Beruflich Dokumente
Kultur Dokumente
ISBN/EAN: 978-90-814105-1-9
Table of content
1 Introduction 14
2 Business value of BI 18
2.5 Track performance and align metrics across the organisation ..................................................................... 28
3.1 BI Foundation................................................................................................................................................ 36
4 Managing BI 48
5.
1
BISt
rat
egy
5.
2
BIDef
ini
ti
on
5.
3
BIDevel
opment
5.
4
BIEx
ploi
tat
i
on
6 BISol
uti
onEngi
neer
ing Avai
labl
einApr
il2010
6.
1
Benef
it
soft
heBIEngi
neer
ingFr
amewor
k
6.
2
Connec
ti
ont
otheBIl
i
fecyc
le
6.
3
St
akehol
derper
spect
iv
es
6.
4
Engi
neer
ingdi
sci
pli
nes
6.
5
TheBIEngi
neer
ingFr
amewor
kModel
7 BestPr
act
i
ces Avai
labl
einMay2010
7.
1
DWHAr
chi
tec
tur
eeval
uat
i
on
7.
2
ETLFr
amewor
k
7.
3
Dat
aQual
i
ty
7.
4
Dat
aVaul
t
7.
5
Di
mens
ional
model
l
ing
7.
6
Dat
aandTex
tMi
ning
8 Fer
rar
iCas
e Avai
labl
einApr
il2010
8.
1
I
ntr
oduct
ion
8.
2
St
rat
egyst
udy
8.
3
Feasi
bil
i
tySt
udy
8.
4
Requi
rement
sAnal
ysi
s
8.
5
Demons
trat
i
on
8.
6
BIAr
chi
t
ect
ure
8.
7
Dat
awar
ehouseReal
i
sat
i
on
8.
8
Repor
ti
ngandanal
yt
icsReal
i
sat
ion
8.
9
Dat
aQual
i
ty
8.
10
Dat
aVaul
tModel
l
ing
8.
11
Di
mens
ionalModel
l
ing
8.
12
Expl
oit
ati
on
10
Preface
Dear reader,
• Do I have the same common language and
Welcome to our newest version of the Logica Business definition of indicators throughout my company?
Intelligence Framework. It has come about thanks to • How do I explain the need for so many reports,
the global collaboration between multicultural, multi dashboards, etc.?
specialist and very experienced people dedicated to • Is my Business Intelligence as mature as my ERP?
improving organisational performance enabled by CPM • If I was to make a call to the helpdesk for every
and BI. When we listen to our customers, or look at our data quality issue, would I need to double the size
own company’s organisation, we still face some of the of the helpdesk?
traditional challenges:
• We have many ways of producing metrics. Today, our goal is to give you our latest vision of the Logica
• We do not have a common language within each Business Intelligence Framework, and so help you to take
part of our organisation. full responsibility for some of the challenges listed above,
• We provide alignment in mergers but we have and some of the other challenges that have to be faced.
experienced challenges in supporting them. We face them knowing that it will not be easy. We have
• We have a wealth of data but have yet to determine shaped visions, crafted strategies, scoped and delivered
what value it adds and how to deal with the quality projects, and then we completely outsourced them; this
control. work represents millions of man hours. Our challenge
is to leverage our former common language that we
I could keep going, and it is an easy job to list all of the used for Business Intelligence and to make it a source of
challenges, but that does not help create value. We additional value for our customers and our people. The
consider every project that involves Business Intelligence main difference is that it is not a book written by a few of
to be an opportunity for transforming and improving our experts but a collaborative effort. It is also the result
the organisation. It is part and parcel of the change of different practices that have been merged together and
management process. Since the 90’s we have been the blended experience from a variety of our constituent
delivering process-oriented projects using pre-built countries. They have been challenged by our best people
applications, such as ERP-related solutions. This greatly and customers; the practices that are used are present
helped to rationalise behaviour involved in such a process, because we are confident that they deliver results in
gave us a sense for how IT departments and new maturity today’s global environment.
levels operate with an industrialised approach, but in
many cases lacked the impact it had on the associated So what can I expect from the Logica Business
decision-making process. If we need proof of this, we just Intelligence Framework? A quick win could be to
need to ask ourselves a few questions. benchmark an approach, a vision, the performance of a
Business Intelligence Competence Center, or simply your
operational costs. Another possibility could be to use the
framework in your own effort of aligning your organisation,
thus avoiding the need of making the investment yourself.
