Quarterly, Vol. 36, No. 3,
pp. 459-484.
Churchill, G.A. Jr./Ford, N.M./Walker O.C. Jr. (1974): Measuring the job satisfac-
tion of industrial salesmen, in: Journal of Marketing Research, Vol. 11, No. 4,
pp. 254-260.
Churchill, G.A. Jr./Ford, N.M./Walker O.C. Jr. (1993): Sales force management.
Homewood.
Cudeck, R./Browne, M.W. (1983): Cross-validation of covariance structures, in:

Multivariate Behavioral Research, Vol. 18, No. 2, pp. 147-167.
Cowles, D./Crosby L.A. (1990): Consumer acceptance of interactive media, in: The
Services Industries Journal, Vol. 10, No. 3, pp. 521-540.
Dowling, M. (1996): Keeping Seasonals, in: Catalog Age, Vol. 13, No. 8, pp. 53-54.
Dwyer, F.R.lSchurr, P.H./Oh, S. (1987): Developing Buyer-Seller Relationships, in:

Eisenberger, R./Huntington, R./Hutchison, S.lSowa, D. (1986): Perceived organ-

izational support, in: Journal of Applied Psychology, Vol. 71, No. 3,
pp. 500-507.
Eisenberger, R./Fasolo, P./Davis-LaMastro, V. (1990): Perceived organizational

support and employee diligence, commitment, and innovation, in: Journal of
Applied Psychology, Vol. 75, No. 1, pp. 51-59.
Feldman, D.C./Doerpinghaus, H.I./Tumley, W.H. (1994): Managing temporary

workers: a permanent HRM challenge, in: Organizational Dynamies, Vol. 23,
No. 2, pp. 49-63.
Frazee, V. (1996): Short-timers are the employees of the future, in: Personnel Jour-
nal, Vol. 75, No. 4, pp. 32-33.
George, J.M./Brief, A.P. (1992): Feeling good - doing good; a conceptual analysis
of the mood at work-organizational spontaneity relationship, in: Psychological
Bulletin, Vol. 112, No. 2, pp. 310-329.
George, J.M.lJones, G.R. (1996): The experience of work and turnover intentions:
interactive effects of value attainment, job satisfaction, and positive mood, in:
Journal of Applied Psychology, Vol. 81, No. 3, pp. 318-325.
Gerbing, D.W.lAnderson, J.C. (1985): The effects of sampling error and model
characteristics on parameter estimation for maximum likelihood confirmatory
factor analysis, in: Multivariate Behavioral Research, Vol. 20, No. 3,
pp. 255-271.
Golden, L. (1996): The expansion of temporary help employment in the US, 1982-
1992: A test of alternative economic explanations, in: Applied Economics,
Vol. 28, No. 9, pp. 1127-1141.
Gouldner, A.W. (1960): The Norm of Reciprocity, in: American Sociological Review,
Vol. 25, No. 4, pp. 165-167.
Graen, G./Novak, M.A.lSommerkamp, P. (1982): The effects of leader-member

exchange and job design on productivity and satisfaction: testing a dual at-
tachment model, in: Organizational Behavior and Human Performance, Vol. 30,
No. 7, pp. 109-131.
Greenwald, J. (1996): Temp agencies offering full-time benefits, in: Business Insur-
ance, Vol. 30, pp. 3-7.
246 Martin Wetze/s / Ko de Ruyter
Holland, R./Hunt, J. (1997): The importance of staying technically competent, in:

Customer Relationship Management, Vol. 2, No. 4, pp. 14-15.
Hu, L./Bentler, P.M. (1995): Evaluating model fit, in: Hoyle R.H. (ed.): Structural
Equation Modeling: Concepts, Issues and Applications, Thousand Oaks.
Hunt, S.D./Morgan, R.M. (1994): Relationship marketing in the area of network

competition, in: Marketing Management, Vol. 3, No. 1, pp. 19-28.
Hutt, M.D. (1995): Cross-functional relationships in marketing, in: Journal of the

Academy of Marketing Science, Vol. 23, No. 4, pp. 351-357.
Jöreskog, K.G. (1971): Statistical Analysis of sets cf Congeneric Tests, in: Psy-
chometrika, Vol. 36, pp. 109-133.
Jöreskog, K.G./Sörbom, D. (1989): LlSREL® 7: A guide to the program and appli-

cation. Chicago.
Jöreskog, K.G./Sörbom, D. (1993): LlSREL® 8: Structural Equation Modeling with

the SIMPLIS™, Hilsdale et al.
Katz, D.lKahn, R.L. (1978): The social psychology of organizations. New York.
Katz, D. (1964): The motivational basis of organizational behavior, in: Behavioral

Science, Vol. 9, No. 2, pp. 131-146.
Konovsky, M.A./Pugh, S.O. (1994): Citizenship behavior and social exchange, in:
Academy of Management Journal, Vol. 37, No. 3, pp. 656-669.
Krackhardt, O./Hanson, J.R. (1993): Informal networks: the company behind the
chart, in: Harvard Buisiness Review, Vol. 10, No. 4, pp. 104-111.
Lee, O.R./Mitchell, T.R. (1994): An alternative approach: the unfolding model of

voluntary employee turnover, in: Academy of Management Journal, Vol. 30,
pp. 51-89.
Lee, O.R. (1996): Why is flexible employment increasing?, in: Journal of Labor Re-
search, Vol. 17, No. 4, pp. 543-553.
Levinson, H. (1965): Reciprocation: The relationship between man and organiza-

tion, in: Administrative Science Quarterly, Vol. 9, No. 5, pp. 370-390.
Livingstone, L.P./Robberts, J.A./Chonko, L.B. (1995): Perceptions of internal and

external equity as predictors of outside salespeoples' job satisfaction, in: Jour-
nal of Personal Selling and Sales Management, Vol. 15, No. 2, pp. 33-46.
MacCallum, R./Roznowski, R.M.lNecowitz L.B. (1992): Model specification in co-

variance structure analysis: The problem of capitalization on chance, in: Psy-
chological Bulletin, Vol. 111, No. 3, pp. 490-405.
Beyond the caJ/ of duty 247
Mackenzie, S.B.lPodsakoff, P.M./Fetter, R. (1993): The impact of organizational

citizenship behavior evaluations of salesperson performance, in: Journal of
Marketing, Val. 57, No. 1, pp. 70-80.
Marsh, H.W./Balla, J.R.lMcOonald, R.P. (1988): Goodness-of-Fit Indexes in Con-

firmatory Factor Analysis: The effect of Sampie Size, in: Psychological Bulletin
Val. 103, No. 3, pp. 391-410.
Marsh, H.W.lBalla, J.R./Tai, H.K. (1996): An evaluation of incremental fit indices: a

clarification of mathematical and empirical properties, in: Marcoulides,
GA/Schumacker, R.E. (eds.): Advanced Structural Equation Modeling: Issues
and Techniques, Mahwah.
Marsh, H.W./Hocevar, O. (1985): Application of Confirmatory Factor Analysis to the

Study of self-Concept: First- and Higher Order Factor Models and Their Invari-
ance Across Groups, in: Psychological Bulletin, Val. 97, No. 3, pp. 562-582.
Mathieu, J.E./Farr J.L. (1991): Further evidence for the discriminant validity of
measures of organizational commitment, job involvement, and job satisfaction,
in: Journal of Applied Psychology, Vol. 76, No. 1, pp. 127-133.
Mathieu, J.E./Zajac, O.M. (1990): A Review and Meta-analysis of the Antecedents,

Correlates, and Consequences of Organizational Commitment, in: Psychologi-
cal Bulletin, Vol. 108, No. 2, pp. 171-194.
McGee, M.K. (1997): Just don't call them temps, in: Information Week, May 12,
pp. 38-40.
Morgan, R.M.lHunt, S.O. (1994): The Commitment-Trust Theory of Relationship

Marketing, in: Journal of Marketing, Vol. 58, No. 3, pp. 20-38.
Murphy, C. (1996): New rales for the brand consultant, in: Marketing, Sept. 26,
p.16.
Netemeyer, R.G./Boles, J.S./McKee, O.O./McMurrian, R. (1997): An investigation

into the antecedents of organizational citizenship behaviors in a personal sell-
ing context, in: Journal of Marketing, Val. 61, No. 3, pp. 85-98.
Nevin, J.R. (1995): Relationship marketing and distribution channels: exploring fun-
damental issues, in: Journal of the Academy of Marketing Science, Vol. 23,
No. 4, pp. 327-334.
Niehoff, B.P./Moorman, R,H. (1993): Justice as a mediator of the relationship be-

tween methods of monitoring and organizational citizenship behavior, in: Acad-
emy of Management Journal, Val. 36, No. 3, pp. 527-556.
Nollen, S.O. (1996): Negative aspects of temporary employment, in: Journal of La-
bor Research, Val. 17, No. 4, pp. 567-582.
248 Martin Wetze/s / Ko de Ruyter
Organ, D.W. (1988): Organizational citizenship behavior; the good soldier syn-
drome, Lexington.
Organ, D.W./Ryan, K. (1995): A meta-analytic review of attitudinal and dispositional

predictors of organizational citizenship behavior, in: Personnel Psychology,
Vol. 48, No. 4, pp. 775-802.
Podsakoff, P.M./MacKenzie, 5.B. (1990): Organizational citizenship behaviors and

sales unit effectiveness, in: Joumal of Marketing Research, Vol. 31, No. 2, pp.
351-363.
Podsakoff, P.M./MacKenzie, 5.B./Hui, C. (1993): Organizational citizenship be-

haviors and managerial evaluations of employee performance: a review and
suggestions for future research, in: Research in Person ne I and Human Re-
sources Management, Vol. 11, No. 6, pp. 1-40.
Rogers, J.K. (1995): Just a temp: experience and structure of alienation in tempo-
rary clerical employment, in: Work and Occupations, Vol. 22, No. 2,
pp. 137-166.
Rothstein, D.5. (1996): Entry into and consequences of nonstandard work ar-
rangements, in: Monthly Labor Review, Vol. 119, No. 10, pp. 75-82.
Rousseau D.M. (1989): Psychological and implied contracts in organizations, in:

Employee responsibilities and rights Joumal, Vol. 2, No. 2, pp. 121-139.
Sheth, J.N.lParvatiyar, A. (1995): The Evolution of Relationship Marketing, in: In-

temational Business Review, Vol. 4, No. 4, pp. 397-418.
Settoon, R.P./Bennet, N./Liden, R.C. (1996): Social exchange in organizations:

perceived organizational support, leader-member exchange, and employee
reciprocity, in: Joumal of Applied Psychology, Vol. 81, No. 3, pp. 219-227.
Sheaves, D.E./Bames, J.G. (1996): The Fundamentals of Relationships: An Explo-

ration of the Concept to Guide Marketing Implementation, in: Advances in
Services Marketing and Management, Vol. 5, pp. 215-245.
Shore, L.M./Wayne, 5.J. (1993): Commitment and employee behavior: comparison

of affective commitment and conti nuance commitment with perceived organiza-
tional support, in: Joumal of Applied Psychology, Vol. 78, No. 5, pp. 774-780.
Silvia, S.E.M./MacCallum, R.e. (1988): Some factors affecting the success of

specification searches in Covariance Structure Modeling, in: Multivariate Be-
havioral Research, Vol. 23, pp. 297-326.
Sparrow, L.B. (1991): Managing the telephone enquiry bureau at British Gas West
Midlands, in: Littlechild, S.C.lShutler, M.F. (eds.): Operations Research in
Management, London, pp. 167-173.
Steenkamp, J.B.E.M.Nan Trijp, H.C.M (1991): The Use of LlSREL in Validating

Marketing Constructs, in: International Journal of Research in Marketing,
Vol. 8, No. 4, pp. 283-299.
Tucker, J. (1993): Everyday forms of employee resistance, in: Sociological Forum,

Vol. 8, No. 1, pp. 25-45.
Tucker, L.R./lewis, C. (1973): The reliability coefficient for maximum likelihood

factor analysis, in: Psychometrika, Vol. 38, pp. 1-10.
Ubaldi, A. (1997): Denmark's focus on employee motivation & education creates

ideal call center environment, in: Direct Marketing, Vol. 59, No. 10, pp. 46-48.
Van Dyne, l./Graham, J.W./Oienesch, R.M. (1994): Organizational Citizenship be-

havior: construct redefinition, measurement, and validation, in: Academy of
Management Journal, Vol. 37, No. 4, pp. 765-802.
Von Hippel, C./Skoglind, J.D./Heneman, Rl. (1997): Temporary employment: can

organizations and employees both win?, in: Academy of Management Execu-
tive, Vol. 11, No. 1, pp. 93-104.
Wayne, S.J./Green S.A. (1993): The effects of leader-member exchange on em-

ployee citizenship and impression management behavior, in: Human Relations,
Vol. 46, No. 12, pp. 1431-1440.
Williams, L.J./Anderson, S.E. (1991): Job satisfaction and organizational commit-

ment as predictors of organizational citizenship and in-role behaviors, in: Jour-
nal of Management, Vol. 17, No. 3, pp. 601-617.
Williams, L.J./Hazer, J.T. (1986): Antecedents and Consequences of Satisfaction

and Commitment in Turnover Models: AReanalysis using Latent Variable
Structural Equation Methods, in: Journal of Applied Psychology, Vol. 10, No. 2,
pp. 219-231.
Wilson, D.T. (1995): An integrated Model of Buyer-Seller Relationships, in: Journal

of the Academy of Marketing Science, Vol. 23, No. 4, pp. 335-345.
Appendix
Construct Sampie Item
Organizational support The THSA really cares about my well-being
Intermediary support The intermediary does everything that helps me to do my

job as good as possible
Job satisfaction I am satisfied with work that I do through the temporary

help service agency
In-role behavior As a temporary employee I fulfill responsibilities specified

in the job description
Citizenship behavior I frequently help co-workers who have heavy workloads
Appendix: Sampie Items for measurement Instruments

Fran90ise Chevalier
Contents
Abstract 253
Introduction 253
2 Research within companies 255

