Sie sind auf Seite 1von 14

The Journal of Product Innovation Management 18 (2001) 258 –271

Patterns of cooperation during new product development among


marketing, operations and R&D: Implications for project performance
Eric M. Olsona,*, Orville C. Walker Jr.b, Robert W. Ruekertc, Joseph M. Bonnerd
a
Professor and Chairman of the Marketing, Strategy, and International Business Department, College of Business, University of Colorado at Colorado
Springs, 1420 Austin Bluffs Parkway, Colorado Springs, CO 80933
b
James D. Watkins Professor of Marketing, Carlson School of Management at the University of Minnesota
c
Professor of Marketing and Associate Dean in the Carlson School of Management at the University of Minnesota
d
Assistant Professor of Marketing and Supply Chain Management in the Eli Broad College of Business at Michigan State University

Abstract
Successful new product development is fundamentally a multidisciplinary process. While this view has helped lead management to the
wide-spread adoption of cross-functional new product development teams, in this study we question whether simply increasing the level of
functional integration is truly a guarantee for enhancing the performance of new products. To assess this we examined patterns of
cooperation between marketing, R&D, and operations at both early and late stages of the new product development process for 34 recently
developed products whose level of innovativeness ranged from high to low. A unique feature of this study is that data were collected from
four sources for each project. This included personal interviews with a project leader and written surveys from marketing, operations, and
R&D personnel on each project.
Findings from this study reveal that: (1) functional cooperation typically increases as the process moves from early to late stages; (2)
cooperation between marketing and R&D is highest during early stages of the process, but for marketing and operations, and for R&D and
operations, cooperation typically increases as the process moves from early to late stages; (3) higher project performance – irrespective of
the level of project innovation – is demonstrated when cooperation between marketing and R&D, and cooperation between operations and
R&D is high during early stages; (4) late stage cooperation between marketing and operations, and R&D and operations is a key determinant
in project performance for innovative products but not for noninnovative products, and; (5) that early stage cooperation between marketing
and operations is associated with superior performance for low innovation projects but is also associated with poor performance for
innovative projects.
Findings from this study demonstrate that the importance of cooperation between specific functional dyads (i.e., marketing - R&D; R&D
- operations; operations - marketing) indeed varies by time (i.e., early vs. late stages), and by the level of innovativeness (i.e.,
new-to-the-world vs. modifications) associated with the new product being developed. © 2001 Elsevier Science Inc. All rights reserved.

1. Introduction action and integration during the new product development


(NPD) process [8,14,41]. For instance, one extensive survey
The creation of successful new products is fundamen- found that cross-functional teams are employed in about
tally a multidisciplinary process. And it seems reasonable to two thirds of all NPD projects, and that their use increases
assume that factors like rapid technological change, flexible to 85% in projects aimed at developing highly innovative
production processes, and global competition are making products or services [12].
close collaboration across functions even more crucial for
the introduction of profitable and timely new products. 1.1. Some unanswered questions
Consequently, firms are embracing coordination mecha-
nisms (e.g., Quality Functional Deployment procedures) A large number of past studies appear to demonstrate the
and organizational structures (e.g., cross-functional teams) positive effects of cross-functional integration on NPD suc-
aimed explicitly at increasing the level of functional inter- cess across a variety of cultural contexts. [9,15,18,22,32,34,
35] However, broad and generalizable conclusions may not
be warranted. A recent meta-analysis of the product devel-
* Corresponding author. Tel.: ⫹1-719-262-3345. opment literature found that the mean correlation between
E-mail address: eolson@mail.uccs.edu (E.M. Olson). functional integration and NPD success – based on a total of

0737-6782/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S 0 7 3 7 - 6 7 8 2 ( 0 1 ) 0 0 0 9 1 - 1
E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271 259

41 reported correlations – is not statistically significant after particular function might play a more crucial role – and
adjusting the individual correlations for sampling error and whether interactions with that function might therefore be
variance in measurement reliability. [17 p.22] As Henard more important – at some stages in the development process
and Szymanski point out, simply increasing the level of than at others. R&D, for instance, may have more to con-
functional integration may not be sufficient to improve new tribute during a project’s early concept development and
product success; the quality, focus and timing of that inte- product design phases than during the later commercializa-
gration may be critical. [17 p.28] This suggests that a tion stage. Such variations in the importance of interactions
number of important questions remain to be explored. with a given function across stages of NPD projects might
For one thing, much of the existing research on the have important implications for choosing permanent versus
effects of functional cooperation have not included all the ad hoc members of the development team and for designing
functions likely to have a major impact on NPD success. appropriate mechanisms for stimulating cooperation and
With a few exceptions [9,13,19,25,35,36], past studies have resolving functional conflicts at different points in the de-
focused solely on the interface between marketing and R&D velopment process.
[6,15,32,33,37]. But while both R&D and marketing often
play critical roles in product development, many important
issues relevant to the technical and market success of a new 2. Purpose
product clearly fall within the domain of operations, such as
the sourcing of components and materials, the production of The research reported here seeks insights relevant to the
prototypes, and the design of efficient production processes above questions by analyzing patterns of cooperation
and quality control procedures. Therefore, as Song, et.al. among marketing, operations, and R&D participants on 34
argue, “. . . an accurate representation of cross-functional different product development projects across a variety of
relationships in NPD must include all three functional per- industries. Information about the performance outcomes of
spectives: R&D, marketing and manufacturing.” [35 p.37] those projects is also examined to determine how patterns of
Examining cooperation levels across all three pairs of func- cross-functional cooperation relate to the success of the
tions would also enable a search for possible differential products developed. Specifically, this study addresses the
effects. Perhaps cooperation between particular pairs of following issues:
functions at particular times in the development process is Y Do levels of cooperation among representatives of the
especially crucial to project performance and timeliness. marketing, operations, and R&D functions change
Another shortcoming of most of the existing studies from the early to the later stages of new product
concerns their level of analysis. Few studies examine dif- development projects?
ferences in interaction or cooperation between individual Y Are differences in the pattern of cooperation among
team members representing different functional areas, or functional departments at different stages in the de-
how such differences affect the outcomes of individual NPD velopment process related to differences in perfor-
projects. Instead, they relate aggregate levels of interaction mance outcomes across projects?
or integration between entire functional departments to Y Are relationships between cooperation patterns and
summary measures of the firm or business unit’s overall performance outcomes moderated by the innovative-
performance in new product development. This aggregated ness of the project?
level of analysis is problematic for two reasons. First, the
amount of uncertainty and the technical challenges facing a In the following section we first define what we mean by
particular functional department may vary across different cross-functional cooperation and then draw on resource
kinds of development projects [11,16,25,29,38]. In particu- dependency theory and previous work in the marketing,
lar, the complexity and the innovativeness of the project product development, and operations literatures to develop
(e.g., line extensions vs. new-to-the-world products) may testable research propositions. We then describe our re-
influence the magnitude of the technical, operational, and/or search methods and measures, analyze our results, and dis-
marketing challenges that confront the firm. The relative cuss our findings. In the last section of the article we point
importance and tractability of those challenges, in turn, out some of the limitations of this study, as well as its
likely moderate the impact of each department’s inputs on implications for future research and for managers charged
the project’s ultimate outcomes. with overseeing or participating in product development
A second reason why a project-level analysis might en- activities within their companies.
hance our understanding of the effects of functional coop-
eration on NPD performance is because development 3. Conceptual framework and propositions
projects typically proceed through a variety of stages: op-
portunity identification, concept development, product de- 3.1. Defining cross-functional cooperation
sign, and so forth – as detailed in Urban & Hauser [39], and
others. Each stage presents a different set of uncertainties While many researchers have attempted to understand
and challenges [2,21]. This raises the question of whether a the impact of cross-functional integration on NPD perfor-
260 E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271