Stéphane Jaubert,
Managing Director
Business Intelligence
Global Practice Leader
11
1
14
1 Introduction
In a world of uncertainty and economic downturn Another important aspect to consider in organising BI is
Business Intelligence has become critical to survive. the continuously changing demand of the organisation for
The importance of monitoring and measurement of the information. That is why we provide a structured approach
performance of your organisation, knowing the customers to manage the BI lifecycle in an cost effective and future
behaviour and complying with regulatory demands has proof manner. Then we will dive deeper into the design
become evident. Well implemented BI solutions deliver and development aspects, presenting our BI engineer
competitive advantage to an organisation. framework. We will end the book with some of the best
practices collected in our Bi community over the years.
With this book we introduce the value of business
intelligence to achieve your business objectives. To give All together a lot of material to read, aimed at a broad
you an understanding of how to achieve this business audience of business management, BI management and
objectives with the use of BI we provide you with a BI practitioners. The table below provides some guidance
clear understanding of what BI is. Knowing that most to what chapter is most relevant for what part of the
organisations already have some BI solutions in place we audience.
focus on cost effective management of BI and provide you
with a clear roadmap on how to lower the Total Cost of
ownership of the current landscape and how to utilise BI
to the best advantage of your organisation.
2. Business value of BI
4. Managing BI
5. BI Lifecycle
6. BI Solution Engineering
7. Best Practices
Recommended, Optional
m An
fra a
fo er fo Increase customer p anced s va
lys
ce
te
or
cu
pl
m
s
st
co
om
ine
m
d
er
an
2
be
isk
m a n a g e m e nt B u s
e nt
ha
y R
vio
ur
bilit
a l
rof i t a
B
O p e r a t i o n a l exc e ll e n c e P ro d
p et i t i v e a
m a n c e
pe p c h u r n
r
e
u
ct le
anc
c s
ers ad
e k e
3
m
c uc
hip
6
r
C
ed
t
0
a
°
r
g
e v
ers Tra at er Trac ew
i
or
s
tric
me
ign
c u s to m in t k of
m r is k a n
or p acy d complian ce Al C
i
L i fe t i m e va l u e
18
2 Business value of BI
This section provides an introduction to Business
Intelligence from a business perspective. Too often we
find Business Intelligence initiatives at customer sites that
are very IT oriented. These initiatives started with good
intentions and for very good reasons. Somewhere down
the road however, business focus is lost and technology
pushes the initiative forward. With this section we offer
some guidelines to keep your Business Intelligence
initiative business-driven. We start with an understanding
of business value based on a scientific model, the
Treacey and Wiersema business value discipline model,
and a pragmatic model, the Logica Enterprise Value
Management model. We define the market pull of
Business Intelligence based on current market intelligence.
We finish this section with a more detailed definition of
common business drivers for a Business Intelligence
initiative in an organisation today, including their
stakeholders and real life customer cases.
2.1 Enterprise Value Management
A business case for Business Intelligence should decision authority to employees that are close to the
always be related to the business strategy and business customer. A typical example of a company with this
objectives of an organisation. To get an understanding of type of strategy would be Rolls Royce.
the business strategy of an organisation we use the value
discipline model of Treacy and Wiersema, as shown in
figure 2.1.