2.1 Objective, Methods and Treatment of Data 255
2.2 Main Results in Companies Studied 256
3 Managing Change 264

3.1 Contrasting Assumptions About People and Organizations 264
3.2 Contrasting Assumptions about Change Implementation 267
3.3 Taking into Account Contradictions and Dilemmas: "Hybrid
Organizations" 269
4 Conclusion 271
References
Quality Management and Organizational Change 253
Abstract
Quality programs are becoming widespread in companies. Nonethe/ess their imple-
mentation is not always easy and often involves several difficulties. This paper pres-
ents the results of the research we have conducted in five companies.
Our research iIIustrates that very rigid, "standardized strategies" for change have led
to difficulties. On the contrary, "heuristic strategies" for change have given rise to
good results. These heuristic strategies for change implementation are essentially
characterized by trying to deal with a number of contradictory phenomena within or-
ganizations rather than attempting to suppress them. A Company's ability to appre-
hend and manage contradictions within organizations determines the deve/opment of
quality programs and their degree of success.
These different approaches to the change process usually stem from very different
underlying assumptions about the nature of people and organizations. Depending on
the way people and organizations are viewed, change implementation will be quite
different. Organizational models filter and focus attention.
Finally, it is not so/ely "what you change" that is important; it is also "how you
change". In other words, the nature of change does not necessarily determine the
change process.
1 Introduction
Since the heyday and aftermath of Taylorism, the success of which was hampered
by the movement's excessive rationalism and inherent dysfunctional nature, compa-
nies were forced to reconstruct their organizational designs. This led to the founding
of "Human Relations", along with new theoretical models, based for the most part on
social psychology and motivational studies, all of which stirred a great deal of corpo-
rate interest and gave rise to a broad range of innovations.
New solutions and models used to solve organizational problems appeared, some of
which focused more particularly on middle management while others were geared
254 Frant;oise Chevalier
toward non-managers. Over the past forty years, companies in France have experi-
mented with all sorts of organizational models from objective - oriented manage-
ment, to new organizational modes of work such as autonomaus groups, then Or-
ganizational Development (O.D.), the most recent models are quality circles, Total
Quality Management and nowadays reengineering. All gave rise to high hopes. Cor-
parate consultants and advisors underscored the importance of these various mod-
els and pushed for their implementation. Same companies tried them out and most
were satisfied at the beginning. Professional periodicals and the press at large im-
mediately publicized their results. A few years later, however, the model was aban-
doned by disappointed or angry management and personnel as they realized that
the solution had proved unsuccessful in the end (Beer/EisenstaUSpector 1990). An-
other model soon superseded the previous one with new experts, new personalized
training and new members who all rallied around a new "banner". This new organ-
izational model raised the same hopes and expectations, before being dropped
same time later. At their worst, these change programs can actually inhibit change,
to a large extent, by promoting skepticism and cynicism among personnel. Cynicism
can inoculate companies against change as employees find it difficult to believe that
wave after wave of programs will make a difference. Many of the people interviewed
said of the latest program, "This, too, will pass."
In that perspective, companies have recently attempted to respond both to competi-

tion and increased demand for quality. Many of them have introduced a new phi-
losophy, a new approach, or a new system to keep up with these challenges. Works
published on quality circles and total quality management have mushroomed. They
have served as guides presenting the methods, techniques and tools necessary'
being as directly operational as possible. In France total quality management hit
major companies both in the private and in the public sectors. Many experienced the
same infatuation with quality programs, although it led to considerably different re-
sults. In some of them successful transformations did take place. But many also en-
countered problems. There were several ca ses in which two of three years after a
CEO. launched a quality program, few things in the way of actual change in organ-
izational behavior and performance had actually occurred. Why do these programs
fail? or, more precisely why are the results so disappointing? Why, when so much
effort, money, and good intentions are invested, do some fall short of the mark?
While others succeed?
2 Research within companies

2.1 Objective, Methods and Treatment of Data
Over the last ten years, we have developed several empirical and longitudinal re-
searches on management and change processes. In these" fieldwork ", we strive to
go beyond strictly sporadic observations by studying phenomena in the long term.
Our key issue is to film changes step by step as they occur instead of providing mo-
mentary snapshots of a given phenomenon at a given moment (Chevalier 1987,
1991; ChevalierlSegalia 1996; Chevalier/Boumois 1998).
The first part of this paper introduces this kind of research within organizations de-
veloping new management programs and more specifically quality programs. The
objectives of this research on quality management is not only to contribute to pro-
viding more knowledge on the way quality programs actually operate in different
companies. It is also more theoretical and consists in wanting to study quality pro-
grams as special organizational forms of management, thus making it possible to
derive a broader analysis from the change process at work within organizations.
The results obtained have been derived from empirical surveys in five different com-
panies realized according to the method of strategie analysis developed by Croz-
ier/Friedberg (1997). Organizational analysis, which came into being with the indus-
trialization of modem societies, initially used a mechanistic model to explain organ-
izational operations and human behavior. Some believed organizational constraints
determined human behavior, while others feit the laUer was mainly based on psy-
chological data. The Strategie Analysis developed by Crozier/Friedberg counters
these mechanistic cause and effect relations. Strategie analysis views man as a free
agent who strives to realize his own specific goals and projects despite many or-
ganizational constraints. An actor's strategy does not answer just to one determined
rationality but to a contingent rationality, limited by the concrete conditions of the
action. In this perspective, an organization is not solely determined by the organiza-
256 Fran~ise Chevalier
ti on chart, the formal rules, or by the psychology of its members, it is a "Iiving body"
composed of free individuals searching for "satisfactory solutions" which take into
account their personal objectives and formal constraints.
Therefore it is necessary to consider perceptions, feelings and attitudes actors have

in regard to the organization in order to understand the real organizational dynamics.
In this empirical and clinical approach, the technique of interviews is a privileged
means of acquiring information beside the gathering of formal data. In our research
the information collected which has enabled us to grasp the way Total Quality pro-
grams actually operate was acquired through 45 individual semi-directive interviews
over a three years period. These interviews on the workplace lasted about two
hours, sometimes three hours. They did not consist in evaluating or criticizing prac-
tices observed but in understanding them; that was the basic principle which guided
our entire approach, which in also a well-known epistemological requirement. All our
interlocutors were assured that their comments would remain absolutely anonymous.
During these surveys, we interviewed upper management, CEOs and employees.
Moreover as observers we attended several quality training sessions, different
meetings related to quality programs and got access to numerous administrative re-
ports.
In order to exploit this mass of data, we used both a sociological and a historicalor
longitudinal dimension which constitutes a major characteristic of our researches. So
the exploitation of data stems fram two main axes: a continuous classification over
time and a final synthesis once we had collected our data. We were thus able to de-
duce our explicative reasoning and hypotheses fram the comments our interlocutors
took the time to share with uso
2.2 Main Results in Companies Studied

The five companies we have studied all belong to the informatic sector. The decision
to retain one specific sector and these companies for study resulted both from op-
portunities as they presented themselves but also and mainly from hypotheses of
research. Our initial contacts and exploratory interviews had indeed enabled us to
note that in this sector different companies were launching Quality Programs, nearly
at the same time and above all they were referring to the same Quality design. We
had, at first, the hypothesis that the model would shape Ihe change process and
would produce similar results. Besides all these companies had Ihe backing and ex-
pertise of consultanl agencies which, at the beginning, reinforced their efforts. The
chronological order for implemenling the quality program was also similar in all the
different firms.
By Ihe way may we mention thai when we speak of a company we refer more pre-
cisely 10 one speeific plant in each if these companies. We must also add that Ihese
companies, beside being part of the same sector, were characlerized by different
histories and different cullures. However all of them referred 10 Ihe same basic
model (see figure 1) which has been widely diffused by (Crosby 1979).
At the end of our sludy what is really is striking to note is that a !wo-fold trend ac-
companied the launching and developing of quality programs. AI Ihe beginning, all
companies referred 10 Crosby's 14-Point Program for bolh defining total quality, for
Ihe structure of implementalion and Ihe necessily 10 inform and train personnel.
However, when we look more closely in lime, Ihe diversity of practices used attracts
Ihe observer's attention. Despite the relative ecumenical character of their struclure
and beyond Ihe orthodoxy which defines Ihem, qualily programs lurned oul in li me a
relalively flexible formula for some companies whereas Ihe others sticked to the ini-
lial model. Over time the situations diverged: three of Ihe companies sludied
moulded the inilial Crosby design and adapled it 10 their own specificities, whereas
the olher !wo remained attached 10 the logic of Ihe normalive model and tried to in-
stitulionalize it.
For example, as soon as Ihe start ups began, some companies drew up a very pre-
eise and rigorous charter presenling "Iolal quality prineiples" and "principles for Ihe
participanls" slating more or less Ihe righls and obligalions of participanls. These
messages were communicated in-house by films, booklets and posters.
258 Franr;oise Chevalier
1. Management commitment. Top management must become convinced of the

need for quality improvement and must make its commitment clear to the en-
tire company. This should be accompanied by a written quality policy, stating
that each person is expected to "perform exactly like the requirement, or
cause the requirement to be officially changed to what we and the customers
really need."
2. Quality improvement team. Management must form a team of department
heads to oversee quality improvement. The team's role is to see that needed
actions take place in its departments and in the company as a whole.
3. Quality measurement. Quality measures that are appropriate to every activ-
ity must be established to identify areas needing improvement.
4. Cost of quality evaluation. The controller's office should make an estimate
of the costs of quality to identify areas where quality improvements would be
profitable.
5. Quality awareness. Quality awareness must be raised among employees.
They must understand the importance of product conformance and the costs
of nonconformance.
6. Corrective action. Opportunities for correction are generated by Steps 3 and
4, as weil as by discussions among employees.
7. Zero defects planning. An ad hoc zero defects committee should be formed
from members of the quality improvement program.
8. Supervisor training. Early in the process, all levels of management must be
trained to implement quality.
9. Zero Defects Day. A Zero Defects Day should be scheduled to signal to em-
ployees that the company has a new performance standard.
10. Goal setting. To turn commitments into action, individuals must establish im-
provement goals for themselves and their groups.
11. Error cause removal. Employees should be encouraged to inform manage-
ment of any problems that prevent them from performing error-free work.
12. Recognition. Public, non-financial appreciation must be given to those who
meet their quality goals or perform outstandingly.
13. Quality councils. Quality professionals and team chairpersons should meet
regularly to share experiences, problems and ideas.
14. Do it all over again. To emphasize the never ending process of quality im-
provement, the program must be repeated.
Fig. 1: Crosby's 14-point program

Source: Crosby (1979)
The cost of quality was measured as precisely as possible and the quality indicators
appeared very numerous. In these companies strategies and development plans
Qua/ity Management and Organizational Change 259
were drawn up. The development phases were to consistently integrate all the per-
sonnel. One hund red percent of the employees and executives were also to partici-
pate in the quality programs according to a predetermined calendar. These compa-
nies wanted to see quality programs branch out into all departments and services.
Likewise extensive !wo or three-day training programs in total quality were set up for
all the members of the company. Every employee and executive had to attend the
training sessions. These companies attached great importance to team work, at the
expense of all other previous managerial features, even if the previous organiza-
tional cultures were quite different. Along the same line, management presentations
were organized and most of the time became real media events. Public, non-
financial appreciation was presented as the only "contribution-retribution" contract
tor those engaged in the quality program.
The particularity of these companies was their willingness to keep up with the differ-
ent steps of wh at they considered as their quality model, with great precision, and to
impose it on their organization. Initially the programs were regarded by their partici-
pants as highly relevant to the companies' needs. However, they failed to change
the companies. According to interviews three years after the programs began, de-
spite enthusiasm trom the first participants, many found it difficult to transfer the
knowledge and interest developed by the programs into action. In fact, these com-
panies did not get the results they were targeting, either in terms of decreasing their
quality problems or in committing their staff. After the first success stories (decrease
of defects, respect of dealing or improvement of customer service) the next results
obtained were perceived "Iess importan!" by the top management. The structure did
not also work as planned and several quality councils did not meet on a regular ba-
sis, leaders no longer followed their schedules and postponed their meetings. Qual-
ity meetings were no longer considered as priorities by participants. The global im-
portance of participation in these quality programs showed considerable disparity
among departments and a certain diversity in the different categories of personnel.
Entire divisions of companies remained recalcitrant to the program. For instance
quality programs had difficulties expanding in the marketing sector, as weil as in
certain production shop floors or administrative sectors. Differences also deepened
be!ween those engaged in quality groups and others. The latter reproached their
260 Fram;oise Chevalier
colleagues for disturbing production with their meetings and wanting to "be noticed
by the boss" in management presentations. Many people also claimed to be torn
between individual performance and team work, between their traditional style of
directive management and the new participative one introduced by the quality pro-
gram. Managers also feit caught between what they described as their daily impera-
tives and quality teams, between productivity and quality, torn between two contra-
dictory systems.
Confronted with these problems, surprisingly these companies reinforced their initial
choices. They reasserted their quality motto, reinforced compulsory training ses-
sions, and stressed team work as the only privileged way. As paradoxical as this
may seem, instead of resolving the problems encountered and improving the situa-
tion, these adjustments only made matters worse and opponents to the programs
were more and more numerous. Whereas the programs were supposed to be long
term ones, being part of a never ending process, three years later they were partly or
even totally abandoned by disappointed managers.
Another form of change is represented by the three other companies which still have
a quality program and whose management declare to be 'satisfied with the results
obtained either interns of quantitative or qualitative indicators. These companies
U
also referred to the same basic quality model but unlike the previous companies,
their approach over time proved more flexible and open. It is remarkable that these
companies started the operation with a few basic principles on quality which were
constantly repeated. But they didn't pay much attention to the Crosby's 14 steps and
to their chronology, unlike other companies. For instance, no official charter defined
the different steps of the program. Nevertheless these companies also experienced
difficulties with the initial quality structure which was rapidly perceived as a "parallel
structure by managers. That's why instead of keeping it in its initial state, the quality
U
improvement team was finally included in the companies steering committees. Qual-
ity training was encouraged and all levels of management invited to participate but
this was never presented as a compulsory rule. Quality groups, such as QCS were
also highly recommended but never imposed; likewise, no quantitative objective for
their branching out was ever provided. For instance as time went by these compa-
nies went on diversifying the concept of team work, trying to adapt it to local condi-
tions. Many different kinds of team work were used. There were homogeneous or
heterogeneous groups chaired either by supervisors or peers. Sometimes employ-
ees were also at the head of problem-solving groups. Moreover, depending on the
subjects, leaders were sometimes appointed by top executives, or co-opted by co 1-
leagues. Ad hoc teams were also formed and widely developed, as were more
regular quality teams. For those not in favor of such participative management, and
who resisted team work, but who were in favor of improving corporate quality, other
means were used, such as individual suggestions through suggestion boxes or sug-
gestion forms. Quality audits, Le., a much more directive approach, were also con-
ducted in some of these sectors. In this way efforts to introduce and develop quality
met with much more commitment from different people and different sectors. The
different systems coexisted and tensions were reduced. Likewise, these companies
integrated their quality programs into their daily operations by modifying their per-
formance appraisal systems, and by example by mentioning involvement in quality
action as a new variable. Individual performance and team work were both taken into
account, as were quantity and quality. Here the problem was to be able to address
different kinds of issues, and not only a new one at the expense of the previous
ones. It was not a question of managing either only quantity or only quality, but of
tackling both.
The study of these companies highlights the reference to a common model at the
beginning and then very different practices and outcomes. How to explain these dif-
ferences? Let us recall that at the beginning of our research we had the hypothesis
that referring to the same model the organizations would undergo similar change
process and get comparable results. Trying to explain these outcomes led us to ex-
plore a field of explicative conjectures stemming from normative literature that pres-
ents requisites for successful TQC programs (and thus underscoring the "negative"
factors which cause failure of programs) as weil as articles which explain the suc-
cess or failure of programs (WilkinsonIWillmatt 1995). Two main types of explana-
tions are usually put forward. The first one concerns management involvement and
the second one the economic situation of the company. The observations made
during our field studies show that quality integration was often accompanied by
management backing, while quality disbandment was accompanied by managerial
262 Franqoise Chevalier
disinvolvement. But such results look like tautologies and do not really explain why
corporate managers persevere in the quality program while others fail to do so. A
second hypothesis, often mentioned, is based on the corporate economic and finan-
cial situation. According to this perspective, asound economy leads to the program
maintenance, as management accepts to continue to invest and back the program,
while, on the other hand, economic difficulties can lead to budget restrietions and
even the total elimination of the programs. By examining the five companies we
found that the economic situation was not necessarily linked to the program success
or failure. The cause and effect relation appeared not to be a determining factor.
Finally what really seems to be discriminating, in our analysis, are the conceptions of
change present in each organization. For the first category of firms, the logic which
prevailed was the continuation of the initial approach despite all the difficulties. This
is not to state that a new corporate philosophy, training, or changes in organizational
structure are always inappropriate. All can play valuable roles. The problems come
when such programs are implemented, notwithstanding the complexity of organiza-
tions, and are used in isolation as a kind of "magie bullet" to spread organizational
change rapidly through the entire organization (Beer/EisenstaUSpector 1990). The
change process in the companies which abandoned the program was geared toward
the strengthening and institutionalization of exclusive measures. Crosby model was
recognized as being valid and had to replace former operational modes. The rea-
soning used here was normative and exclusive. Those initiating change wanted to
help others without reaily asking for their opinions. It enlightened a deliberate intent
to stick to the initial framework, without really taking into account previous organiza-
tional designs, existing systems and questions or oppositions from people. Thus the
lack of involvement in quality programs was viewed as a problem of resistance to
change by people whose habits served as blind conservatism. In order to overcome
such resistance, constant persuasion, information and training was used. The type of
change endorsed by these companies is one which altempts to channel and melt
actors into the same mold. The lalter are given one alternative: integration or exclu-
sion. This is a frontal make or break model, offering conversion or excommunication.
Opposition does not go unheard, but when it is taken into consideration, it is both to
refuse its legitimacy and to reaffirm the initial conceived model.
For the second category of companies change was achieved by progressively