mance, they have seldom agreed on what functional inte- eration rather than the attitudinal construct of collaboration
gration means or how it might best be measured. They have for two reasons. First, behaviors are easier for management
employed a variety of operational definitions, ranging from to quantify, influence, and change than attitudes. For in-
simple measures of the frequency of interaction between stance, the frequency of interaction can be shaped by the
two departments to multidimensional measures of collabo- organizational structure (e.g., the creation and composition
rative behaviors and attitudes. (See Kahn 18, pp.138 –140 of NPD teams), formal policies, and investments in com-
for a more extensive review). munications infrastructure. Similarly, the amount of infor-
At one extreme, some early studies focused largely on mation and resource sharing between functions can be in-
the frequency or amount of communication between the fluenced by the adoption of clearly defined development
members of different departments [for example, 13,29]; the processes, procedures, and reward systems. [22] On the
rationale being that frequent interaction is necessary for other hand, interdepartmental attitudes and emotions are
effective functional integration [40]. The fact that two func- more intangible and difficult for management to regulate or
tions communicate frequently, however, does not guarantee change. [18 p.139] Thus, our focus on the behavioral di-
they will exchange any useful information, and subsequent mensions of cooperation across functions should help make
research suggests that the frequency of functional interac- participants’ responses somewhat more objective, and the
tion, by itself, does not bear a significant relationship to managerial implications of the findings more clear-cut.
NPD outcomes [18]. Thus, frequent interaction may be a A second advantage of our focus on behavioral cooper-
necessary but not a sufficient condition for functional inte- ation is that it ensures a relatively conservative test of our
gration. conceptual propositions. Any relationships observed be-
Therefore, a number of researchers have tried to gauge tween behavioral cooperation and project performance
the substantive content – as well as the amount – of inter- would likely be further strengthened by attitudinal affinities
action between functions in order to assess the degree of across departments, such as a shared vision, common val-
constructive cooperation or teamwork inherent in their re- ues, and collective goals [18].
lationship [7,15,32,33]. For example, Song, et.al. define
functional cooperation as the amount of “. . . interdepen- 3.2. The relationship between functional dependence and
dency and information sharing between the various organi- cooperation
zational units.” [35 p.37] For the most part, measures of this
“cooperation” construct are behavioral; focusing on the na- The classic studies of Burns and Stalker [5], Lawrence
ture and frequency of cooperative or conflicting actions and Lorsch [20], and others suggest that the greater the
undertaken between functional departments. environmental uncertainty an organization faces, the greater
At the other end of the continuum, some recent studies the functional specialization it is likely to incorporate within
employ more complex, multidimensional measures which its internal structure. Such specialization enables the firm to
attempt to assess both the behavioral and attitudinal nature segment uncertainty by function and to charge groups of
of relationships across functions [18,19]. Kahn labels this specialists with resolving – or buffering the rest of the
broader construct “collaboration,” and defines it as “. . . an organization from – uncertainties related to a specific func-
affective, volitional, mutual/shared process where two or tional domain. Thus, R&D specialists can focus on resolv-
more departments work together, have mutual understand- ing problems related to newly emerging technologies, while
ing, have a common vision, share resources, and achieve marketers deal with uncertainties concerning customer pref-
collective goals.” [18 p.139] In other words, it represents the erences and market demand.
affective, unstructured nature of functional relationships as Resource dependency theory goes on to suggest that the
well as the substantive actions engaged in by the partici- locus or content of the uncertainty an organization faces
pants. also influences the relative importance of each functional
There are some advantages to examining both the behav- specialist’s role in resolving that uncertainty. This, in turn,
ioral and attitudinal dimensions of functional integration. helps shape the pattern of interfunctional dependence – as
The two dimensions may have different effects on NPD well as interaction and power – within the firm. [3,26]. For
project performance and carry different management impli- instance, R&D specialists are likely to play a more active
cations. In this study, however, we focus solely on the and influential role in shaping a firm’s strategic decisions
behavioral dimension of functional integration, which – when it operates in a turbulent and evolving technical en-
consistent with the terminology of Song, et.al [35]. and vironment. On the other hand, operations specialists are
others – we label “cooperation.” Our operational definition likely to be more influential in firms pursuing value-driven
of cooperation includes both the frequency of interaction competitive strategies that place a premium on finding ways
and the amount of information and resources shared be- to improve operating efficiency and customer value.
tween a pair of functions involved in an NPD project: As Ruekert, et.al. [29], Gupta, et.al [16], and others point
marketing - R&D, marketing - operations, and R&D - op- out, however, functional specialization and departmental-
erations. ization also creates coordination difficulties. Thus, they ar-
We chose to focus on the behavioral construct of coop- gue that as a firm’s functional specialization increases in
E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271 261

response to greater environmental uncertainty, the need for sents a major “go-no go” decision point in most projects and
integration and cooperation across functional departments brings with it major changes in the kinds of issues and
also increases. When applied to a specific NPD project, the decisions faced by the development team [39].
above resource dependency arguments suggest that team
members from other functions will perceive themselves to 3.3a. Overall levels of cooperation across stages. While
be more dependent upon the R&D department – and there- there is no published evidence to draw upon, it seems likely
fore have greater motivation to interact and foster cooper- that the total frequency of interaction and level of cooper-
ation with R&D – when the project involves relatively high ation across the various functions involved in a NPD project
levels of uncertainty concerning technological issues sur- increase as the project moves from early to later stages in
rounding the product’s design and performance. Note, too, the development process. As the conceptual parameters of
that the level of such technical uncertainty may vary both the new product are worked out during the early stages, the
across projects (e.g., a highly innovative product with new- decisions made impose constraints which the various func-
to-the-world technology vs. a modification of an existing
tions must embrace and adapt to in the later stages as they
product or core technology) and across stages within a given
attempt to convert those parameters into a physical product
project.
that can be efficiently manufactured and effectively mar-
Similarly, team members should interact and cooperate
keted. Thus, their mutual interdependence increases as does
more intensively with marketing and sales representatives
their need for information sharing and coordination of their
when uncertainty concerning customer preferences, levels
of demand, or competitor reactions in the marketplace is activities. Also, as the project moves from concept to phys-
high. Finally, dependence upon and cooperation with oper- ical product there are more opportunities for evaluation and
ations will probably be greatest when there are major con- feedback concerning the product’s technical performance,
cerns about process technology or the firm’s ability to meet manufacturability, and customer appeal. As additional in-
production schedules, cost parameters, or desired quality formation is gained from the testing of prototypes, pilot
levels. production runs, beta sites, and other evaluation activities in
a project’s later stages, the need for the various functions to
3.3. Patterns of functional dependence and cooperation share and adapt to that information also increases. Conse-
across project stages quently:
H1: The overall level of cooperation among representa-
Development projects progress through a series of tives of the R&D, marketing and operations functions is
stages, such as opportunity identification, concept develop- significantly higher in the later stages of a NPD project
ment, product design, process design, and commercializa- than in its early stages.
tion [11,28,39]. Each stage presents a different set of func-
tional uncertainties and challenges [2]. Therefore, a given 3.3b. Levels of cooperation between pairs of functions. As
functional department is likely to play a more crucial role in mentioned, the types of uncertainty faced by members of a
reducing or resolving uncertainty surrounding its particular NPD project team – and therefore their dependence upon
area of expertise during some stages of a project than others. and likely cooperation with a given functional department –
Consequently, levels of dependence on a particular function can vary across stages of the project. Therefore, while the
– and the level of interaction and cooperation with that overall level of functional cooperation may increase from
function – may also shift across the stages of an NPD early to later stages, the relative level of cooperation be-
project.
tween specific pairs of functions may follow different pat-
Patterns of dependence and cooperation among functions
terns.
probably vary at every stage in a development project. To
Specifically, representatives of the marketing department
help make data collection more tractable and to more clearly
should have knowledge or access to information about cus-
focus our discussion of research propositions and results,
however, we have condensed the five or six stages found in tomer preferences and competitive offerings that are partic-
most existing models of the NPD process down to two: ularly relevant for resolving design and positioning issues in
early stages versus later stages. Operationally, early stages the early stages of a project. Similarly, because R&D or
encompass activities concerned with product conceptualiza- product engineering bears primary responsibility for trans-
tion and evaluation, such as idea generation, concept devel- lating available technology into a viable design with work-
opment, and business assessment. Later stages involve ac- able features and desirable performance attributes, it medi-
tivities related to the production and commercialization of ates knowledge and information resources that are
the new product, such as prototype design and development, important for resolving design issues during a project’s
process design, product testing, and market introduction. early stages. On the other hand, such competencies and
This two-stage perspective makes sense theoretically since resources may be less relevant for solving the kinds of
the shift from conceptualization activities to the physical operational problems that arise during pilot production and
production and commercialization of a new product repre- commercialization. Consequently:
262 E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271