building or formulating new behavioral trends or operational rules. Rather than try to
convert opponents to the new approach, these companies took their opposition into
consideration. Quality programs were implemented in a much more flexible way.
Former models and organizational designs and systems were not entirely replaced
by new ones, but attempts were made to implement complex systems. The approach
was quite eclectic. In these companies managing quality implied managing diversity.
This diversification strategy is interesting for more than one reason: on the one
hand, it avoids a break with the past and limits negative reactions among certain
workers and managers; on the other hand, it is more apt to take into consideration
the variety of micro-cultural motivations and specificities linked to individual charac-
teristics and working conditions. Such approaches take into account the heteroge-
neous motivations of people at work that is underscored by several researchers.
They are based on segmentation and require different organizational designs in cor-
relation with different levels of involvement in work. Reexamining management from
this point of view is not always an easy task, as both managers and personnel have
been accuslomed 10 receiving unilateral messages in the pas!. These companies
continued to assert their quality policy, bul were quite open as far as tools, tech-
niques, and means were concerned. This relales to the important concept of requi-
site variety (ConanUAshby 1970). The law of requisite variety "states that the variety
within a system must be at least as great as the environmental variety against which
il is attempling to regulate itself. Put more succinctly, only variety can regulate vari-
ety" (Burckley 1968). The reasoning here is sociological and multitudinous; rather
Ihan choosing between two aspects of a given alternative, both are incorporated 10
form just one. Dilemmas and contradictions are recognized as natural and inherent
parts of management procedures. A hitch is fused into an intricate corporale stralegy
rather Ihan simply done away with. No dogmatic or definite ostracizing mechanism
exisls here-simply a new dynamic trend which develops slowly but surely. Change is
dealt with as a process of successive apprenticeships which take into account oppo-
sitions, and evolves over time. It doesn't rejecl previous rules or behaviors for the
sake of new ones.
264 Franr;;oise Chevalier
Finally, the trick seems to be able to manage the switch from the end cf one kind of
organization to the start of another. And when ccmpanies try to manage this switch
in a revolutionary perspective, putting forward new models, longing to impose them
withcut taking into account the previous ones and the diversity of the organization,
they run out of steam. Whereas companies which emphasize change in an evolu-
ticnary and swirling perspective, get better results.
3 Managing Change
What is in question here is not so much the nature of change as the implementation
of change itself. It is not solely "what you change," it is also "how you change." In
other words, we mean the nature of change does not necessarily determine the
change process.
Actually, once the nature of change is decided, it may be managed in various ways.
Many change implementations are prescriptive and normative: they are based on rules
and steps. 'They are based on strategies which are explicit, purposefully develcped, and
planned in advance. And this often tums out to be both unrealistic and not empirically
representative of managed change efforts. As with organizational strategy, both "in-
tended" and "realized" (Mintzberg 1976), managed change has both intentional (explic-
itly planned for) and realized (emerged out cf the situation) aspects.
3.1 Contrasting Assumptions About People and

Organizations
Much of the planned change carried out by organizations in this century, including
most contemporary work, falls into this category. In most cases, the change is
worked out by "technocrats", then legislated from the top and implemented through
the hierarchy.
This approach to the change process usually stems from underlying assumptions
about the nature of people and organizations Schein (1988) has classified them as
fallows, and it is interesting to note that the categories follow each other in a sort of
historical procession, starting from the time of the industrial revolution.
+ The rational economic assumption. People are primarily motivated by

economic needs.
+ The social assumption. People are essentially motivated by social relation-
ships. Issues of leadership style and group behavior are therefore of great
importance.
+ The self-actualizing assumption. People seek to be mature on the job and
are capable of being so. External controls and pressures are likely to be
seen as reducing people's needs for autonomy and therefore will affect their
motivation. Given a chance people will voluntarily integrate their own goals
with those of the organization.
To a certain extent man is treated as a semi-conscious mechanism responding to a

variety of impulses in a somewhat predictable way. These assumptions have given
rise to a set of prescribed organizational models and designs such as Scientific
Management, Job Enrichment, Autonomous Work Groups, Participative Manage-
ment and nowadays to Quality Management. When properly applied, these models
are supposed to produce productivity, job satisfaction, high performance and quality.
Here organizations and people are viewed as presumably rational, accountable, or-
derly and rather predictable units. Organizing is assimilated to using solutions.
These assumptions appear excessively simplistic, each implying a universalistic
"one best way" approach (figures 2 and 3).
Human Nature
Motivation Theories
Motivation Theories in Application

Prescribed Organizational Models
Output
Fig. 2: Understanding people and organizations through normative approach

• People are motivated by specific motivation faetors.

Organizafion can be mode/ed according to these factors.
• There is a linear mechanica/ relation between people and organization.
~
"ready to wear" solutions approach
~
"managerial fads"
Fig. 3: Nonnative and didactical approach of people and organization
On the contrary, organization can be conceived of as clusters of people joined by a

great variety of links (Tichy 1983). People are both formally struetured (prescribed),
such as in departments, in a f10w chart, and informally struetured (emergent), such
as coalitions. Prescribed organizations are typically represented in organizational
charts. Clear distinetions are made here between prescribed and emergent organi-
zations to emphasize the point that, within organizations, there exist a multitude of
interpersonal work arrangements which arise out of many possible types of relation-
ships. Only a portion of the organizational structure is prescribed. Thus, unplanned
structures and behavior patterns gene rally emerge in all organizations.
Organizations develop extensive informal or emergent structures and processes as

a result of human interaction. These networks of relationships and processes
emerge because individuals with their "bounded rationality" (March/Simon 1958)
tend to formulate, reformulate, and interpret the mission: understand, abide by,
and/or change the prescribed organization and organizational processes; use,
abuse and alter the technology; and differentially respond to changing environmental
conditions. As a result, a new set of unplanned and often unanticipated structures
and processes emerge (see figures 4 and 5).
I \lI
Human beings and behavior Organizations
• Diversity of changing Motivation • Each organization differs from
one another
• PersonnellGroup Objectives and • Each organization is both formally
PersonnellGroup Strategies structured (prescribed) and infor-
mally structured (emergent)
• Bounded Rationalitv
!)\ I
Fig. 4: Underslanding people and organizalion Ihrough a conlingenl approach
• People are motivated by very different means. Their motivations change

over time. They develop individual and group projects. They have strategies
and "bounded rationality."
Organizations cannot be entirely modeled. Each organization is different
from one another. Each organization is both formally structured (prescribed)
and informally structured (emergent).
• There is a dialectical organic relation between people, organization and en-
vironment.
"problem-focused" approach
~
"custom-tailored" organizations
Fig. 5: Conlingenl approach of people and organizalions
3.2 Contrasting Assumptions about Change

Im plementation
Given these views of people and organizations, the implications for the implementa-
tion of change are quite profound (see figure 6). The implementation of changes will
be based on the manager's intuitions and more or less implicit models of organiza-
tion. Presumptions about people and organizations are at the core of all organiza-
tional change. They provide guidelines for selecting diagnostic information and for
arranging the information into meaningful patterns. For a manager or consultant,
these assumptions funciion much like a physician's model of the human system. The
physician conducts tests and collects vital information with the model as a guide.
After the diagnosis is made, the model guides the selection of appropriate medical
intervention. In a similar fashion, the organizational model is used to guide the con-
duct of a diagnosis and the development of change strategies (Tichy 1983).
Many companies are still using a mechanistic and linear model, first applied to man-
aging physical work, and superimposing it into the new mental model of today's or-
ganization of knowledge. It is "the fallacy of programmatic change" (Beerl
EisenstatlSpector 1990). Most managers get into trouble because they literally for-
get, to think in circles. They continue to believe that there are such things as unilat-
eral causation, independent or dependent variables, origins and terminations. Ex-
amples are everywhere: leadership style affeets produetivity, parents socialize chil-
dren, stimuli affect responses, ends affect means, desires affect aetions. Those as-
sertions are wrong because each of them demonstrably also operates in the oppo-
site direetion: produetivity affects leadership style (Lowin/Craig 1968), children so-
cialize parents (Osopky 1971), responses affect stimuli (Gombrich 1960), means
affeet ends (Hirschman/Lindblom 1962), aetions affect desires (Born 1967). In every
one of these examples causation is circular, not linear. And the same thing holds
true for most organizational events (Weick 1979).
Whereas "programmatic strategy" for change implementation is based on conven-

tional knowledge, assumptions and overlooking the reality of human transaetions,
"heuristic strategy" takes into account the functioning and the regulations of social
systems. It is particularly charaeterized by its ability to deal with tensions and con-
tradictions. In fact it views them as inherent to organizations. A heuristic change pro-
cess appears as a kind of ambivalent activity which bears some resemblance to the
process of Janusian thinking defined as "actively conceiving !wo or more opposite or
antithetical ideas, concepts, or images simultaneously" (Rothenberg 1976). This
style of thinking, which Rothenberg associates with creative activities, was named
after the Roman god Janus, the god with !wo faces who perceived the world in oppo-
site direetions simultaneously.
The formalized, standardized strategies for a change process aet as a kind of ex-
ogenous force whereas the heuristic strategies for a change process act much more
as endogenous forces which allow for differences and take into account the charac-
teristics and specificities of each company. The latter change implementation proc-
ess is a swirling, endless one.
Programmatic Strategy Heuristic Strategy
• conventional knowledge and as- • the unpredictable reality of human

sumptions on organizations transactions
• looking at organizations as • looking at organizations contingent

modelized torms torms
• putting forward preconceived solu- • stressing the importance of diagnosis

tions and problems
• insisting on rules and procedures, • insisting on change as an open and

tool boxes flexible process
• rejecting contradictions • recognizing and managing contradic-

tions as natural mutual adjustments
• managing procedures and pieces • managing a dynamic and an on going

process
Fig. 6: Contrasting assumptions about change implementation
3.3 Taking into Account Contradictions and Di-

lemmas: "Hybrid Organizations"
Here, those in charge of change navigate continual tension between rational and
organic approaches, mechanistic and organic solidarity, individualism and commu-
nalism. In that sense companies appear to be hybrid, or complex as stated by
Morgan (1986) and Guillen (1994):
"We live in a world that is becoming increasingly complex. Unfortunately,

our styles of thinking rarely match this complexity. We oflen end up per-
suading ourselves that everything is more simple than it actually is, deal-
ing with complexity by processing thai il does not really exist. This is very
evident in Ihe way fad and fashion dominale approaches to organizational
analysis and problem-solving, an inleresl in one type or set of techniques
quickly giving way 10 another. [... ] Organizations are generally complex,
ambiguous, and paradoxical. The real challenge is 10 learn to deal with
complexity" (Morgan 1986).
"In the murky world of ideological and technical practice, however, the
most adequate metaphor seems to be one of layers or sediments of dif-
ferent organizational paradigms, not one of periodically dominant para-
digms that totally replace old approaches and, in turn, disappear com-
pletely with the success of the next paradigm" (Guillen 1994).
The inability of people in organizations to tolerate equivocal processing may weil be

one of the most important reasons they have trouble. It is their unwillingness or in-
ability to meet differences, contradictions and dilemmas that most of the time pro-
duces failure, non adaptation, psychological costs, problems, and so on. Organiza-
tions continue to exist only if they maintain a balance between flexibility and stability.
Jones and Gerard describe similar ideas by the concept of "basic antinomy". The
"basic antinomy" is between openness to change and the desire to preserve past
wisdom. 80th processes are necessary.
Companies where quality programs are successful appear as "hybrid organizations"

(see figure 7). They combine very different aspects which are usually presented as
contradictory, such as centralization and decentralization, short-term orientations
and long-term orientations, hierarchical management and management of peers,
quantity and quality. They take into account traditional hierarchical design while en-
couraging empowerment and self-organization. All these dimensions do exist in or-
ganizations, the problem is not so far to reject one at the expense of the other; the
problem is to embrace both, even if one is, at a certain time, more stressed than the
other. You don't do one or the other absolutely; you do what is appropriate.
One could say the companies might experience entropy, or such a disorder that it
would weaken thern. Actually the companies which have long successful quality pro-
gram have minimized this risk by integrating all their different forms of management
into their existing management and strategie system. In this way, managers can un-
lock many of the paradoxes of management and handle the inherent ambiguities.
There it appears to be important to look not for just one correct answer, but for many
correct answers over time within the same reinforced quality goals which serve as a
kind of magnetic force.
Quality Management and Organizatianal Change 271
Ta view organizations as "hybrid forms" appears 10 speak much more 10 the day-to-
day realities of organizations and reveals a better understanding of their real way of
functioning.
• prescribed slructure and emergent structure

quality and quantity
hierarchical management and management of peers
• centralization and decentralization
• directive leadership and democratic leadership
• short-term orientations and long-term orientations
• lightly controlled slructures and low controlled slructures & uniformity
and diversity
• formal and informal
slraight forwards and melaphors
•
Fig. 7 Taking into account contradictions and dilemmas "Hybrid Organizations"
4 Conclusion
What is most striking is definitely Ihe importance of the change implementation con-
ceptions and Iheir structural power on Ihe results of quality programs.
The change implementation conceptions appear to be twofold: one is rigid while Ihe
other is flexible, one allows for adjustments while the other does not, and one
authorizes the appropriation of change by different groups of actors while the other
does not. Although the normative model used by all companies at the outset gener-
ates enthusiasm, it is the way in which each company progressively modifies the
model, the specific manner or the specific change process in action which finally
generates the differences in results. The integration of quality programs is thus es-
sentially conditioned by Ihe degree of "openness" of the change process, its possi-
ble articulation within pre-existing management systems, and its capacily 10 incorpo-
rale differences or conlradictions.
Our research underlines, in particular, how a company's ability 10 apprehend and

manage change is deeply rooted in its assumplions on people and organizalion. Or-
ganizational models filter and focus attention on essentials.
Most change programs don't work as weil as they were supposed to because they
are guided by theoretical approach es to people, organizations and change imple-
mentation which are tao exclusive, thus, unable to cape with differences, variety and
antipodal requirements.
Quality Management and Organizationa/ Change 273
References
Beer, M./Eisenstat, R.A./Spector, B. (1990): Why change programs don't produce

change, in: Harvard Business Review, Vol. 68, No. 6, pp. 158-166.
Beer, M./Spector, B. (1993): Organizational diagnosis: Its role in organizational

learning, in: Journal of Counseling and Development, Vol. 71, No. 7/8,
pp. 642-650.
Berger, l.A./Sikora, M.J. (1994): The Change Management Handbook, New York.
Chevalier, F. (1987): Les cercles de qualita a bout de souffle?, in: Annales des
Mines, Garer et Comprendre.
Chevalier, F. (1991): From quality Circles to total quality, in: The International Jour-
nal of Quality and Reliability Management, Vol. 8, No. 4, pp. 9-24.
Chevalier, F. (1991): Cercles de qualita et Changement Organisationnel, Paris.
Chevalier, F./Segalla, M. (1996): Organizational Behaviour and Change in Europe -

Case Studies, London.
Chevalier, F./Boumois, F. (1998): Doing research with foreign Colleagues: a proj-'

eel life cycle approach, in: Journal of Managerial Psychology, Vol. 12, No. 3/4
pp. 206-213.
Cole, E.R. (1995): The Death and Life of the American Quality Movement, Oxford.
Crosby, P.B. (1979): Quality is Free, New York.
Crozier, M.lFriedberg, E. (1977): L'acteur et le systeme, Le Seuil.
Duck, J.D. (1993): Managing change: The art of balancing, in: Harvard Business
Review, Vol. 71, No. 6, pp. 109-118.
Friedberg, E. (1993): Le pouvoir et la regle, Le Seuil.
Guillen, M.F. (1994): Models of Management, Work, Authority and Organization in a

Comparative Perspective, Chicago.
Handy, C. (1993): Understanding Organizations, Oxford.
Hurst, D.K. (1984): Of boxes, buffies and effective management, in: Harvard Busi-
ness Review, Vol. 62, No. 3, pp. 78-88.
March, J.G./Simon, H.A. (1958): Organizations, New York.
Morgan, G. (1986): Images of Organizations, Newbury Park.
Morin, P. (1988): Pratiques du changement organisationnel, in: Revue Franc;:aise de

Gestion, No. 68, pp. 60-66.
Moss K.lStein, R.lBarry, A./Tick, T.D. (1992): The Challenge of Organizational

Change, New York.
Nutt, P. (1986): Tactics of Implementation, in: Academy of Management Journal,

Vol. 29, No. 2, pp. 230-261.
Pettigrew, A./Whipp, R. (1991): Managing Change for Competitive Sueeess, Ox-

ford.
Ouinn, J.B. (1980): Strategies for Change: Logieal Inerementalism, Homewood, IIli-
nois.
Schein, E.H. (1988): Organizational Psychology, 3d ed., Englewood Cliffs, New Jer-
sey.
Tichy, N.M. (1983): Managing Strategie Change. Technical, Political and Cultural
Dynamics, New York.
Tichy, N.M.lHomsteln, M./Nisberg, J. (1977): Organization diagnosis and interven-

tion strategies: Developing emergent pragmatic theories of change, in: Burke,
W.w. (ed.): Current Issues and Strategies in Organization Development, New
York.
Weick, K.C. (1979): The Social Psychology of Organizing, Reading, Mass.
Wilkinson, A./Willmatt, H. (1995): Making Quality Critical. New Perspectives on

Organizational Change, London.
Wilson, D.C. (1992): A Strategy of Change. Concepts and Controversies in the

Management of Change, London.
Sales-Related Benefit-Analysis of Service Quality Investments
- The Moderating Role of Customer Relationship Type
Manfred Bruhn and Dominik Georgi
Contents
Abstract 277
Introduction 277
2 Sales-Related Benefit Categories of Service Quality 278
3 Determining Sales-Related Service Quality Benefits 281
4 Customer Relationship Type as a Determinant of the

Repurchase Benefit 288
4.1 Effects of Relationship Strength 288
4.2 Effects of Relationship Length 291
4.3 Empirical Examination of the Effects of Relationship Type 292
4.3.1 Operationalization cf Variables 293
4.3.2 Results 293
5 Implications 296
References
Sa/es-Re/ated Benefit-Analysis of Service Quality Investments 277
Abstract
Service quality management is acknowledged as a major key to service provider's
economic success. In contrast to traditional marketing aetivities, it aims at generating
economic success by satisfying and retending current customers. However, as other
aetivities of service corporations, quality activities should be assessed against their
profitability. This paper presents an approach for quantifying the benefit of quality
activities generated by increasing customer satisfaction and customer retention.
Furthermore, because of several moderating faetors the effects of quality manage-
ment on customer behavior are not linear in nature. An important moderating factor
is the type of the current customer relationship. Therefore, this paper also analyzes
this construct's moderating effect on the relationship between customer satisfaction
and customer repurchase behavior.
1 Introduction
In theary and practice, there is a camman befieve in the service success chain. By
increasing custamer satisfaction higher customer retention is expected. Due ta the
higher number of custamers retended and cammitted tawards the service provider,
economic success should result from increased sales and lowered casts (Reich-
heldlSasser 1990, pp. 105-106; Stauss 1992; Heskett et al. 1994, p. 51). Examples
of companies which were not successful economicafly in spite af high quafity expen-
ditures seem ta disprove the universal validity of this chain (GranUShani/Krishnan
1994, pp. 25-26; RusUZahoriklKeiningham 1994, pp. 2-5). Furthermare, there are
common sense reasons against generafizing the hypothesized relationships. It
daesn't seem feasible that every quality measure leads ta higher custamer satisfac-
tion, higher customer retention and higher econamic success. Several moderating
factars are the reasan for the non-generafizability of quality acitivities' success
(Homburg/Faßnacht 1998; Bruhn 1998). Among athers, the type of customer rela-
tionship invalved impacts the profitability of quality activities. Strength and length af
customer relationships are major determinants af the service success chain.
278 Manfred Bruhn / Dominik Georgi
Therefore, quality measures should be understood as quality investments instead of

expenditures. In consequence, they should be assessed concerning their profitability
before and after carried out (Bruhn 1998). One crucial problem of measuring service
quality profitability in terms of a consistent cost-benefit-analysis is determining serv-
ice quality investments' benefits. In the services sector, the relative importance of
customer related benefits of service quality resulting in higher sales due to customer
satisfaction is immense (RusUZahoriklKeiningham 1995, pp. 58-59).
This paper first explains the sales-related benefit categories of service quality. Sec-
ond, an approach for determining service quality benefits is presented. Third, the
moderating role of customer relationship type for the satisfaction-purchase-
relationship - a core hypothesis in the service success chain - is analyzed.
For demonstrating the approach of determining quality benefits and to analyze the
effects of relationship type, data from an empirical study with the German Vereins-
und Westbank are applied. The Vereins- und Westbank is a bank that offers about
450.000 customers a wide range of services (Schreiber/Zimmer 1996, pp. 664-667;
Zimmer 1997). In the study, more than 10.000 customers were asked by telephone
interviews. The data of 9.342 respondents were considered in the analyses.
2 Sales-Related Benefit Categories of Service

Quality
In order to explain the positive economic consequences of service quality systemati-
cally, several categories of service quality's benefit can be distinguished. The
categories can be differentiated by two dimensions: the customer behavior that leads
to economic consequences and the direction of these consequences.
The relevant customer behavior can be categorized in dependence of the focal

customer whose behavior leads to the consequences. In general, the sales effect
can be generated by the behavior of a curren! or a potential customer. Current cus-
tomers' behavior leads to the retention benefit of service quality. Current customers'
behavior can further be differentiated by its impact on the customer relationship. Ex-
istence-related behavior refers to a relationship maintenance, while intenseness-
Sales-Related Benefit-Analysis of Service Quality Investments 279
related behavior concerns the modification of customer relationships' intenseness.

Potential customers' behavior concerns an indirect customer acquisition due to
word-of-mouth communications of current customers.
Concerning the consequences' direction a (enhanced) desirable and a (avoided)
not desirable component can be distinguished. The enhanced component specifies

the realization of the respective customer behavior. The avoided component con-
cerns the avoidance of a non-realization of the respective behavior. Thus, this di-
mension distinguishes (enforced) desirable customer behavior and (avoided) not
desirable customer behavior.
By means of these dimensions, six benefit categories of service quality result (see
figure 1). The major determinant of the underlying customer behaviors is customer
satisfaction which represents the degree of fulfilling the customers' requirements
concerning the service.
~
Current customers Potential customers
behavior
Retention benefit
Communication
benefit
Direction Existence-related Intenseness-related
(Enhanced) Repurchase Expansion Recommendation

Desirable benefit benefit benefit
Cancelling- Weakening- Waming-

(Avoided)
avoidance avoidance avoidance
Not desirable
benefit benefit benefit
Fig. 1: Sales-related benefit categories 01 service quality benelit
A repurchase benefit is generated when current customers intend to use the serv-
ice again in the same intenseness. By avoiding the corresponding not desirable be-
havior, Le. the customer's relationship breakdown, the service provider can realize a
breaking off-avoidance benefit. Satisfied customers are more probable to buy

services again than dissatisfied customers. Because of the high risk connected to
using services, satisfied customers are less willing to change their provider. How-
ever, dissatisfied customers are more motivated to look for another service provider
(FornelllWernerfelt 1987, p. 343; Woodside/Frey/Daly 1989, p. 8; Boulding/Kalral
Staelin/Zeithaml 1993, p. 23; SchmenglerfThieme 1995, p. 131; Liljander/Strandvik
1995, pp. 51-53; Nader/Johnson/Bühler 1996, pp. 706-707).
If current customers even plan to intensify their relationship to the provider, an ex-
pansion benefit is generated. The sales due to a relationship expansion can be the
consequence of a higher sales volume or a higher price realized. The volume eftect
can be generated by cross buying behavior and/or by higher buying frequencies of
satisfied customers. Empirical results prove the role of customer satisfaction for
cross buying intentions (Bruhn 1998) as weil as for increased buying frequency in-
tentions (Danaher/Rust 1996). The price eftect results from a higher price tolerance
of satisfied customers. The relationship between customer satisfaction and price tol-
erance also could be shown empirically (Anderson 1996, p. 270; Zeithaml/Berry/
Parasuraman 1996, S. 41). While satisfied customers are probable to expanse their
relationships, dissatisfied customers not breaking down their relationship can be ex-
pected to at least weaken their relationships by doing less cross buying, lowering
their buying frequencies and/or being more price sensitive. Thus, by avoiding those
behaviors due to lowering customer dissatisfaction, the weakening-avoidance
benefit can be realized.
Indirect new customer acquisition - entailed by more positive or less negative word-
of-mouth communications of current customers - leads to a recommendation bene-
fit or avoiding-warning benefit. The more satisfied a customer is, the more prob-
able is his recommending the focal service provider to friends and colleagues
(Boulding et al. 1993, p. 23; Diller 1997a, p. 37; Bruhn 1997). This relationship could
be shown empirically (Bruhn 1998). However, dissatisfied customers can be ex-
pected to warn their friends and colleagues concerning the focal provider (Singh
1988, p. 93; CroninlMorris 1989, p. 42; BlodgettIWakefield/Barnes 1995, p. 34).
Since services are intangible in nature and hard to judge before consumption, word-
of-mouth-communications playa major role in service buying decisions
Sa/es-Re/ated Benefit-Analysis of SeMee Quality Investments 281
(George/Berry 1981, p. 53; Zeithaml/Berry/Parasuraman 1993, p. 9; GierliStichl

Strohmayr 1997, p. 30).
3 Determining Sales-Related Service Quality

Benefits
The described customer behaviors build the basis for determining service quality
benefits. The determining proeess contents of three steps:
(1) determining the volume component of benefits,

(2) determining the value component of benefits and
(3) attribution of benefit figures to service quality.
(1) Detenmining the Volume Component of Benefits
The volume component represents the number of customers who show the de-
scribed benefit related behaviors. The data needed to determine the volume compo-
nent can be gained by customer surveys about customer satisfaction on the one
hand and behavioral intentions on the other. Volume components for the different
benefit categories obtain four elements: number of current customers, rate of
(dis)satisfied customers, benefit rate for each category and benefit degree for each
category (see figure 2).
Since only the current customers can judge a provider's service quality, the total
number of current customers is the basis of the customer behaviors to be ana-
Iyzed. Even the communication benefit which is finally realized by potential custom-
ers' behavior cannot be generated without the increase (decrease) of current cus-
tomers' recommendations (warnings).
Total Rat.af
aenofit Nwnberof aenofit
numb.,ot (dis)satis- BenefitdegrH Exptlnation of v.ri_bles
c:ategory customers rolll
N, number of repurcn... customelS

Repurchase
-
·.
N number af customeJS
benef~ N, N s r. l'1II:edsatisfaction
'. repun::tll......
Cancelling- nuntlerd CMCltlling customers

avoidance
benef~
Nb N d bd - N
•
number of CUSlDm8f$
rite 01 dtsatilfacbon
CIIncdngrate
"
N, nu,,**, cf expansion custcmers
Expansion
N. N s e. d9.
,N numberofc:ustamers
rat,ofsatistaetion
benefit '.dg" ..nsion dI9'ae
ellplnsionratA
N" number of Mak«tlng customers