H2: The level of cooperation between marketing and products developed by the firm (2), management’s satisfac-
R&D, relative to the level between other pairs of func- tion with its technical design and functional performance,
tions, is significantly greater during the early stages of and (3) the degree to which it achieves its sales objectives.
NPD projects than during the later stages.
Efficiency measures are concerned with the amount of re-
While marketing has much expertise relevant to resolv- sources required to complete the project. Money and time
ing design and positioning issues in the early stages of a are certainly among the most constrained – and therefore
project, it is also likely to have primary responsibility for important – resources necessary for developing new prod-
developing an effective marketing program during the later ucts. Therefore, we also examined the relationships between
stages to communicate product benefits to the target market, patterns of functional cooperation and (4) the project team’s
stimulate demand, and achieve adequate distribution. Sim- adherence to its budget, and (5) the project’s cycle time.
ilarly, operations typically has primary responsibility for – While these measures were initially addressed as sepa-
and the expertise to solve – process engineering and pro- rate items in this study, subsequent tests of dimensionality
duction problems. Thus, it usually plays a crucial role in and reliability revealed that they were indeed highly inter-
producing prototypes, developing production methods, and
related. Consequently, we follow the lead of other research
achieving desired quality and output levels for commercial-
in marketing strategy and new product development and
ization. While operations may be asked for advice on the
consider performance at the aggregate level [14,17,26,31,
producibility of alternative designs, relative costs, and so
33,34].
forth during earlier conceptualization stages, its competen-
cies may be seen as less relevant for resolving early design Recall that we drew upon the tenets of resource depen-
and positioning issues. Consequently: dency theory to formulate hypotheses about expected pat-
terns of cooperation between pairs of functions over the
H3: The level of cooperation between marketing and
course of a NPD project. But what if the expected patterns
operations, relative to the level between other pairs of
functions, is significantly greater during the later stages do not hold within a given project team? Suppose team
of NPD projects than during the earlier stages. members misperceive the relevance of a particular func-
tional area for helping them to resolve the uncertainties they
Finally, the above arguments suggest that R&D and face at a particular stage, or that structural or procedural
operations representatives are likely to interact and share
constraints make it difficult for them to attain the desired
information and advice relatively more frequently during
level of cooperation with a particular function. Will the
the later stages of a project when conceptual designs from
effectiveness of such a team be compromised? Resource
the lab must be transformed into workable and producible
dependency theory suggests that the answer is most likely
products. Thus:
yes. We expect, therefore, that project teams whose patterns
H4: The level of cooperation between R&D and opera- of functional cooperation best conform to the amount and
tions, relative to the level between other pairs of func-
locus of uncertainty they are likely to experience at different
tions, is significantly greater during the later stages of
NPD projects than during the earlier stages. project stages, will also generate the most effective and
efficient outcomes. Specifically:
3.4. Functional cooperation and performance outcomes
H5a: The level of cooperation between marketing and
R&D during the early stages of a NPD project is posi-
While observing the patterns of cooperation between
tively related to project performance.
various functions at different project stages may provide
some insights into how NPD teams work, the really inter- H5b: The level of cooperation between marketing and
esting question is whether and how particular patterns of R&D during the later stages of a NPD project is not
functional cooperation are related to a project’s subsequent significantly related to project performance.
performance outcomes. Before we can address this ques- H6a: The level of cooperation between marketing and
tion, however, we must first specify the performance dimen- operations during the early stages of a NPD project is not
sions on which a NPD project might be evaluated. significantly related to project performance.
H6b: The level of cooperation between marketing and
3.4a. Performance dimensions. Existing research has exam-
operations during the later stages of a NPD project is
ined the impact of functional interaction and cooperation on
positively related to project performance.
a number of dimensions of a firm’s NPD performance [25].
These measures can be considered in terms of effectiveness H7a: The level of cooperation between R&D and oper-
and efficiency. Measures of project effectiveness are con- ations during the early stages of a NPD project is not
cerned with the quality of the resulting product and its significantly related to project performance.
ultimate success in the marketplace. Therefore, we exam- H7b: The level of cooperation between R&D and oper-
ined the relationships between patterns of functional coop- ations during the later stages of a NPD project is posi-
eration and (1) the new product’s quality relative to other tively related to project performance.
E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271 263