Weakening- N number of custom&rs
avoidance Nw N d Wd d9w • ra.
I'ItlofdtultilfKtion
benefit ....
w. 'NUkening
...... k..llng dIgree
N. number cf acquu..d cultonWrs
·"'.
N number cf customel$
Recommen· rateolsatisfadion
dation N. N s res cP s Pp '. recommendltion rat.
communlCltlan partners of
benef~
.. Slltisfied customeJs
prot.biIityGtpul'Chase
"'. nurnberofnataequiredcllSto,,*,
Waming-
avoidance
benefrt
Nn• N d wad CPd Pnp ...
N
•
"'.
nurnberafcustomers
rateotc:issabsfaction
waming ....
communication partners cf
... disutisfled cuslDmtls

ptOt:8bllity cf non-putehase
Fig. 2: Equations tor determining service quality benefrts
Furthermore, only the relevant behaviors of (dis)satisfied customers is to be consid-

ered. Therefore, the rate of (dis)satisfied customers is of major importance tor
isolating the benefit relevant customer behaviors. Customer satistaction is relevant
to the enhanced customer behaviors, while the rate of customer dissatisfaction con-
cerns avoided behaviors.
For each benefit category, customers with benefit relevant behaviors can be isolated
by the respective benefit rate. The benefit rate represents the percentage of cus-
tomers who behave as supposed by category definition. Thus. for the enhanced
customer behaviors, those satisfied customers who repurchase the service, expand
their relationship and/or recommend the company are to be examined. The respec-
tive benefit rates are repurchase rate, expansion rate and recommendation rate. For
the avoided customer behaviors, those dissatisfied customers breaking off or weak-
ening their relationships and/or warning against the provider are considered. The
respective benefit rates are breaking off rate, weakening rate and warning rate.
While existence-related behaviors - Le. repurchasing or breaking off a relationship -

are bipolar in nature, the other four benefit-relevant behaviors can differ in degree.
For example, when a customer repurchases a service, the company seils the same
service units to this customer as before. However, when a customer expands his
relationship, this expansion must not equal the unit (ar an amount dividable by this
unit). Thus, for the latter categories, the benefit degree has to be considered. For
expansion and weakening-avoidance benefits, the degree of sales modification with
those customers is considered. For the communication benefit, the benefit degree is
compounded of !wo elements. The number of communication partners of current
customers represents the number of persans the (dis)satisfied customer talks to
about the service provider. The behavior probability represents the probability by
which a customer with a recommendation (warning) buys (avoids buying) the focal
service.
By multiplying the three - for the existence related benefit - or four elements - for
the other categories - respectively, the total number of customers per category re-
sults.
Operationalization of Variables
Most of the data needed to calculate the volume component were obtained in the
empirical study with the Vereins- und Westbank (see figure 3):
• Customer (dis)satisfaction was found out by asking the respondents "How

satisfied are you with the bank overall?" on a three-point-scale with the
points "absolutely satisfied", "satisfied" and "dissatisfied". Since also "only"
satisfied customers often have the tendency not to show the described posi-
tive behavioral intentions (Stauss/Neuhaus1995), only the top box custom-
ers are counted for the rate of customer satisfaction. This rate is 69%, while
31 % of the respondents are less than "very satisfied".
• Concerning their repurchase intentions respondents were asked to reply to
the sentence "I would choose the bank as my bank again." on a three-point-
scale from "I absolutely agree" to "I don't agree at all". For absolutely satis-
fied customers, the repurchase rate is 85%.
• In order to find out the breaking off intentions respondents were asked
"Have you already thought about breaking off your relationship to the bank?"
on a two-point-scale ("yes" or "no"). For dissatisfied customers, the breaking
off rate equals 50%.
• Conceming positive word-of-mouth intentions respondents had to reply to
"I would recommend the bank to others." on a three-point-scale from "I ab-
solutely agree" to "I don't agree at all". 86% of absolutely satisfied respon-
dents agreed to the statement.
• Conceming negative word-of-mouth intentions respondents were asked
"If you ever had a problem with the bank, have you talked about it to oth-
ers?". 18% of dissatisfied respondents agreed.
Because not all of the necessary data were obtained in the study, some data in fig-
ure 3 are fictive in order to demonstrate the benefit calculation completely:
• For both the expansion rate and the weakening rate a value of 10% is as-
sumed. The benefit degree is supposed to be two thirds and will be consid-
ered in the next section.
• Conceming the number of communication partners, it is assumed that
satisfied customers talk to 10 persons about the bank, while dissatisfied
customers wam 20 persons. These data resemble values found in other
studies (Goodman/Grimm 1990).
• Finally, the probability of purehase of a customer with a recommendation
is assumed to be 2%, while the probability of non-purchase is assumed to
be4%.
By entering both the found and the assumed data in the respective equations the
numbers of customers for each benefit category can be quantified.
Sales-Related Benefit-Analysis of SeNice Qua/ity Investments 285
Not acquired customers (N".J :20 .088 16 .236
V.kIII CompcIIIIfIt ,~.:,.!" ~ . .~ :. t;:"~ ,...,iii:,LJ ·.;:t:.:c...c·~,~ ~'~~ f Beneflt figur.

Repurchase benefit
Repurchase customers 263.925 27:2.769

Profit per customer 150 DM 150DM
Absolute benefit figu re 39 .588 .750 DM 40.915.350 DM 1.326.600 DM
Canceling-avoidance benefit
Breaking off customers -69.750 -56 .375
Profit per customer 150DM 150DM
Absolute benefit figure -10.462 .500 DM -8.456.250 DM 2.006.250 DM
Expansion benefit
Expansion customers 31 .050 32 .090

Profit per customer 100 DM 100 DM
Absolute benefit figure 3.105.000 DM 3.209.000 DM 104.000 DM
Weakening-avoidance benefit
Weakening customers -1 3.950 -11 .275
Profit per customer 100 DM 100 DM
Absolute benefit figure -1.395.000 DM -1 .127.500 DM 267.500 DM
Recommendation benefit
Newty acquired customers 53.406 55.195
Profit per customer 1:20 DM 120DM
Absolute benefit figure 6.408.720 DM 6.623 .400 DM :214.680 DM
Waming-avoidance benefil
Not acquired customers -:20.088 -16.236

Profit per customer 1:20 DM 120 DM
Absolute benefit figure -:2.410.560 DM -1.948.320 DM 462 .240 DM
Total benefit 4_277.270 DM

I Ftom I,. Ihe number ot customers ls not lh. adu:l1one, beeause only satistac:tlon drtven c:ustorner nuctuatlon Is conskfered.
From t l • Öle number ot CU$tomers equalt N.. •. Nt. +N,

Fig.3: Determin ing quality benefits
(2) Determining the Value Component of Benefits
Not every benefit unit - e.g. not every service sold to a retended customer - can be
connected with the costs resulting from quality activities. This cost-related attribution
problem is due to the necessity of further activities by the service provider which are
not related to a specific quality investment (Bruhn 1998). For example, there are
general service processes necessary in order to realize the expansion benefit.

Therefore, the described customer behaviors have to be valued by a financial figure
that isolates the real revenue effect of those behaviors. This can be done by using
the profit per customer conceming the focal services.
Because of the heterogeneity of customer relationships, there are some difficulties

using one figure as the profit per customer for all customers. The type of customer
should rather represent a differentiation criterion for the profit per customer to
choose. Two sub criterions can be applied. First, customers should be segmented by
characteristics which are determinants of the customers' profit potential. Examples
for such characteristics are the customers' income, their financial assets as weil as
the number and type of services they use actually. Second, for the communication
benefit categories, different profits per customer should be applied than for the re-
tention benefit categories. Since a service provider has relatively high investments in
the beginning of a customer relationship, the profit per new customer is probably
lower than the profit per current customer.
By valuing the volume component by the profit per customer, the absolute benefit
figures result. These figures indicate which profits a company makes with customers
in the respective benefit category. Because of data protection reasons, the benefit
calculation for the Vereins- und Westbank is based on fictive profits per customer.
For current customers, a profit per customer of 150 DM is assumed. Because of the
benefit degree of !wo thirds for the expansion and weakening-avoidance benefits,
those profits per customer equal 100 DM. Profits per customer for new acquired
customers is 120 DM. The resulting absolute benefit figures for the six benefit cate-
go ries are shown in figure 3.
However, the sum of these figures does not represent the actual benefit of service
quality. To calculate the actual benefits, the absolute benefit figures have to be con-
nected to service quality activities.
(3) Attribution of Benefit Figures to Service Quality
Next to the cost-related attribution problem of service quality there exists an activity-
related attribution problem. Not every benefit relevant customer behavior can be
connected to service quality activities. For example, if a service provider didn't en-
gage in quality management, there still would be customers not breaking off their
relationships. Conversely, the absolute benefit figures cannot be attributed to spe-
cific quality measures. They rather are the result of all marketing activities in the
pas!.
Therefore, in order to isolate the major part of the irrelevant effects a time compari-
son is applied (Stauss/Seidel 1996). The core of the time comparison is the differ-
entiation of Iwo phases called "absence of activities" and "presence of activities".
The "absence of activities" represents the period before the quality investment is
realized, while the "presence of activities" begins at the investment time. By com-
paring the absolute benefit figures in the absence and presence of activities for each
category the total quality benefit can be calculated (Bruhn 1998).
If a service provider has the necessary data for Iwo points of time belween which a
quality investment was carried out, an ex-post profitability analysis can be under-
taken. By applying the described time comparison to this empirical data, the benefits
of the quality investments can be quantified.
If the relevant data is gained right before a quality investment, an ex-ante profit-
ability analysis can be realized. For this, several estimates about the development
of benefit relevant figures can be made. The main driver for higher quality benefits
as weil as the main objective of quality activities is an improvement of customer sat-
isfaction. Thus, assuming a certain improvement in customer satisfaction represents
the basis for an ex-ante profitability analysis of service quality investments. By en-
tering the target rate of customer satisfaction in the described equations, the abso-
lute benefit figures for the presence of activities can be calculated. By comparing the
absolute figures for presence and absence of activities, the relative figures resul!.
The sum of the relative benefit figures of all categories constitute the total expected
benefit of service quality.
In the empirical example, a target rate of customer satisfaction of 74 per cent is

assumed. Thus, the target rate of dissatisfied customers is 26 per cent. Supposing
the constancy of all other rates, the absolute benefit figures for the presence of ac-
tivities can be calculated (see figure 3). By summing up these figures, the total
benefit can be calculated and, in the example, is about 4.3 Mio. DM.
The dependence of service quality benefits from relationship type is to be demon-

strated exemplary for the repurchase benefit.
4 Customer Relationship Type as a

Determinant of the Repurchase Benefit
Although the positive relationship between customer satisfaction and repurchase
intentions is accepted in theory and practice, it depends on several moderating fac-
tors (Bruhn 1998). In order to build the basis for further analysis, this relationship
has to be proven for the current data set.
Hypothesis 1
H,: Higher customer satisfaction leads to higher repurchase intentions.
One factor that leads to variation in the satisfaction-repurchase-relationship is the

type of customer relationship. Two characteristics of a customer relationship are re-
lationship strength and relationship length (see figure 4).
4.1 Effects of Relationship Strength

In this study, relationship strength is defined as the relative priority role of a certain
service provider in the service source set of a customer. The service source set de-
scribes the set of service providers a customer buys a certain service from. Thus,
relationship strength is an involvement-related construct considering the competitive
environment of the service provider. Relationship strength has direct and moderating
effects on repurchasing behavior.
Relationship
strength
j + (H"I
Dlrect :- (H 2b)
effects
Customer + (H 1) + Repurchase
._------.,. Moderating
effects
satisfaction intentions
NegatIVe
(-) (if any) effect
- (H 3b)
Relationship
~(-I(H"I
length
Fig. 4: Hypothesized relationships between customer satisfaction, repurchase intentions, relation-

ship strength and relationship length
Several researchers propose that involvement is a necessary condition for brand

loyalty (Engel et al. 1990, p. 30; Peter/Olson 1996; Bloemer/Pauwels/Kasper 1997).
Reasons for this are a more extensive information search, emotional bonds and
higher awareness concerning attribute differences. The relationship between in-
volvement and loyalty could be shown empirically as weil as the relationship be-
tween involvement and commitment, an element of loyalty (Beatty/Smith 1987). De-
fining the concept of involvement in a brand instead of adecision process, the effect
of brand involvement on brand loyalty also could be shown (Bloemer/Pauwelsl
Kasper 1997). Loyalty can be seen as a multi-dimensional construct, namely con-
sisting of affective, cognitive as weil as conative consumer responses. In the context
service quality benefits, the conative responses, namely the repurchase intentions,
are of interest. Thus, if involvement in a brand has a positive effect on brand loyalty,
it can be assumed that it also has an effect on repurchase intentions.
Next to involvement in the brand, the competitive environment is seen as en ele-

ment of relationship strength. In a study about the antecedents of repurchase be-
havior's erosion on a market level, it could be shown that the brand leader in a sec-
tor has less repurchase erosions than competing providers (EasUHammond 1996, p.
168). Applying this brand order on a market level to a relative brand order on an in-
dividuallevel, it can be supposed that the relative brand order influences repurchase
intentions.
Since both elements of the defined construct relationship strength correlate posi-
tively with repurchase intentions, it can be hypothesized that there is a positive rela-
tionship between relationship strength and repurchase intentions.
Hypothesis 2a
H2a Relationship strength has a positive effect on repurchase intentions.
Since customer satisfaction is the major antecedent of repurchase intentions, rela-

tionship strength can further be seen as a mOderating variable for the satisfaction-
repurchase-relationship instead of a direct antecedent of repurchase intentions. The
strenger a customer perceives his relationship to the previder, the weaker should be
the relationship between customer satisfaction and repurchase intentions. Perceived
relationship strength can have several reasons, above all positive experiences with
the provider. These positive experiences, e.g. indicated by high customer satisfac-
tion, resulted in re-choosing the focal provider in the pas!. In the present, perceived
relationship strength indicates high involvement in the provider in his services and a
perceived superiority of the provider to its competitors. Thus, these factors' influence
on repurchase intentions leads to a weaker effect of customer satisfaction. However,
customers with weak relationship strength probably were more satisfied with other
providers. Consequently, they feel more involved in the competitors. If they are sat-
isfied by the focal provider in the present, they might overthink the perceived priority
role of this provider. At least, they will intend to repurchase its services. From this
follows that relationship strength should have a negative impact on the satisfaction-
repurchase-relationship.
Hypothesis 2b
H2b A satisfaction improvement of customers with a strong relationship to the

provider will lead to a lower improvements in repurchase intentions than a
satisfaction improvement of customers with a weak relationship.
4.2 Effects of Relationship Length

Relationship length simply is defined as the duration of a customer's relationship to
a provider. Common sense reasons and empirical evidence show that relationship
length positively effects sales (ReichheldlSasser 1990, pp. 105-106; Blattbergl
Deighton 1996, pp. 138-139). This argumentation is the basis for the claim, that
customer retention is more profitable than customer acquisition.
However, there is a lack of analysis of relationship length's effect on repurchase

intentions. In the special case of the membership duration in a customer club, it
could be shown that members with longer membership are readier to repurchase the
brand than members with a shorter membership (Diller 1997b).
But, a positive relationship between relationship length and repurchase intentions

would imply that a service provider didn't have to invest in marketing activities for
customers with a long relationship. There are common sense reasons against this
assumption. Thus, it is feasible that relationship length can't correlate positively with
repurchase intentions.
Hypothesis 3a
H3a The relationship length has - if any at all - a negative effect on repurchase
intentions.
In the context of this study, the moderating effect of relationship length on the satis-
faction-repurchase-relationship is of even more interest. Even though customer sat-
isfaction still effects repurchase intentions of customers with a longer relationship, it
is assumed that the satisfaction effect is weaker for those customers than for cus-
tomers with a shorter relationship. Reasons for this assumption can be found ana-
Iyzing both short and long relationship customers.
Short relationship customers can be differentiated by their experience with the

respective service category. If a customer has no (or less) experience with the cate-
gory, the focal provider is possibly the first provider he buys the service from. In this
case, a satisfied customers has no motivation to change the provider. If a customer
has already experience with the service category, he probably has changed the
service provider at least once. Furthermore, there are probably reasons for his pro-
vider change(s). In this context, it can be supposed that one major reason will be his
dissatisfaction with the other provider(s). Thus, when the respective provider per-
forms better than his competitors resulting in a high customer satisfaction, the cus-
tomer has no reasons to change the provider again. 80th arguments imply a rela-
tively high satisfaction-repurchase-relationship for short relationship customers.
Regarding long relationship customers stable and unstable customers can be dis-
tinguished. Stable customers made their decision to stay with the respective provider
in any case - unless they will be disappointed heavily and repeatedly. For those
customers, there are reasons to repurchase other than customer satisfaction, such
as trust and commitment. Even though such factors are influenced by (earlier) satis-
faction, the current role of customer satisfaction for repurchase intentions is rela-
tively weak. Unstable customers are variety seeking customers who always look for
variety in their service use. Even if a variety seeking customer is satisfied with a pro-
vi der he can switch to another provider. 80th arguments lead to the assumption that
the effect of customer satisfaction on repurchase intentions of long relationship cus-
tomers is relatively weak.
Hypothesis 3b
H3b A satisfaction improvement of customers with a shorter relationship to the

provider will lead to a higher improvement in repurchase intentions than a
satisfaction improvement of customers with a longer relationship.
4.3 Empirical Examination of the Effects of

Relationship Type
In order to analyze the formulated hypotheses empirically, data of the Vereins- und
Westbank study are applied again.
4.3.1 Operationalization of Variables

The customer satisfaction and repurchase intentions variables were operation-
alized in the way as described in the section "Determining the Benefit of Service
Quality". Data for both variables were obtained on a three-point-scale.
Furthermore, relationship strength was defined as a bipolar variable. Respondents

were asked which relationship strength they perceive by the quest ion "Would you
describe the bank as your main bank or a minor bank of yours?". If a respondent
views the bank as his main bank, a strong relationship is assumed, while customers
viewing the bank as their minor bank are interpreted as weak relationship customers.
Finally, relationship length was measured on a six-point-scale with the points "Iess
than one year", "1 to 3 years", "4 to 6 years", "7 to 10 years", "11 to 20 years" and
"21 years and more". For complexity reasons, this variable was re-coded to a two-
point-variable with the points "short relationships (up to 10 years)" and "Iong rela-
tionships (11 and more years)".
4.3.2 Results
For hypothesis testing, correlation analysis and two-sample significance tests were
applied.
Concerning the satisfaction-repurchase-relationship, a significant positive corre-

lation can be found. The correlation coefficient equals 0.57 and is significant on a
99% level (see figure 5, Part A). Thus, the positive relationship between customer
satisfaction and repurchase intentions can be accepted. Consequently, the basis
hypothesis H, can be accepted.
294 Manfred Bruhn /Dominik Georgi
Hypo- Part A: Correlation between customer satisfaction and repurchase intentions

thesis Customer group
Coefficienta
H, All customers 0.57
Weak relationships 0.60
H2b
Strong relationships 0.53
Short relationships 0.58
H3b
Long relationships 0.55
• for aD coeffidents pcO.01
Part B: Comparison of means

Hypo-
thesis Customer group Mean of
p F
repurchase intentions
Strong relationships 2.64
H,. 0.000 579
Weak relationships 2.33
Short relationships 2.61
HJa 0.000 15
Long relationships 2.54
Fig. 5: Direct and moderating effects of relationship strength and length
Next, the effects of relationship strength were examined. Direct positive effects of
relationship strength on repurchase intentions and negative moderating effects on
the satisfaction-repurchase-relationship were hypothesized.
For the direct effects, means of repurchase intentions were tested for significant
differences between the strong relationship customers - those for whom the bank is
the main bank - and Ihe weak relationship customers - those for whom the bank is a
minor bank. The respective results report significant differences (see figure 5,
Part B). While main bank customers have mean repurchase intentions of 2.64, minor
bank customers state repurchase intentions of 2.33. Levene's test for equality of
variances reports p = 0.000 and F = 579. Strong relationship customers have high er
repurchase intentions than weak relationship customers. Thus, hypothesis H2a can
be accepted.
Furthermore, a negative moderating effect of relationship strength on the satis-

faction-repurchase-relationship is assumed. In order to test this hypothesis, correla-
tion coefficients of the satisfaction-repurchase-relationship were compared between
Sales-Related Benefil-Analysis of Service Quality Investments 295
the main bank and the minor bank customers (see figure 5, Part A). Correlation
analysis shows that the minor bank coefficient is 0.60 (p < 0.01), while the main bank
coefficient equals 0.53 (p < 0.01). Although the significance of this difference cannot
be evaluated, hypothesis H2b seems to be acceptable. Thus, an improvement in sat-
isfaction of strong relationship customers might lead to lower improvements in repur-
chase intentions than a high er satisfaction of weak relationship customers.
Next to the effects of relationship strength, the role of relationship length was hy-
pothesized. Again, direct and moderating effects are to be examined.
While relationship strength has a positive influence on repurchase intentions, the

direet effeet of relationship length on repurchase intentions was assumed to be
negative - if there is any at all. Similarly to the relationship strength analysis, signifi-
cance of means differences of purchase intentions is examined. Dividing customers
in two groups in dependence of their relationship length, means of repurchase inten-
tions of short and lang relationship customers can be compared. While short rela-
tionship customers stated mean purchase intentions of 2.61, lang relationship cus-
tomers have purchase intentions of 2.54 (see figure 5, Part B). Levene's test for
equality of variances proves the significance (p=O.OOO; F=39) of these differences.
Thus, the longer a relationship lasts, the lower are the repurchase intentions. Hy-
pothesis H3• can be accepted.
Furthermore, the moderating effeet of relationship length on the satisfaction-

repurchase-relationship can be analyzed by a comparison of correlation coefficients.
For short relationship customers, there is a coefficient of 0.58 (p < 0,01) for the sat-
isfaction-repurchase-relationship (see figure 5, Part A). For lang relationship cus-
tomers, the respective value is 0.55 (p < 0.01). Thus, hypothesis H3b also seems 10
be acceptable. For customers with longer relationships, an improvement in customer
satisfaction leads to weaker improvements in repurchase intentions than for custom-
ers with shorter relationships.
Conclusively, relationship strength affects repurchase intentions positively, while

relationship length's impact on repurchase intentions is negative. Furthermore, the
296 Manfred Bruhn /Dominik Georgi
longer a relationship lasts and the stronger the relationship is, the weaker is the cor-
relation belween customer satisfaction and repurchase intentions.
5 Implications
A profitability-oriented service quality management intends to invest in quality
activities which lead to highest possible quality benefits. It could be shown that the
customer relationship type influences the satisfaction-repurchase-relationship. Only
considering the existence-related quality benefit indicating sales with satisfied and
retended customers, the profitability effects of quality activities vary according to the
customer relationship. Thus, a service provider should weight his quality invest-
ments according to the respective relationship type.
On the basis of the found results, management of short relationship and weak re-
lationship customers should focus heavier on customer satisfaction. For those
customers, the repeated fulfillment of customer expectations is of major importance.
Clearly, results should not imply that managing long and strong relationships
should not focus on customer satisfaction. But, there seem to be other drivers of re-
purchase intentions which have a relatively high importance in comparison to the
other relationship types. For lang relationship customers, a customer's repurchasing
decision might be affected by variety see king. Therefore, a provider has to da more
than satisfy their long relationship customers. In order to manage the repurchasing
behavior of those customers competitive factors should be identified, such as service
differentiation. Furthermore, a strong relationship customer's buying decision might
depend more on other constructs than satisfaction, such as commitment and trust.
Thus, for those customers, managing these constructs should have more weight
than for weak relationship customers.
Next to these management implications, there are some research implications to

consider. Even though the presented approach for determining service quality bene-
fits can show a tendency of the financial consequences of quality activities, there is
further research needed for the analysis of the relationships in the service success
chain.
In this paper, the link between quality investments and customer satisfaction
was operationalized by a time comparison. This method allows isolating most of the
effects of not quality-related measures established before the quality investment. In
order to isolate effects of not quality-related measures carried out after the quality
investment the method has to be extended. For example, the specific quality invest-
ment can be operationalized as a satisfaction driver. By integrating other satisfaction
drivers in the calculation, the improvement in customer satisfaction that is caused by
the quality investment can be determined.
Furthermore, the link between customer satisfaction and customer retention

should be analyzed more specifically. There are other company-driven determinants
of customer retention than customer satisfaction, e.g. customer credit cards. By inte-
grating these determinants in the calculation, the proportion of an increase in cus-
tomer retention that is due to customer satisfaction can be estimated more exactly.
Besides, the understanding of the link between customer retention and economic
success ought to be deepened. Concerning the current customers' behavior, not
only behavioral intentions but also real behavior should be integrated in the calcula-
tion. Regarding to the potential customers' behavior, research has to be done in or-
der to quantify the links between recommending/warning of current customers and
choosing/not choosing the provider by potential customers more precisely.
Finally, next to the examined effects of the relationship type on the satisfaction-
repurchase-relationship, there are other moderating variables determining the
whole service success chain, such as variety seeking intentions, the customers'
profit potential and the service range of the provider. Further analyses are necessary
in order to find more evidence about the influences of those moderating variables.
When there is more knowledge about these moderating effects, a customer behav-
ior-oriented customer segmentation can be established. Segmenting customers ac-
cording to the service success chain can help identifying those customers concern-
ing whom quality activities are most profitable.
298 Manfred Bruhn IDominik Georgi
References
Anderson, E.W. (1996): Customer Satisfaction and Price Tolerance, in: Marketing
Leiters, Vol. 7, No. 3, pp. 265-274.
Beatty, S.E.lSmith, S.M. (1987): External Search Effort. An Investigation across

several Product Categories, in: Journal of Consumer Research, Vol. 14,
pp. 83-95.
Blattberg, R.lDeighton, J. (1996): Manage Marketing by the Customer Equity Test,

in: Harvard Business Review, Vol. 74, No. 4, pp. 136-144.
Blodgett, J.G.iWakefield, K.L.lBarnes, J.H. (1995): The Effects of Customer Serv-

ice on Consumer Complaining Behavior, in: Journal of Services Marketing,
Vol. 9, No. 4, pp. 31-42.
Bloemer, J.M.M.lPauwels, K.H.lKasper, H.D.P. (1997): There is more to Loyalty

than just Satisfaction. The Effects of Satisfaction and Involvement on Brand
Loyalty and Dealer Loyalty, paper at the AMA-Conference in Dublin, Ireland.
Boulding, W.lKalra, A.lStaelin, R.lZeithaml, VA (1993): ADynamie Process

Model of Service Quality. From Expectations to Behavioral Intentions, in: Jour-
nal of Marketing Research, Val. 30, No. 1, pp. 7-27.
Bruhn, M. (1997): Qualitätsmanagement für Dienstleistungen, 2nd edition, Heidel-

berg.
Bruhn, M. (1998): Wirtschaftlichkeit des Qualitätsmanagements. Qualitätscontrolling

für Dienstleistungen, Heidelberg.
Cronin, J.J.lMorris, M.H. (1989): Satisfying Customer Expectations. The Effect on

Conflict and Repurchase Intentions in Industrial Marketing Channels, in: Jour-
nal of the Academy of Marketing Science, Val. 17, No. 1, pp. 41-49.
Danaher, P.J.lRust, R.T. (1996): Indirect Financial Benefits from Service Quality, in:
Quality Management Journal, Val. 3, No. 2, pp. 63-75.
Diller, H. (1997a): Was leisten Kundenclubs? Ein Testbeispiel, in: Marketing ZFP,
Val. 19, No. 1, pp. 33-41.
Diller, H. (1997b): Wh at Da Customer Clubs Achieve?, paper at the AMA-

Conference in Dublin, Ireland.
East, R.lHammond, K. (1996): The Erosion of Repeat-Purchase Loyalty, in: Mar-

keting Leiters, Val. 7, No. 2, pp. 163-171.
Engel, J.F.lBlackwell, R.D.lMiniard, P.W. (1990): Consumer Behavior, 6th edition,

Chicago.
Sales-Related Benefit-Analysis of SeNice Quality Investments 299
Fornell, C.; Wernerfelt, B. (1987): Defensive Marketing Strategy by Customer

Complaint Management. A Theoretical Analysis, in: Journal of Marketing Re-
search, Vol. 24, No. 4, pp. 337-346.
George, W.R.lBerry, L.L. (1981): Guidelines for the Advertising of Services, in:
Business Horizons, Vol. 24, No. 4, pp. 52-56.
Gierl, H.; Stich, A.; Strohmayr, M. (1997): Einfluß der Glaubwürdigkeit einer Infor-
mationsquelle auf die Glaubwürdigkeit der Information, in: Marketing ZFP,
Vol. 19, No. 1, pp. 27-31.
Goodman, J.A./Grimm, C.J. (1990): A Quantified Case for Improving Quality Now!,
in: Journal of Quality and Participation, Vol. 14, No. 3, pp. 50-55.
Grant, R.W./Shani, R./Krishnan, R. (1994): TQM's Challenge to Management The-

ory and Practice, in: Sioan Management Review, Vol. 35, No. 4, pp. 25-48.
Heskett, J.L.IJones, T.O./Loveman, G.W.lSasser, W.E./Schlesinger, L.A. (1994):