3.5. The moderating effects of project innovativeness H8c: The positive impact on project performance from
high levels of cooperation between R&D-operations dur-
By definition, projects involving the development of in- ing later stages of the product development process (as
novative products or services are likely to involve substan- hypothesized in H7b) will be greater for innovative
tial amounts of uncertainty concerning a variety of func- projects than for less innovative projects.
tional issues [11]. And the firm’s personnel usually have
little relevant experience or prior learning to draw on when In hypotheses 8ac we predict that higher levels of coop-
resolving those functional uncertainties. This combination eration between marketing and R&D at early stages of the
of high uncertainty and lack of relevant experience tends to product development process, and between marketing and
increase the amount of interdependence – and therefore the operations, and R&D and operations at the later stages of
need to share information, expertise, and other resources – the product development process will result in higher prod-
across functions [16,25,29]. uct performance levels for highly innovative new products
Greater functional cooperation should improve the qual- than for new products of low innovation. However, this then
ity of the new product and improve its chances for success begs the question, what about the other side of the equation?
in the marketplace regardless of how innovative the product What should we expect the impact on product performance
is and how much experience the firm can draw upon in its to be with regard to high levels of cooperation between
development. However, uncertainty, functional interdepen- marketing-R&D during later stages of the product develop-
dence, and the need for cooperation are likely to be greatest ment process and between marketing-operations, and R&D-
for the most innovative projects. Therefore, we anticipate operations during the earlier stages of the product develop-
that when a project team has little past experience to draw ment process when considered in terms of new product
upon (i.e., the development of highly innovative products) innovation levels?
the need for functional cooperation will be greater. Simi- It has been generally accepted that teamwork leads to
larly, the amount of relevant experience the project team can the development of better products. However, at least one
bring to bear should also moderate the efficiency of the study has demonstrated that this is not always the case
project. In projects requiring innovative problem solving [25]. This finding suggests that higher levels of cooper-
where the team has little relevant experience, high levels of ation may not always be beneficial, and may under some
interaction and functional cooperation throughout the circumstances prove detrimental. However, there is little
project may actually decrease total cycle time and develop- specific evidence or theory upon which to build an argu-
ment costs. The reason is that frequent sharing and feedback ment that high levels of cooperation between functional
of information, analyses, opinions, and advice may help the dyads at specific points in the new product development
various functions avoid incompatible decisions that lead to process might actually prove detrimental to the project’s
later conflicts and problems, and reduce the subsequent time
success. Therefore, we anticipate that higher levels of
and expense necessary to revisit and rework those decisions.
cooperation between marketing-R&D during later stages
On the other hand, in projects involving relatively mod-
of the process, and between marketing-operations, and
est changes to existing products where team members have
R&D-operations during the earlier stages of the process
a great deal of relevant past experience to rely upon, the risk
will generate no greater performance impact on the de-
of uninformed decisions leading to costly and time consum-
velopment of high innovative products when compared
ing subsequent revisions and recriminations are probably
with the development of low innovations products. As
lower. Therefore, the large expenditures of time and effort
such, we have adopted the null position in our final
required for high levels of functional cooperation early in
hypotheses. Namely:
the project are less likely to be recouped via reduced rework
and recycling later.
The above arguments are summarized in our hypotheses. H9a: The level of cooperation between marketing-R&D
Namely: during later stages of the product development process
(as predicted in H5b) will have no greater impact on
H8a: The positive impact on project performance from
innovative projects than for less innovative projects.
high levels of cooperation between marketing-R&D dur-
ing early stages of the product development process (as H9b: The level of cooperation between marketing-oper-
hypothesized in H5a) will be greater for innovative ations during early stages of the product development
projects than for less innovative projects. process (as predicted in H6a) will have no greater impact
on innovative projects than for less innovative projects.
H8b: The positive impact on project performance from
high levels of cooperation between marketing-operations H9c: The level of cooperation between R&D-operations
during later stages of the product development process during early stages of the product development process
(as hypothesized in H6b) will be greater for innovative (as predicted in H7a) will have no greater impact on
projects than for less innovative projects. innovative projects than for less innovative projects.
264 E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271

4. The study projects. We specifically requested projects that differed in


terms of their innovativeness, and which had been com-
4.1. Multiple respondents pleted within the past six to twelve months. A total of 34
project teams ultimately provided usable data from both the
The primary focus of this study is on the processes and project leader and team members from all three functions.
outcomes associated with individual NPD projects rather
than on the aggregate product development performance of 4.3. Measures
an entire firm or division. However, within each project we
were interested in examining variables that cut across two As mentioned, project managers were asked to assess
different levels of analysis. Many of our variables, such as their project’s performance on effectiveness and efficiency
the performance outcomes generated by the new product dimensions, and the amount of relevant past experience the
and the amount of relevant experience team members could firm could draw upon in developing and commercializing
bring to bear during its development, do reside at the project the new product. Project performance was assessed by five
level. But the study also focuses on the amount of cooper- items where the final product was rated on a 7-point Likert
ation between pairs of functions over the course of a project; scale. Effectiveness measures included: (1) the new prod-
a dyadic construct which operates at the interdepartmental uct’s quality relative to that of other products recently de-
level. veloped by the firm (2), management’s satisfaction with the
We adapted our research approach to the demands of product’s design and performance, and (3) the degree to
these multiple levels of analysis by obtaining data from two which sales objectives were reached. Project efficiency was
types of informants for each project. First, we collected examined on both financial and temporal dimensions. First,
information about the project’s innovativeness, the relevant managers were asked to assess on a 7-point Likert scale the
experience each function brought to the project, and its degree to which the project came in under or over its budget.
ultimate performance outcomes from a “project manager” They also assessed the project’s temporal efficiency by
or “team leader” identified by upper management as the evaluating the time required to complete the project relative
person with the most direct responsibility for the project. to the anticipated time frame for both (1) the early stages of
These project managers were personally interviewed by one the development process, including idea generation, concept
of the authors using a structured questionnaire. They were testing, and business assessment; and (2) the later stages of
also asked to identify project team members from R&D, the process, including prototype development, product test-
marketing, and operations. Those team members were then ing, and commercialization. These two items were com-
sent written questionnaires asking for information about the bined and used as an overall assessment of the project’s
perceived dependence and level of cooperation between cycle speed. While these items were originally envisioned
their department and each of the other two functions at both as separate measures for new product effectiveness and
early and late stages of the project. To encourage participa- efficiency, subsequent analyses of dimensionality and reli-
tion, all of the informants were assured that their responses ability revealed they were highly correlated. As a conse-
would be confidential and not shared with management or quence, we followed the lead of other researchers in this
other team members in disaggregated form, and they were area [14,17,26,31,33,34] and merged these individual mea-
promised an advanced copy of the study’s results. sures into a single aggregate measure of project perfor-
mance.
4.2. The sample Project managers were also asked to evaluate the inno-
vativeness of their projects by estimating the amount of
Given the number of informants and the sensitivity of experience their firms had with products similar to the ones
some of the data we sought, we solicited participation via being developed. They were asked to rate their firm’s level
personal contact with top management. Nine firms – mostly of relevant prior experience in terms of: (1) marketing and
members of the Fortune 500 – agreed to participate. A sales activities (2), manufacturing methods and activities,
diverse array of industries were represented, including both and (3) technology and R&D/engineering activities on a set
consumer and industrial products, such as electronic mea- of 7-point Likert scales. Because these experience measures
surement equipment, commercial food products, women’s were highly correlated across all three functional areas [as
fashions, commercial small engine appliances and equip- discussed in footnote #1 and shown in Table 1], the three
ment, high tech glass products, consumer health and beauty items were combined into a single measure of relevant prior
products, industrial water and filtration systems, office prod- experience.
ucts, and medical supplies. The firms ranged in age from 12 The three functional representatives on each team were
to over 100 years, with annual revenues from $50 million to asked to provide an assessment of the amount of coopera-
more than $10 billion. tion between their function and each of the other two de-
In order to increase the diversity of projects in the study partments during both the early and later stages of the
and help maintain the anonymity of the respondents, we project. Cooperation at each stage was measured with three
asked each participating firm for access to multiple NPD 7-point Likert scales concerning: (1) the frequency of com-
E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271 265