Dienstleister müssen die ganze Service-Gewinn-Kette nutzen, in: Harvard
Business Manager, Vol. 16, No. 4, pp. 50-61.
Homburg, C./Faßnacht, M. (1998): Kundennähe, Kundenzufriedenheit und Kun-

denbindung bei Dienstleistungsunternehmen, in: Bruhn, M.lMeffert, H. (eds.):
Dienstleistungsmanagement. Grundlagen, Konzepte, Erfahrungen, Wiesbaden
1998, in print.
Liljander, V./Strandvik, T. (1995): The Nature of Customer Relationships in Serv-

ices, in: Swartz, TA/Bowen, D.E.lBrown, S.w. (eds.): Advances in Services
Marketing and Management. Research and Practice, Vol. 4, London,
pp. 141-167.
Nader, G.lJohnson, M.D.lBühler, W. (1996): Kundenzufriedenheit und Unterneh-

menserfolg. Eine kausalanalytische Betrachtung am Beispiel des Privatkredit-
geschäfts von Universalbanken, in: Österreichisches Bankarchiv, No. 9,
pp. 702-710.
Peter, J.P./Olson, J.C. (1996): Consumer Behavior and Marketing Strategy, Chi-
cago.
Reichheld, F.F.lSasser, W.E. (1990): Zero Defections. Quality Comes 10 Services,

in: Harvard Business Review, Vol. 68, No. 5, pp. 105-111.
Rust, R.T./Zahorik, A.J.lKeiningham, T.L. (1994): Return on Quality. Measuring

the Financiallmpact of Your Company's Quest for Quality, Chicago.
Rust, R.T.lZahorik, A.J.lKeiningham, T.L. (1995): Return on Quality (ROQ). Mak-

ing Service Quality Financially Accountable, in: Journal of Marketing, Vol. 59,
No. 2, pp. 58-70.
Schmengler, H.J./Thieme, M. (1995): Die Bedeutung eines Bonusprogramms im

Marketing einer Luftverkehrsgesellschaft, in: Marketing ZFP, Vol. 17, No. 2,
pp. 130-135.
Schreiber, D./Zimmer, K. (1996): Effizientes Beschwerdemanagement durch EDV-

Unterstützung, in: Die Bank, Vol. 36, No. 11, pp. 664-667.
Singh, J. (1988): Consumer Complaint Intentions and Behavior. Definitional and

Taxonomicallssues, in: Journal of Marketing, Vol. 52, No. 1, pp. 93-107.
Stau55, B. (1992): Dienstleistungsqualität contra Kostensenkung?, in: Betriebswirt-

schaftliche Blätter, Vol. 41, No. 2, p. 111-115.
Stau55, B.lNeuhaus, P. (1995): Das Qualitative Zufriedenheitsmodell, Discussion

Paper No. 66 of the Department for Business Administration Ingolstadt, Catho-
lic University cf Eichstätt, Ingolstadt.
Stau55, B./Seidel, W. (1996): Beschwerdemanagement. Fehler vermeiden, Leistung

verbessern, Kunden binden, München.
Woodside, A./Frey, L./Daly, R. (1989): Linking Service Quality, Customer Satisfac-

tion and Behavioral Intention, in: Journal of Health Care Marketing, Vol. 9,
No. 12, pp. 5-17.
Zaichkowsky, J.L. (1985): Measuring the Involvement Construct, in: Journal of

Consumer Research, Vol. 12, pp. 341-352.
Zeithaml, V.A./Berry, L.L./Parasuraman, A. (1993): The Nature and Determinants

of Customer Expectations of Service, in: Journal of the Academy of Marketing
Science, Vol. 21, No. 1, pp. 1-12.
Zeithaml, V.A./Berry, L.L./Parasuraman, A. (1996): The Behavioral Consequences

of Service Quality, in: Journal of Marketing, Vol. 60, No. 2, pp. 31-46.
Zimmer, K. (1997): Kundenbindungsmanagement im Filialvertrieb. Umsetzung der

"IM-PULSE" in den Filialen der Vereins- und Westbank, Presentation at the
Summer University 1997 in Berlin, Hamburg.
List of Abstracts 301
List of Abstracts
National Customer Satistaction Indices: A Criticallnvestigation trom an

Application Perspective
Ursula Hansen and Thorsten Hennig-Thurau

become a central area of research in service management and marketing. Despite
an considerable amount on scientific and societal discussion, a massive lack of
meta-theoretical research in the area of NCSI development and usage exists. Based
on the identification of several levels of NCSI based satisfaction comparisons the
authors formulate a catalogue of general criteria which function as a frame of refer-
ence for the critical analysis and evaluation of NCSls. With the American Customer
Satisfaction Index and the German Customer Satisfaction Barometer, two well-
known indices are than evaluated with regard to the criteria pr.eviously worked out.
Several praposals are generated corresponding to the results of the analysis, which
can be used as a guideline for future improvements of NCSI methodology and us-
age.
Dynamics in Service Quality Perceptions - A Longitudinal Approach to

Evaluate the Outcome Dimension in Service Quality Perceptions
Claudia Klausegger and Dieter Scharitzer
During the past decade much conceptual and empirical research has been done re-
garding customers' perceptions of process- and outcome-related dimensions of
services. In analyzing a number of empirical studies, we noticed that most of the
research designs were based on crass-sectional measurement approaches. Those
attempts usually neglect the potential dynamics in the customers' evaluation of the
services delivered over time.
This article presents the results of an empirical longitudinal project based on data
collected at two different points in time. In our research an ex-post evaluation of the
educational quality of an Austrian Business College was done by means of impor-
tance and performance ratings fram the students' point of view at two different times
after finishing the undergraduate course.
302 List of Abstracts
The Link between Customer Satisfaction and Dealer Satisfaction - The Case of
the German Car Industry -
Alexander Decker
The satisfaction of car dealers with their suppliers has gained same remarkable at-
tention in the German car industry in recent years. A major reason for this phenome-
non may be seen in the apparent link between dealer satisfaction and customer sat-
isfaction. Although this relation seems apparent, the positive link between dealer
and customer satisfaction is not proven. For example, there exists no empirical evi-
dence for the causality of the relation. Therefore, the goal of this paper is to gain
some theoretical insights into the relationship on the basis of preliminary findings
and additional industry related considerations.
The findings of the analysis point out, that on the basis of preliminary empirical find-
ings there is only a weak support for the assumption of a positive relationship. In
addition, the direction of the connection is yet unclear. According to a more theoreti-
cal viewpoint, it seems to be more likely that the two constructs are at least positively
related. An additional calculation will show, that none of the two constructs has a
higher impact on the other. Limitations of these findings are going to be shown as
weil as implications for future research in this area.
Criticality of Criticallncidents in Customer Relationships

Bo Edvardsson and Tore Strandvik
In this article the traditional service encounter oriented view on critical incidents is
challenged. The authors propose instead an approach based on a customer rela-
tionship view. The main research questions are: "In what way do critical incidents
affect the customer relationship? How critical are critical incidents? When and why
are critical ineidents critical?". A tentative framework for analyzing critical ineidents in
customer relationships is developed and presented. The framework is built on two
main contextual dimensions; (1) the situational context which can be divided into an
internal, actor-based and an external, environmental context and (2) the time dimen-
sion which captures the history, the present state and the future of the relationship.
Critical Incidents in Dutch Consumer Press: Why Dissatisfied Customers Com-

plain with Third Parties
Jan A. Schulp
Customer complaints published in the main Dutch consumers' monthly were studied
using the Critical Ineident Technique. Out of 750 incidents two sampies on brand
promotions and retail promotions of 52 and 137 items respectively were studied in
more detail.
Ust of Abstracts 303
Dissatisfiers in the incidents have been classified according to attributes of the pro-
motions and the employee behavior. Customer behavior is described and classified
in so far as it leads to complaining with the Consumentenbond. Dissatisfaction is
mainly about the intangible aspects of the transactions. An aggravating factor in
promotions is the enhanced expectation which is disconfirmed. Dissatisfaction is
caused both by attributes of the promotions and the interaction between employees
and customers due to negative customers affects and negative perception of service
fairness. Differences between the outcome in retail promotions and brand promo-
tions clearly reflect the difference in personal contact intensity. The incidence of
complaining with the Consumentenbond without previous complaining with the serv-
ice supplier is highlighted as a managerial problem. In designing a promotion, ethical
viewpoints should be taken into account.
Achieving Service Quality through the Application of Importance-

Performance Analysis
Robert Govers and Frank M. Go
A survey was conducted amongst different types of consumers to measure the mar-
keting effectiveness of the Dutch domestic tourism organization, The Stichting Toer-
isme en Recreatie AVN. Specifically the purpose of the present study is to investi-
gate how the AVN might improve its ability to manage consumer expectations and
determine which attributes the organization should devote its attention and re-
sources to, in order to improve, meet, or exceed consumer expectations. First, the
importance of attributes is being determined through a management survey. Next,
the performance is measured through multiple consumer surveys. Finally, the differ-
ences between importance and performance are analyzed in an effort to identity the
organizations strengths and weaknesses.
The Sequentiallncident Technique for Innovations (SITI) -

A Tool for Generating Improvements and Ideas in Service Processes
Alexander Decker and Harald Meissner
service innovations are usually generated in a haphazard way. There is a lack of
application of market research techniques in this area as weil. Therefore, the pur-
pose of the paper is to show the development and application of a research tool,
which should help generating service innovations systematically.
304 List of Abstracts
In order to fulfill theoretically deduced requirements regarding the special charac-

teristics of services and service quality, the Sequential Incident Technique was
modified into the Sequential Incident Technique for Innovations (SITI). This paper
describes the method in detail and gives deeper insights into the results of its first
application. Study findings suggest that especially negative, but also positive inci-
dents can trigger innovative suggestions from the customers' side. The development
of coherent service innovation concepts from these suggestions, however, remains
left to the research team. Implications for the future use of the SITI as a research
tool as weil as limitations of the study are discussed.

Martin Wetzeis and Ko de Ruyter
Call centers can be considered as rapidly growing communication channels that

companies use to be more customer-oriented. In many cases, temporary staft, em-
ployed by specialized temporary helping service agencies, is used within call cen-
ters. These agencies playa pivotal role in motivating and establishing desirable be-
haviors of the agenls. In this paper we examine exchange relationships of lemporary
call cenler agents with their agency as weil as Ihe intermediary working for the tem-
porary employment agency as antecedenls to job satisfaction and in- and extra-role
behavior. The results indicate that perceived intermediary support has a direct eftect
on in-role behavior and thai perceived organizational support has a direct relation-
ship with extra-role behavior as weil as an indirect impact through job satisfaction of
temporary call center agents.
For the relationship between intermediary support and in-role behavior it could be
surmised that working in call centers requires bolh technical skills and 'people' skills.
The lemporary work agencies generally provide training focusing on information and
communication technology as weil as training focusing on customer orientalion. The
intermediary plays a facililating role in taking care of the training needs of the tempo-
rary slaft. Furthermore, organizational support by the temporary help service agency
positively aftects extra-role behavior by the temporary staft or 'ta go the extra mile'.
For the relationship between the temporary help service agency and their clients, the
willingness of temporary staft 10 go beyond traditional job requirements may consti-
tute a crucial competitive advantage.
List of Abstracts 305

Franyoise Chevalier
Quality programs are becoming widespread in companies. Nonetheless their imple-

mentation is not always easy and often involves several difficulties. This paper pre-
sents the results of the research we have conducted in five companies.
Our research illustrates that very rigid, "standardized strategies" for change have led
to difficulties. On the contrary, "heuristic strategies" for change have given rise to
good results. These heuristic strategies for change implementation are essentially
characterized by trying to deal with a number of contradictory phenomena within or-
ganizations rather than attempting to suppress them. A Company's ability to appre-
hend and manage contradictions within organizations determines the development
of quality programs and their degree of success.
These different approaches to the change process usually stem from very different
underlying assumptions about the nature of people and organizations. Depending on
the way people and organizations are viewed, change implementation will be quite
different. Organizational models filter and focus attention.
Finally, it is not solely ''what you change" that is important; it is also "how you
change". In other words, the nature of change does not necessarily determine the
change process.
Sales-Related Benefit-Analysis of Service Quality Investments -

The Moderating Role of Customer Relationship Type
Manfred Bruhn and Dominik Georgi
Service quality management is acknowledged as a major key to service provider's

economic success. In contrast to traditional marketing activities, it aims at generating
economic success by satisfying and retending current customers. However, as other
activities of service corporations, quality activities should be assessed against their
profitability. This paper presents an approach for quantifying the benefit of quality
activities generated by increasing customer satisfaction and customer retention.
Furthermore, because of several moderating factors the effects of quality manage-
ment on customer behavior are not linear in nature. An important moderating factor
is the type of the current customer relationship. Therefore, this paper also analyzes
this construct's moderating effect on the relationship between customer satisfaction
and customer repurchase behavior.
Biographies 307
Biographies
Manfred Bruhn
Manfred Bruhn conducted his studies in Business Administration at the University

of Münster, Germany and graduated in 1974. After his Doctorate (Dr. rer. pol.) in
1977 (Supervisor: Prof. Dr. Dr. h.c. Heribert Meffert, Director of the Institute for
Marketing, University of Münster/Germany), he received his University Lecturing
Qualification in 1985. From 1983 to 1995 he was head of the Chair of Marketing
and Retailing at the European Business School in Oestrich-Winkel, Germany.
Since 1995, he is now head of the Chair of Marketing and Business Administration
at the Center of Economics at the University of Basel, Switzerland.
He provides teaching experience since 1979, teaching at the following institutions:

University of Münster, Germany; European Business School in Oestrich-Winkel,
Germany; University of Jena, Germany; Business University in Vienna, Austria;
University of Basel, Switzerland.
Manfred Bruhn's current research areas are Marketing Management, Service Mar-
keting, Quality Management for Services, Integrated Communications, Internal
Marketing.
Fran\(oise Chevalier
Since 1986 Franc;:oise Chevalier works with the Group HEC, Jouy-en-Josas Cedex,
France. From 1990 to 1994 she acted as a course coordinator for human resource
management. Today she is an associate professor at the Chair of Management and
Human Resources with the same Groupe. There, she is also a member of the re-
search committee. In addition, in 1995 she was a research associate at UC
Berkley, USA. From 1993 to 1996 she lead the "Organizational Behaviour" group of
the CEMS (Community of European Management Schools).
Franc;:oise Chevalier's current research areas are human resource management,

new methods of social organization and organizational change and sociology in
organizations.
308 Biographies
Alexander Decker
In 1994 Alexander Decker graduated from the University of Bayreuth with a Di-
ploma in Business Administration. Then, he was working as a free research assis-
tant at the chair of Business Administration and Automobile Marketing at the Otto-
Friedrich University of Bamberg for one year. As the project leading research as-
sistant he was in charge for the development of a research tool to measure dealer
satisfaction in the German car industry and for the first empirical survey in 1995.
Since 1996 he is a research assistant with Prof. Dr. Bernd Stauss at the Chair of
Business Administration and Marketing respectively Services Management at the
Catholic University of Eichstätt, Germany.
He has been doing research in several areas, but especially with respect to satis-
faction and/or the car industry. Consequently, his doctoral thesis deals with the
dealer satisfaction as a goal in vertical marketing of the German car industry.
80 Edvardsson
Ba Edvardsson is professor and director at the Service Research Center (CTF),