Table 1 Table 2
Multi-item scale reliability estimates and descriptive statistics Differences in informant perceptions of cooperation between pairs of
functions
Measure Items Coeffcient Mean Standard
alpha deviation Marketing Operations R&D
manager manager manager
Relevant prior experience 3 .97 3.95 2.23
Project performance 5 .88 4.11 .93 Marketing/operations—early 3.01 3.24
Dyadic cooperation measures Marketing/operations—late 3.94 4.19
Marketing/operations—early 3 .76 3.01 .78 Marketing/R&D—early 4.65 3.92*
Marketing/operations—late 3 .70 3.94 .80 Marketing/R&D—late 4.37 4.25
Marketing/R&D—early 3 .87 4.65 1.09 Operations/R&D—early 4.45 3.95*
Marketing/R&D—late 3 .79 4.37 .89 Operations/R&D—late 5.27 5.35
Operations/marketing—early 3 .83 3.24 .78
Operations/marketing—late 3 .87 4.19 .88 * Significant difference in means at the p ⬍ .05 level; n ⫽ 34.
Operations/R&D—early 3 .79 4.45 .82
Operations/R&D—late 3 .81 5.27 .75
R&D/marketing—early 3 .89 3.92 .85 factors was constrained to unity, and, in the other model, the
R&D/marketing—late 3 .80 4.25 .77 correlation between factors was freed 4. The results showed
R&D/operations—early 3 .87 3.95 .90 no significant difference between the ␹2 statistics (␹2 dif-
R&D/operations—late 3 .88 5.35 .92
ference ⫽ 0.03). Therefore, the results of both the explor-
atory and confirmatory factor analyses suggest that the
project performance measure exhibits unidimensionality.
munication between the two departments (2), the amount of
information, advice, and technical assistance they ex- 5.2. Consistency across respondents
changed, and (3) the level of work (e.g., raw materials,
specifications, and other resources) transferred. Given that we used multiple respondents to examine
cooperation levels between functions, an obvious question
is whether those respondents’ perceptions were consistent
5. Results with one another. As Table 2 indicates, there was strong
agreement, on average, between the members of each dyad
5.1. Reliability of the measures concerning their cooperation levels, with two exceptions:
marketing and operations respondents both perceived their
The summary measure of relevant prior experience, the functions’ cooperation with R&D during early project
measures of cooperation between functions, and the mea- stages to be significantly greater than did the R&D repre-
sure of project performance were all multi-item measures. sentatives. This suggests that R&D departments may per-
Coefficient Alpha was used to assess their internal reliabil- ceive themselves to be more independent and self-contained
ity. These coefficients – together with the mean and stan- during the early stages of NPD projects – with less need to
dard deviation for each measure – are shown in Table 1. For share information, advice, or assistance with other functions
most of our measures alpha exceeded 0.80, and all sur- – than either of the other two departments. This finding was
passed the minimum threshold of 0.70 recommended by even more surprising in view of the fact that relevant ex-
Nunnally [23]. perience levels were highly correlated across the three func-
In addition, a test of the unidimensionality of the project tions. In other words, R&D had no more (or less) experience
performance measure was performed. Since the measure to draw upon or share than did the other functions.
consisted of items that may be more indicative of a project’s
In spite of these two differences in the magnitude of
effectiveness, such as sales and product quality, and other
cooperation perceived by R&D representatives versus ev-
items that may be more indicative of the project’s effi-
erybody else when those perceptions were averaged across
ciency, such as timeliness and budget adherence, this test
all projects, correlations of cooperation scores among the
was deemed necessary. First, we performed an exploratory
factor analysis with all the project performance items. The individual members of each project team were positive and
results indicated only a single factor with an eigen value statistically significant. The correlations were 0.45 (p ⬍ .01)
greater than 1. Second, using LISREL, we compared the between marketing and R&D team members, 0.41 (p ⬍ .05)
results of two 2-factor confirmatory factor analyses. One between marketing and operations, and 0.53 (p ⬍ .01)
factor consisted of the effectiveness items: quality relative between R&D and operations. Thus, for analytical parsi-
to that of other products, management’s satisfaction with the mony and clarity of discussion, cooperation scores from
product’s design and performance, and achievement of sales each pair of functional respondents on a team were averaged
objectives. The second factor consisted of the efficiency to produce a single cooperation measure for the dyad. Such
items: under or over project budget, and time required to measures were calculated for both early and late project
complete project. In one model, the correlation between the stages.
266 E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271

Table 3 Table 4
Patterns of functional cooperation across project stages The influence of functional cooperation and project performance:
Step-wise regression analysis
I. Cooperation across all functions (n ⫽ 102) (t-value beta coefficients shown with (p-values)):
Average cooperation score: Average cooperation score: t-Value
early stages late stages Cooperation between: Project performance

Marketing/R&D—early 3.20 (.003)


3.87 4.57 10.31**
Marketing/R&D—late *
II. Proportion of total cooperation accounted for by each pair of Marketing/operations—early *
functions Marketing/operations—late 2.16 (.039)
% of total % of total t-Value R&D/operations—early 2.05 (.049)
cooperation score: cooperation score: R&D/operations—late *
early stages late stages Adjusted R2 .55

* Insignificant coefficient estimates at the p ⬍ .05 level.


Marketing—R&D .37 .31 ⫺7.71**
Marketing—operations .27 .30 3.56**
R&D—operations .36 .39 3.69**
early and late project stages as independent variables, and
** ⫽ p ⬍ .001. project performance as the dependent variable. The results
of this analysis across all 34 projects are shown in Table 4.
5.3. Patterns of functional cooperation
5.4a. Cooperation between marketing and R&D. We pre-
dicted in H5a that high levels of cooperation between mar-
Overall cooperation across stages. As expected, average
keting and R&D during the early stages of a project would
levels of cooperation across the three functions increased as
be positively related to project performance. On the other
NPD projects moved from the early concept development
hand, H5b argued that marketing-R&D cooperation levels
stages to the more information-rich production and com-
during later project stages would be unrelated to project
mercialization stages. The top section of Table 3 indicates
effectiveness. The data in Table 4 strongly support both sets
that average cooperation scores increased from 3.87 in the
of predictions.
early project stages to 4.57 in later stages (t ⫽ 10.31;
p ⬍ .001), providing support for H1.
Cooperation between pairs of functions. As anticipated, 5.4b. Cooperation between marketing and operations.
overall levels of functional cooperation increased from early Based on the assumption that the competencies and exper-
to later project stages. However, we predicted that the rel- tise of operations personnel would be most relevant for
ative amount of cooperation between different pairs of func- resolving uncertainties arising during the later stages of
tions would differ across stages. In other words, different NPD projects, we predicted (H6b) that cooperation levels
functions are expected to play a role in resolving different between marketing and operations during those later stages
kinds of uncertainty at different stages of an NPD project. would be significantly related to project performance, but
Thus, the dependence of one functional department on an- (H6a) that cooperation during the earlier stages would bear
other, and its motivation to cooperate with that other de- no relationship to project performance. As indicated in
partment, are likely to vary across stages. Table 4, both hypotheses were supported.
The second section of Table 3 displays the relative pro-
portion of total functional cooperation accounted for by 5.4c. Cooperation between R&D and operations. For rea-
each pair of functions at both early and later project stages. sons similar to those outlined in the previous section, we
As expected, marketing-R&D accounted for a larger pro- predicted (H7b) that cooperation levels between R&D and
portion of total cooperation in the early stages than in the operations during the later project stages would be signifi-
later stages. On the other hand, marketing-operations and cantly related to project performance, but (H7a) that their
R&D-operations both accounted for a larger proportion of cooperation during the earlier stages would bear no rela-
the functional cooperation that occurred during later stages tionship to project performance. As indicated in Table 4,
than earlier in the project. These findings support the pre- neither hypothesis was supported. Counter to our prediction,
dictions of H2, H3, and H4. higher levels of cooperation between R&D and operations
during early stages of the development process proved to be
5.4. Patterns of functional cooperation and project significantly related to project performance.
performance
5.5. The moderating effects of project innovativeness
To examine whether different patterns of cooperation
between pairs of functions are related to project perfor- While high levels of cooperation between marketing -
mance, we conducted a series of step-wise regressions treat- R&D and operations - R&D during early project stages
ing cooperation scores for each pair of functions at both were positively related to project performance, the unan-
E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271 267

Table 5 Table 6
The influence of the interaction between project innovativeness and The influence of cooperation on project performance at low and high
functional cooperation on project performance levels of project innovativeness
(t-value beta coefficient estimates shown): (simple slope coefficient estimates at one standard deviation above and
below mean of project innovativeness)
Cooperation variables Project performance
Cooperation variables High project Low project
Full Reduced
innovativeness innovativeness
interaction interaction
t-values t-value
model I model II
Marketing/operations—early ⫺2.177* 2.923**
Project innovativeness ⫺2.092* ⫺2.007*
Marketing/operations—late 3.394** ⫺.58
Mktg.-R&D (early) 2.059* 3.135**
R&D/operations—late 4.395** ⫺.873
Mktg.-R&D (late) .338
Adjusted R2 .747 .747
Mktg.–ops. (early) ⫺.13 ⫺.928
Mktg.-ops. (late) 1.266 1.873 The results from six separate regressions following the Aiken and West
R&D-ops. (early) ⫺.069 (1991) procedure for each of the three significant interactions from Table
R&D-ops. (late) 2.586* 2.947** 5 at low and high levels of project innovativeness.
Project innovativeness ⫻ Mktg-R&D (early) .145 * p ⬍ .05.
Project innovativeness ⫻ Mktg-R&D (late) .46 * p ⬍ .01.
Project innovativeness ⫻ Mktg.-ops (early) ⫺1.627 ⫺3.281**
Project innovativeness ⫻ Mktg.-ops (late) 2.254* 2.802**
Project innovativeness ⫻ R&D-ops (early) ⫺.268 5.5a. Cooperation between marketing and R&D. We argue
Project innovativeness ⫻ R&D-ops (late) 2.596* 4.046** that the increased uncertainty and functional interdepen-
Adjusted R2 .708 .747 dence associated with innovative NPD projects where firms
* p ⬍ .05. have little relevant prior experience to draw upon will likely
** p ⬍ .01. strengthen the impact of cooperation between pairs of func-
tions during the most crucial project stages. In the case of
marketing and R&D, we predicted in H8a that cooperation
swered question is whether those relationships are moder- between the two functions during early project stages would
ated by an individual project’s characteristics; particularly be more positively related to project performance in rela-
by the project’s innovativeness. To answer this question, we tively innovative projects than in less innovative ones. The
tested for the interaction effects of cooperation and project results reported in Table 5 fail to support this prediction.
innovativeness on project performance. More specifically, Early-stage cooperation between marketing and R&D is not
we first regressed the project performance variable on the significantly more important for highly innovative projects
cooperation variables plus all of the six interactions vari- than for less innovative projects. However, as predicted in
ables formed by multiplying the cooperation variables by H9a, marketing-R&D cooperation during later project
the project innovativeness variable. The results of this re- stages is not related to project performance for either inno-
gression are shown in Table 5 under the full interaction vative or noninnovative projects.
model column (Model I). We then removed insignificant
interaction and direct cooperation variables, and ran a sec- 5.5b. Cooperation between marketing and operations. In
ond regression. The results are shown in Table 5 under the the case of marketing and operations, we predicted in H8b
reduced interaction model column (Model II). Note that the that cooperation between these two functions during the
insignificant marketing-operations (early) cooperation vari- later project stage would be more positively related to
able was not removed from the reduced regression model project performance in relatively innovative projects than in
because there was a significant interaction found between less innovative ones. The results reported in Table 6 provide
this variable and project innovativeness. Leaving the direct support for this prediction. At high levels of project inno-
effects term in the regression model strengthens the test of vativeness, late-stage cooperation between marketing and
the interaction term coefficient. operations is positively related to project performance (␤ ⫽
In addition to testing for interaction significance, we also 3.39; p ⬍ .01), and, at low levels of project innovativeness,
performed the Aiken and West [1] procedure to test the late-stage cooperation was not related to project perfor-
effects of cooperation on project performance at high and mance. Therefore, H8b is supported. Surprisingly, while we
low levels of project performance. This was done for each predicted in H9b that cooperation between marketing and
of the three significant interactions found in the reduced operations during the early project stages was not related to
model regression results. This procedure entails testing for project performance for either innovative or noninnovative
the significance of the coefficient of the cooperation vari- projects, the data indicate that cooperation is actually neg-
ables at both one standard deviation above the mean of atively related to project performance (␤ ⫽ ⫺2.18; p ⬍ .05)
project innovativeness, and at one standard deviation below for innovative projects. However, for noninnovative
the mean of project innovativeness. The results of this series projects, cooperation between marketing and operations
of regressions are reported in Table 6, and will be related to early in the project was found to be positively related to
the hypotheses in the following sections. (␤ ⫽ 2.92; p ⬍ .01) project performance.
268 E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271

5.5c. Cooperation between R&D and operations. We pre- 6.2. Functional cooperation and performance outcomes
dicted in H8c that cooperation between R&D and operations
during the later project stages would be more positively While average cooperation patterns across pairs of func-
related to project performance in innovative projects than in tions largely conformed to the expectations of resource
less innovative ones. The results in Table 6 indeed indicate dependence theory, those patterns did not always bear the
that late-stage cooperation between R&D and operations in expected relationship to performance of a given NPD
innovative projects is positively related to project perfor- project. As we predicted, substantial cooperation between
mance (␤ ⫽ 4.395; p ⬍ .01) while the relationship between marketing and R&D during the early stages is strongly
late-stage cooperation in noninnovative projects was not associated with positive project performance. Counter to
significant. Therefore, H8c is supported. We also predicted our expectations, substantial cooperation between R&D and
operations during the early stages is also positively related
in H9c that cooperation between R&D and operations dur-
to project performance. When examined across all projects,
ing the early project stages would have no greater effect on
the data suggest that higher levels of cooperation between
project performance in innovative projects than in low in-
functions generates stronger new product performance than
novation projects. This hypothesis was also supported.
does lower levels of cooperation. However, there are certain
times when cooperation between specific functions is more
critical to the project’s ultimate performance than other
6. Discussion and implications times.
The most obvious conclusion to be drawn from this 6.3. The moderating effect of project innovativeness
study’s findings is that the relationship between cross-func-
tional cooperation and NPD project performance is more Another conclusion that can be drawn from this study’s
complex than the simple “more is better” prescription found findings – one that has both conceptual and managerial
in most previous studies. Instead, the impact of cooperation implications – is that the innovativeness of a given project,
on performance depends on which functions cooperate, at and thus the amount of relevant past experience the firm can
what stage in the development process they cooperate, and draw upon, influences the relationship between particular
on the innovativeness of the project itself. patterns of cross-functional cooperation and project perfor-
mance. While our detailed hypotheses concerning the mod-
6.1. Patterns of functional cooperation erating effects of project innovativeness received only par-
tial support, we nonetheless observed substantial
Across all projects, patterns of functional cooperation differences between innovative and less innovative projects.
largely conform to what one would expect based on the
tenets of resource dependency theory. For one thing, overall 6.3a. Patterns of cooperation and performance in innova-
levels of functional inter dependence – and therefore cross- tive projects. The importance of when and the degree to
functional cooperation – increase as a project moves from which functional groups involved in the development of a
its conceptual early phases to more concrete later stages new product cooperate appears to be impacted by the degree
where there is a greater need for information sharing, feed- of innovation in the new product being developed. How-
back, and coordination of activities and decisions across ever, not necessarily in the ways we might have expected.
functions. Indeed, high levels of cooperation between marketing and
More interestingly, the relative amount of cooperation R&D, and between operations and R&D at early stages, and
between marketing and R&D at late stages of the product
between specific pairs of functions changes as a project
development process proved no more important in deter-
evolves and different kinds of information and expertise are
mining project performance for high innovation projects
needed to resolve the various types of uncertainty that
than they did for low innovation projects. Of course this
emerge. Given that marketing and R&D tend to be reposi-
does not mean that high levels of cooperation are not im-
tories of customer, competitor, and technical information
portant in high innovation projects, just that they appear to
and expertise that are particularly germane during the early not be more so than in low innovation projects.
conceptual stages of a development project, those two func- However, that dynamic changes during the later stages of
tions account for a relatively large portion of functional the process as greater levels of cooperation between mar-
cooperation during those stages. On the other hand, both keting and operations, and between R&D and operations
marketing and R&D depend on operations for expertise appears to be a critical factor in determining project perfor-
relevant for resolving the process engineering, product qual- mance in high innovation projects. Once again, this does not
ity, and production cost and volume issues typically en- mean that cooperation between members of these dyadic
countered later in a project. And as expected, relative co- pairs is not without some importance in low innovation
operation levels between operations and the other two projects, only that it is not as critical a determining factor in
functions increased significantly during later project stages. project performance as it is for high innovation projects.
E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271 269