University of Karlstad, Sweden. He has published 12 books focusing on different
aspects of service management and published 30 articles and a number of re-
search reports. He is often invited to speak to executives in service companies in
Sweden and other countries.
Dominik Georgi
Dominik Georgi has conducted his studies in Business Administration at the Euro-
pean Business School, Oestrich-Winkel, Germany between 1990 and 1995. During
this time he spent about one year (1992) abroad studying at Ecole Superieure de
Commerce in Dijon, France, as weil as at "Thunderbird" American Graduate School
of International Management in Glendale, Arizona, U.SA
Since January 1996, he is a research assistant with Prof. Dr. Manfred Bruhn at the
Chair of Marketing and Business Administration at the Center of Economics
(WWZ) , University of Basel/Switzerland.
His research areas include the following topics: service marketing, measurement of
service quality, quality management for services, profitability of quality manage-
ment.
Biographies 309
Frank M. Go
Frank Go is professor and chairman of the Centre for Tourism Management, Rot-
terdam School of Management, Erasmus University, since September 1996. Before
that he was working since 1970 in the United States and Canada, where he was
associated with the business management schools of Universities in Toronto and
Calgary. From 1992 to the summer 1996 he was professor and chairman of the
school for tourism management at the Hong Kong Polytechnic University.
His research interests are directed towards the relationship between marketing,
entrepreneurship and innovation and tradition in the tourism industry. Regularly he
acts as chairman or speaker at conferences and fulfils executive positions in busi-
ness, higher education institutes and governing bodies.
Robert Govers
In November 1996 Robert Govers was appointed as consultantlproject manager

with the Centre for Tourism Management (CTM) of the Rotterdam School cf man-
agement, Erasmus University. Within the CTM the research focus is on net-
work/value chain development. After his graduation in 1995 he was active as a
guest lecturer for nine months, at the Department of Travel and Tourism at the
Technikon Witwatersrand, Johannesburg, South Africa. Within the CTM he has
been engaged in writing strategie plans for travel and tourism related businesses
and organizations and in organizing and participating in workshops and projects on
the use of ICT within the travel and tourism industry.
Apart from that his interests are in the areas of market research, the relationship
between ICT and marketing in the travel industry and regional tourism development
based on a partnership with the environment and the cultural and historical heri-
tage.
310 Biographies
Ursula Hansen
Ursula Hansen studied Business Administration and Social Psychology at the uni-
versities of Vienna, Austria and Kiel, Germany. After her graduation she worked as
assistant with Prof. Dr. Eugen Leitner, she finished her doctoral thesis in 1968 and
qualified for lecturing in university in 1974. During this time she completed a re-
search project in the United States. In 1975 she became head of the chair Market-
ing1: for Market and Consumption at the University of Hanover, Germany. In addi-
tion to this she worked as a visiting professor at the university of Leipzig, Germany
(1989/1990) to get involved in the (Iocal) development of scientific work in the field
of Business Administration.
In 1991 she contributed significantly to the founding of the university-linked Institute

for Market-Environment-Society (Institut für Markt-Umwelt-Gesellschaft (imug) e.V.)
in which she is still involved as a member of the scientific board.
Her interests focus on consumer behavior, retail marketing, after-sales marketing,

ecological-orientated marketing, product design, consumer policy, ethics of mar-
keting and history of marketing.
Thorsten Hennig-Thurau
Thorsten Hennig-Thurau studied Business Administration at the University of

Lüneburg, Germany and graduated from there (Diplom-Kaufmann) in 1994. From
July 1994 to January 1996 he work as assistant at the Institute for Market-
Environment-Society where his interests were focused on relationship and after-
sales marketing as weil as empirical marketing research.
From November 1994 to March 1998 he was Research Assistant with Prof. Dr.
Ursula Hansen at the Department of Marketing at the University of Hanover, Ger-
many. Since he finished his doctoral thesis in February 1998 he is now Scientific
Assistant and preparing for his post-doctoral thesis at Prof. Hansen's chair.
Claudia Klausegger
Claudia Klausegger studied Business Administration at the University of Economics

and Business Administration of Vienna, Austria. Since 1991 she is a University As-
sistant and lecturer for marketing with Prof. Dr. Fritz Scheu eh at the chair of Mar-
keting at the same University. She was a presenter in a course on environmental
management that was organized by the University of Economics and Business Ad-
ministration of Vienna, Austria. Besides, until 1997 she was as weil editor of the
magazine "der markt".
Her research interests are on complaint management, customer retention, NPO

marketing and environmental management.
Biographies 311
Paul Kunst
Paul Kunst studied business administration at the Erasmus University in Rotter-

dam, the Free University in Amsterdam and the University of Amsterdam, where he
graduated in 1983. From 1983 to 1989 he was an Assistant Professor at the Fac-
ulty of Economics and Business Administration at Maastricht University.
During this time, he began preparatory work for a dissertation based on contral
systems in organizations. He has also been involved in studies concemed with
works councils, as weil as studies on the relationships between strategy, structure,
culture and contra I. Paul Kunst is now a consultant and part-time Senior Research
Fellow at MERIT, where his main research interest concentrates on quality man-
agement, management accounting and human resource management.
Paul Kunst started this year as a Visiting Professor at the Limburgs Universitair
Centrum (LUC) in Diepenbeek, Belgium.
Jas G.A.M. Lemmink
Jos Lemmink has been visiting professor at University of Southem Queensland,

Australia and Limburg University Center, Belgium. He holds a degree in Business
Administration fram the University of Groningen and a Ph.D. from the University of
Limburg. He was a market researcher for Dutch Telecom for four years and was a
consultant for a number of large companies.
Currently, he is professor with the Department of Marketing & Marketing Research at

the Faculty of Economics and Business Administration, Maastricht University. He initi-
ated (with Paul Kunst) the publication of this series of bocks on Quality Management in
Services on the basis cf aseries of workshops. Furthermore he published 7 books and
more than 100 articles on marketing research methods and techniques, quality man-
agement in services, and service marketing and management in general. His work has
been published amongst other outlets in Intemational Joumal cf Research in Market-
ing, Joumal of Economic Psychology and Advances in Services Marketing and Man-
agement. His research interests concem service marketing and management,
analysis of service processes and marketing and new media.
Harald Meissner
Harald Meissner studied Economics at the University of Bochum, Germany and

graduated with a diploma in economics (Diplom-Ökonom) in 1995. Since 1996 he
has been working as research assistant with Prof. Dr. Bemd Stauss at the Chair of
Business Administration and Marketing respeclively Services Management at the
Catholic University of Eichstätt, Germany.
Harald Meissner is currently writing his doctoral thesis on the topic of Managing
Customer Dialogues via email.
312 Biographies
Ko de Ruyter
Ko de Ruyter is currently associate professor of Marketing and Marketing Research

at Maastricht University and director of the Maastricht Academic Center for re-
search in Services (MAXX). He has been a visiting professor at Maryland State
University and Purdue University. He holds degrees in Linguistics and Political Sci-
ences from the Free University Amsterdam and the University of Amsterdam. He
received his Ph.D. in Management Sciences from the University of Twente. He has
published five books and over 130 refereed articles. His research interests concern
service quality, relationship marketing and customer satisfaction and dissatisfac-
tion.
Dieter Scharitzer
Dieter Scharitzer studied Business Administration at the University of Economics

and Business Administration of Vienna, Austria. Since 1989/1990 he is university
assistant with Prof. Dr. Fritz Scheuch at the chair of Marketing at the same univer-
sity. He is also lecturer for Surgery Management at the Veterinary University of Vi-
enna, Austria. Dieter Scharitzer's scientific interest is focused on the evaluation of
service quality, marketing for professional services and controlling in retailing.
Jan A. Schulp
Jan Schulp studied biology at the Free University of Amsterdam, the Netherlands.
He graduated with a ph.d. degree in molecular genetics in 1972. From 1972 -
19899 he was a lecturer of Educational Methods for Science in Primary Education
at the Oe Him Teachers Training College Sneek, the Netherlands. From 1987 -
1992 he was a lecturer for Hygiene and Nutrition at the Hotel Management School
of Leuwarden, the Netherlands. In this time he oriented towards the research field
of services marketing and service quality. From 1992 - 1995 he was the head of
the department of Service Management at the Faculty of Economics at the
Christelijke Hogeschool, Noord-Nederland (CHN) Leuwarden, the Netherlands.
Since 1996 Jan Schulp is a Senior Consultant at the Institute for Service Manage-
ment. He is involved in the curriculum development for the CHN and for extern al
relations, both in the Netherlands and abroad. His research projects focuses espe-
cially on customer satisfaction.
Biographies 313
Bernd Stauss
Bernd Stauss completed his studies in Business Administration at the University of

Hamburg, Germany. Later, he was a Research Assistant with Prof. Dr. Ursula Han-
sen at the University of Hanover, where he completed his doctoral thesis and his
habilitation. He went on to the University of Innsbruck, where he was a visiting
professor. In 1989, Bernd Stauss became Head of the Chair of Business Admini-
stration and Marketing at the faculty for Economics and Business Administration in
Ingolstadt, part of the Catholic University of Eichstätt, Germany. Since 1997 he his
the Head of the Chair of Business Administration and Services Management at the
same school.
His main interests are in the areas of services marketing, service quality, Total
Quality Management, internal marketing within service companies, measurement of
customer satisfaction and complaint management.
He is a member of the American Marketing Association and the International Serv-

ice Quality Association and a founding member of the International Academy of
Services Research and Education. He also is a member of the editorial board of the
European Journal of Marketing and the International Journal of Service Industry
Management.
Tore Strandvik
Tore Strandvik IS director of Center for Relationship Marketing and Service Man-
agement (CERS) at HANKEN - the Swedish School of Economics and Business
Administration. This research and knowledge center has about 20 full-time re-
searchers focusing on academic research concerning "The influence of customer
relationships and business networks on the competitiveness of firms". Tore
Strandvik is also associated to the Swedish School of Economics and Business
Administration, Helsinki, Finland since 1981 as a lecturer and assistant professor
and 1987-98 associate professor at the Department of Marketing of the school. He
is currently involved in research projects concerning customer perceived service
quality, the dynamics of customer relationships and relationship termination. Arti-
eies have been published in several international journals and books. He has ac-
tively participated in international conferences and workshops concerning service
quality, relationship marketing and managerial cognition. He is also a frequent
speaker at management seminars and workshops in Scandinavia on service qual-
ity, customer satisfaction and service marketing and management as weil as rela-
tionship marketing and management.
314 Biographies
Martin G.M. Wetzeis
Martin Wetzeis is an Assistant Professor with the Department of Marketing & Marketing
Research at the Faculty of Economics and Business Administration, Maastricht Univer-
sity. In 1993 he obtained his Master's Degree (with distinction) at the Faculty cf Eco-
nomics and Business Administration of the University of Limburg. After finishing his
dissertation titled Service Quality in Customer-Employee Relationships in July 1998 he
became an Assistant Professor with Department of Marketing & Marketing Research.
His main research interests are: service quality, customer satisfaction/dissatis-

faction, customer value, quality management in service organizations, services
marketing, marketing research, cross-functional cooperation and relationship mar-
keting.

Paul Kunst, Jos Lemmink, Bernd Stauss Auth., Paul Kunst, Jos Lemmink, Bernd Stauss Eds. Service Quality and Management

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Paul Kunst, Jos Lemmink, Bernd Stauss Auth., Paul Kunst, Jos Lemmink, Bernd Stauss Eds. Service Quality and Management

Hochgeladen von

Copyright:

Verfügbare Formate

Kunst/Lemmink/Stauss (Eds.

Der Wandel von der Industrie- zur Dienstleistungsgesellschaft ist de

Springer Fachmedien Wiesbaden GmbH

Service quality and management / Paul Kunst ... (eds.). -

Alle Rechte vorbehalten

© Springer Fachmedien Wiesbaden 1999

Lektorat: Ute Wrasmann

Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in diesem

ISBN 978-3-8244-6780-8 ISBN 978-3-322-90981-7 (eBook)

Dynamics in Service Quality Perceptions - A Longitudinal

National Customer Satisfaction Indices:

The Link between Customer Satisfaction and Dealer

Criticality of Critical Ineidents in Customer Relationships

Critical Ineidents in Dutch Consumer Press: Why Dissatisfied

Achieving Service Quality through the Application of

The Sequential Ineident Technique tor Innovations (SITI) -

Beyond the Call of Duty: The Impact of Perceived Support

Quality Management and Organizational Change

Sales-Related Benefit-Analysis of Service Quality Investments -

List of Abstracts 301-305

Claudia Klausegger and Dieter Scharitzer

3 Research Design and Hypotheses 9

4 Preliminary Results - Hypothesis Testing 15

5 Summary and Research Outlook 21

In our research quality perceptions of students of an Auslrian "Fachhochschule"

• To add a new evaluation instrument to the methods already used in the

3 Research Design and Hypotheses

• Differentiating carefully the phases in the service production and consump-

• Approprialely selecling dimensions and items to adequately describe the

(Fornell 1992, Anderson/Forneli/Lehmann 1994) and "Das Deutsche Kun-

For practically applying the results obtained by satisfaction measurement in market

3.2 Empirical Design of the Survey

J L ____•• 1" stage of data eolleetJon (SeptJOet. 96):

J L ____•• 2"" stage of da/a eolleetJon (July/Aug. 97):

Specialized Course Program

Fig. 1: Longitudinal study Analysis

4 Preliminary Results - Hypothesis Testing

Personafity Training Courses

o Materials Management & Logistics j

Fig. 2: Quality Portfolio of the Undergraduate Course Subjects

4.2 Longitudinal Analysis of Subjects

Above all, courses in marketing, computer science, personnel management!

Personality Training Courses"I~~l'::.-=::ok~~:= ~

Fig. 3: Importance ot Subjects trom the student's point ot view

Fig . 4: Significant Differences in Ihe Importance ot subjects

4.3 Detailed Analysis of the Dynamics of

Fig. 5: Evaluation of Economics after Survey 1 and Survey 2

1st Survey 2nd Survey

Fig. 6: Subject Priority of Economics

5 Summary and Research Outlook

In our study, ex-post analysis of outcome-related quality dimensions of educational

Since we analyzed several measurement techniques for course evaluation in the

namics in customers satisfaetion.

Anderson, E.W.lFomell, C.lLehmann, D.R. (1994): Customer Satisfaction, Market

Boström, G.-O./Wilson, T. (1993): Changing Quality Perceptions Among Providers

Danaher, P.J.lMattsson, J. (1994): Cumulative Encounter Satisfaction in the Hotel

Fisk, R.P. (1981): Toward A Consumption/Evaluation Process Model for Services,

FarnelI, C. (1992): ANational Customer Satisfaction Barometer: The Swedish Expe-

Grönroos, C. (1993): Toward a third phase in service quality research: Challenges

Haller, S. (1995): Beurteilung von Dienstleistungsqualität, Wiesbaden.

Liljander, V.lStrandvik, T. (1994): The Nature of Relationship Quality, in: EIASM

Meyer, A./Mattmüller, R. (1987): Qualität von Dienstleistungen - Entwurf eines pra-

Parasuraman, A.lZeithaml, V.A./Berry L.L. (1988): SERVQUAL: A Multiple-Item

Scharitzer, D. (1994a): Dienstleistungsqualität - Kundenzufriedenheit, in: Topritz-

Personality Training Courses"Il'::.-=::ok:= ~