Perhaps the most surprising finding of this study is the 7. Limitations and future research directions
observation that high levels of cooperation between market-
ing and operations during early stages of the product devel- As is usually the case with survey research, this study has
opment process were actually associated with significantly several limitations. First, our sample – though diverse in the
lower levels of project performance in high innovative types of products and firms represented – is relatively small.
projects and, in marked contrast, significantly higher levels This was in part a consequence of our determination to
of project performance for low innovation projects. This include only projects for which we were able to obtain
finding suggests that in low innovation projects it is very completed questionnaires from all three functional respon-
important for marketing and operations personnel to get dents as well as from the project manager. We felt that this
together at the outset of the development process in order to was an acceptable trade-off, however, given (1) that the full
determine the most efficient way to produce the desired range of functions was seldom represented in previous re-
modifications and then, as the project progress to later search, and (2) our desire to minimize perceptual bias in the
data.
stages, to get out of each other’s way. In contrast, it appears
The use of a cross-sectional methodology to examine a
that during the development of a high innovation projects it
process which evolves through multiple stages over time is
is best to let marketing and R&D determine the market’s
another obvious limitation which constrained our ability to
latent needs and the basic technology required to address
infer the direction of causality. Similarly, asking respon-
these needs before operations becomes enmeshed in the dents to provide retrospective evaluations introduced the
details of how to actually produce such a product. Once possibility of both the simple forgetting of relevant infor-
product parameters have been set it becomes very important mation as well as the possible posthoc justification for either
for R&D and marketing to cooperate with operations. project success or failure. However, as mentioned, we at-
While our data do not provide a definitive explanation as tempted to reduce the possibility of perceptual bias by using
to why these different dyadic patterns of functional coop- multiple respondents from each project, and by collecting
eration are related to superior performance in innovative performance data from different respondents than the data
projects, one possibility is that paying too much attention to on functional cooperation. Nevertheless, our understanding
operational concerns and practicalities early in a project of how and why different temporal patterns of functional
may constrain the conceptual flexibility and creativity of the cooperation affect NPD effectiveness and efficiency would
project team. Such constraints, in turn, may result in a clearly benefit from future research employing larger sam-
product which is less technologically advanced and/or less ples and longitudinal methods.
responsive to customer needs than it could have been.
The implications of these findings for improving the 7.1. Future research
management of innovative NPD projects are also not en-
tirely clear. Operations – like the marketing and R&D func- This study also raised more interesting questions than it
tions – embodies knowledge and expertise that is clearly was able to answer. First, while this research focused on the
essential to the ultimate success of any NPD project and overall level of cooperation among functions during the
should therefore be represented on the core project team. NPD process, it tells us little about the nature or behavioral
But when a project is highly innovative, our results suggest content of that cooperation. Given the recent emphasis on
that either: (1) such representation should not begin until the market-oriented management and the development of
conceptual direction of the project is fairly well established, “learning” organizations [10,30,31], it might be particularly
useful for future research to examine the role of each func-
or (2) the team’s decision processes should be managed in
tion in tying the NPD team to interested parties and sources
some way such that cooperation between operations and the
of expertise – such as customers, suppliers, or distributors –
other functions – and the amount of decision influence
outside the boundaries of the firm, as well as across internal
wielded by operations – begin at relatively low levels but
departments. In other words, cooperation might profitably
expand over the course of the project. One way in which this be extended beyond the members of a project team and the
might be accomplished is to employ a modified stage-gate functional departments they represent to include other ex-
procedure which includes fewer operations-related criteria – ternal constituents with whom a function has strong ties.
or assigns such criteria lower weights – for winning ap- Thus, functional representatives may play important roles as
proval for the project to advance through the early “gates,” gatekeepers and coordinators of knowledge flows among –
but which increases the number or weight of such criteria at as well as within – organizations [2,24].
later decision points. Of course, certain industries may rely More critically, while this study provides some prelimi-
significantly on innovations from manufacturing, and, as a nary insights into the patterns of functional cooperation
result, these suggestions should be tempered in such cases. associated with the performance of different types of NPD
An interesting follow-on research study might look at fac- projects, it says nothing about how managers might best
tors that might increase the need for stronger manufacturing encourage or stimulate the desired patterns of cooperation.
involvement in early stages of innovative projects. While our results suggest that increased cooperation be-
270 E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271

tween marketing and operations representatives during the Acknowledgment


earliest stages of relatively incremental, noninnovative
projects is strongly related to superior project performance, The authors gratefully thank the Marketing Science In-
they offer no clues concerning the ways in which managers stitute for its financial support of this project.
might improve upon the currently very low levels of coop-
eration between those two functions in such projects. What
is needed, then, is research focused on how different coor- References
dination mechanisms, management interventions, tracking
[1] Aiken LS, West SG. Multiple Regression: Testing, and Interpreting
and assessment procedures, and reward and incentive poli-
Interactions, Newbury Park. Sage, 1991.
cies impact the frequency, timing and content of coopera- [2] Ancona DG, Caldwell DF. Bridging the boundary: external activity
tion between particular functions at particular stages in the and performance in organizational teams. Administrative Science
development process. Quarterly 1992;37:634 – 65.
Finally, the construct of project innovativeness – and the [3] Aldrich HE, Pfeffer J. Environments of organizations. Annual Re-
view of Sociology 1976;2.
closely related notion of relevant prior experience – were [4] Bagozzi R, Yi Y, Phillips LW. Assessing Construct Validity in
shown to have a moderating effect on how functional co- Organizational Research. Administrative Science Quarterly 1991;36:
operation was related to project outcomes. It would there- 421–58.
fore be useful to have a more comprehensive understanding [5] Burns T, Stalker G. The Management of Innovation. London: Tavis-
of the domain of “relevant” experience for each function, tock Publications, 1961.
[6] Carroad PA, Carroad CA. Strategic interfacing of R&D and market-
and of which aspects of those domains most strongly impact ing. Research Management 1982;25:28 –33.
cooperation patterns and project outcomes. How is experi- [7] Clark KB, Fujimoto T. Product Development Performance: Strategy,
ence and expertise shared and processed by development Organization, and Management in the World Auto Industry. Boston:
teams? How is such information stored within organizations Harvard Business School Press, 1991.
[8] Cooper RG, Kleinschmidt EJ. Benchmarking the firm’s critical suc-
for use in future projects, and how is it judged to be relevant
cess factors in new product development. Journal of Product Inno-
or useless for resolving particular issues on those projects? vation Management 1995;12:374 –91.
Such questions are central to improving our understanding [9] Ettlie JE. Integrated design and new product success. Journal of
of how innovation happens and how it may be helped or Operations Management 1997;15:33–55.
hindered by cooperation among functional specialists. [10] Garvin DA. Building a learning organization. Harvard Business Re-
view, 78 –91 (July/August, 1993).
[11] Griffin A. Metrics for measuring product development cycle time.
Journal of Product Innovation Management 1993;10:112–25.
[12] Griffin A. PDMA research on new product development practices:
Notes updating trend, and benchmarking best practices. Journal of Product
Innovation Management 1997;14:429 –55.
[13] Griffin A, Hauser JR. Patterns of communication among marketing,
1. The following discussion implicitly assumes that all
engineering, and manufacturing – a comparison between two new
three functions have similar amounts of relevant product teams. Management Science 1992;38:360 –73.
knowledge, and experience to bring to bear on a given [14] Griffin A, Hauser JR. Integrating mechanisms for marketing and
NPD project. Thus, projects that are innovative, and R&D. Journal of Product Innovation Management 1996;13:191–215.
unusual in a technical sense are expected to present [15] Gupta AK, Raj SP, Wilemon D. The R&D-marketing interface in
high-technology firms. Journal of Product Innovation Management
new challenges, and uncertainties not only to the R&D
1985;2:12–24.
function but to the operations, and marketing people [16] Gupta AK, Raj SP, Wilemon D. A model for studying the R&D-
as well. Similarly, products that are simple modifica- marketing interface in the product innovation process. Journal of
tions or improvements on an established technological Marketing 1986;50:7–17.
platform are not expected to be seen as very innova- [17] Henard DH, Szymanski DM. Why are some new products more
successful than others? A meta-analysis of empirical evidence. Un-
tive or challenging from an operations or marketing published working paper, Texas A&M University, College Station,
perspective either. This assumption was empirically Texas; 1999.
validated by examining the correlation among the [18] Kahn KB. Interdepartmental integration: a definition with implica-
“relevant prior experience” scores for: (1) marketing tions for product development performance. Journal of Product In-
novation Management 1996;13:137–51.
and sales activities, (2) manufacturing methods and
[19] Kahn KB, McDonough EF. An empirical study of the relationships
activities, and (3) technology and R&D activities. As among co-location, integration, performance, and satisfaction. Jour-
shown at the top of Table 1, the correlation among nal of Product Innovation Management 1997;14:161–78.
these scores was .97. Note, too, that the projects we [20] Lawrence PR, Lorsch JW. Organization, and Environment: Managing
studied were broadly distributed in terms of the Differentiation, and Integration. Boston: Harvard University Press,
1967.
amount of relevant experience available within the [21] Madhavan R, Grover R. From embedded to embodied knowledge:
firm. On a 7-point scale the mean experience score new product development as knowledge management. Journal of
was near 4, with a standard deviation slightly over 2. Marketing 1998;62:1–12.
E.M. Olson et al. / The Journal of Product Innovation Management 18 (2001) 258 –271 271

[22] Menon A, Jaworski BJ, Kohli AK. Product quality: impact of inter- [38] Souder WE, Song SM. Analyses of US and Japanese management
departmental interactions. Journal of the Academy of Marketing Sci- processes associated with new product success and failure in high and
ence 1997;25:187–200. low familiarity markets. Journal of Product Innovation Management
[23] Nunnally JC. Psychometric Theory. New York: McGraw-Hill Co., 1998;15:208 –23.
1978. [39] Urban GL, Hauser JR. The Design, and Marketing of New Products.
[24] O’Connor GC. Market learning and radical innovation: a cross case Englewood Cliffs, N.J: Prentice-Hall, Inc., 1993.
comparison of eight radical innovation projects. Journal of Product [40] Van de Ven AH, Ferry DL. Measuring, and assessing organizations.
Innovation Management 1998;15:151– 66. New York: John Wiley, and Sons, 1980.
[25] Olson EM, Walker OC, Ruekert RW. Organizing for effective new [41] Wheelwright SC, Clark KR. Revolutionizing Product Development:
product development: the moderating role of product innovativeness. Quantum Leaps in Speed, Efficiency, and Quality. New York: The
Journal of Marketing 1995;59:48 – 62. Free Press, 1992.
[26] Pfeffer J, Salancik GR. The External Control of Organizations: a Re-
source Dependence Perspective. New York: Harper, and Row, 1978.
[27] Pinto MB, Pinto JK. Project team communication and cross func- Biographical Sketches
tional Cooperation in new program development. Journal of Product Eric M. Olson (Ph.D., University of Minnesota) is Professor and Chairman
Innovation Management 7: 200 –212(1990); of the Marketing, Strategy, and International Business Department at the
[28] Rosenau MD. Phased approach speeds up new product development. University of Colorado at Colorado Springs. His research has focused on
Research & Development 1988;30:52– 6. marketing strategy, functional interaction in new product development, and
[29] Ruekert RW, Walker O. Marketing’s interaction with other functional design management. His research has been published in the Journal of
units: a conceptual framework, and empirical evidence. Journal of Marketing, Strategic Management Journal, Harvard Business Review,
Marketing 1987;51:1–19. Business Horizons, Design Management Journal, and Marketing Manage-
[30] Slater SF, Narver JC. Market orientation and the learning organiza- ment where he serves as the Case-In-Point section editor.
tion. Journal of Marketing 1995;59:63–74.
[31] Slater SF, Narver JC. Customer-Led, and Market-Oriented; Let’s Not Orville C. Walker, Jr. (Ph.D., Ohio State University) is the James D.
Confuse the Two, Strategic Management Journal, 19 (10), 1001– Watkins Professor of Marketing and the Director of the Ph.D. program in
1006, 1998. the Carlson School of Management at the University of Minnesota. His
[32] Song XM, Parry ME. The R&D-marketing interface in Japanese research has focused on marketing strategy and sales force management
high-technology firms. Journal of Product Innovation Management and has been published in the Journal of Marketing, Journal of Marketing
1992;9:91–112. Research, and Strategic Management Journal among others. He has coau-
[33] Song XM, Parry ME. R&D-marketing integration in Japanese high- thored three books on these topics. He is the recipient of the Journal of
technology firms: hypotheses, and empirical evidence. Journal of the Marketing Research’s O’Dell award.
Academy of Marketing Science 1993;21:125–33. Robert W. Ruekert, (Ph.D., University of Wisconsin) is Professor of
[34] Song XM, Parry ME. What separates Japanese new product winners Marketing and Associate Dean in the Carlson School of Management at the
from losers? Journal of Product Innovation Management 1996;13: University of Minnesota. His research has focused on issues of marketing
422–39. strategy and management and been published in the Journal of Marketing,
[35] Song XM, Montoya-Weiss MM, Schmidt JB. Antecedents and con- Journal of Marketing Research, and Strategic Management Journal among
sequences of cross-functional cooperation: a comparison of R&D,
others. Professor Ruekert and Professor Walker shared the 1987 Maynard
manufacturing and marketing perspectives. Journal of Product Inno-
Award for outstanding theory development article in the Journal of Mar-
vation Management 1997;14:35– 47.
keting.
[36] Song XM, Thieme J, Xie J. The impact of cross-functional joint
involvement across development stages: an exploratory study. Jour- Joseph M. Bonner (Ph.D., University of Minnesota) is Assistant Professor
nal of Product Innovation Management 1998;15:289 –303. of Marketing and Supply Chain Management in the Eli Broad College of
[37] Souder WE. Managing relations between R&D and marketing in new Business at Michigan State University. In 1998 he was a recipient of a
product development projects. Journal of Product Innovation Man- Juran Fellowship at the University of Minnesota. His research emphasis is
agement 1988;5:6 –19. in the area of new product development.

Das könnte Ihnen auch gefallen