In this model three different strategies for an organisation Best
Product
are characterised:
Business value of BI
19
20
and quality. A BI solution for the marketing organisation case - containing benefits, costs and the ROI - is a
of such an organisation will indentify customers that are critical success factor in mobilising key players to drive
looking for the low prices in the first place. An organisation the initiative. The business goals and the strategy define
that focuses on operational excellence will probably invest what the organisation wants to reach and how this
less budget, and therefore a high end solution will not should be accomplished. The strategy will be in line with
be realistic. It is however critical for any organisation to the value discipline that the organisation chooses for. A
monitor and manage its enterprise values. These values business plan - complying to the goals and strategy - is
are based on the business strategy in combination with specified in which objectives and policies are set. Budgets
the risk and compliancy demands, every organisation is are provided to the business functions throughout
facing today. The Logica Enterprise Value Management the organisation. For a compelling business case it is
(EVM) model, as shown in figure 2.2, enables you to necessary that the anticipated benefits contribute to the
do so. This model balances business risks (KRI’s) and objectives of the corresponding business owner while the
costs stay within the budget. If more than one business
owner is involved in a BI initiative, then it is necessary to
find a business owner at a higher level in the organisation.
In that case it becomes important to link the benefits to
Engineer Manage Measure
Value Change Performance the business strategy as well because a common goal
must be found. Also, it is very difficult to implement a
Organisational Model
KRI’s
& Primary Measures KPI’s demanding initiative, while the people involved are focused
on something different. Therefore, the BI initiative should
Manage Ensure Assure also be in line with the chosen value discipline or business
Risk Continuity Compliance priority.
MS
– market leaders across a wide range of industries
1500
have to build and enhance their knowledge of
1000
customers’ needs, preferences and behaviour.
500
zz Organisations need to restructure quickly to cope with
0
changing market conditions. The speed at which they 2005 2006 2007 2008 2009 2010 2011 2012
Business value of BI
21
22
BI implementations rarely meet these expectations. Based on these demands we identified a number of
Organisations are looking for, and desperately need, BI business drivers that justify BI initiatives nowadays:
solutions which can deliver information based on a ‘single
view of the operational data’, rather than multiple views zz Track risk and compliance;
and therefore inconsistent information. Enable an organisation to be accountable and
‘in control’ by providing auditable and traceable
In summary, the market demand for BI is based on the information flows from operational systems to
following: regulatory reporting.
zz Extract more value from customer interactions;
zz The impact of globalisation is putting pressure on Leverage the full potential of customer data in
C-level executives to make informed decisions, and an organisation to analyse and predict customer
on the CIO to provide accurate, consistent and timely behaviour and value.
information from a dynamic organisation. zz Track performance and align metrics across the
zz Additional information must be gathered and organisation;
presented to stakeholders so as to comply with a Support the performance management initiatives in an
growing body of regulations. compliance is a key organisation by collecting, analysing and presenting
issue, as it affects the share price and ability to trade. the required metrics.
zz Organisations need to adapt to operational issues
with appropriate decisions based on up-to-date and In the next sections these business drivers will be defined
correct information, which is rarely obtainable in the in more detail, including their stakeholders and a real life
form required from current BI implementations. customer case.
zz Managers are used to having access to on-demand
information from the internet, and likewise expect
internal information to be readily available.
zz Reducing the costs of BI has become a top priority – it
is not unusual for an enterprise’s BI systems today to
be supplied by many different vendors, resulting many
standards.
zz Environmental concerns are also encouraging IT
departments to streamline their data processing
requirements.
2.3 Track risk and compliance
Compliance Stakeholders
Risk Manager, Chief Compliance Officer Wants to mitigate the risk of non-compliance. Needs reliable data. Needs an audit trail
CEO, Marketing Director Wants to maintain reputation and avoid brand damage. Wants to improve brand value.
CFO, Finance Director Wants to avoid penalties. Needs to maintain licence to operate.
Organisations have the obligation to inform stakeholders shows the regulations that have been introduced and the
such as regulatory bodies (national banks, government), increasing rate with which new regulations have been
the shareholders and the public about their performance. introduced in the recent past.
Several regulations like Basel II, Solvency II, the Sarbanes-
Oxley Act (SOX) and MIFID were introduced only in the The core requirement of Risk and Compliance initiatives
last seven years. The timeline presented in figure 2.4 is to prove that certain business processes are under
Risks of
Web September 11 IAS/IFRS
Clinical Laboratory
Improvement Usage
Amendments Data Standards
21 CFR
Part 11 Solvency II
Dot-com Integration of
Sarbanes- Business
Insurance ‘Shakeout’
Oxley Act Systems
Manager
1965 1985 1990 1998 2000 2002 2004 2006 2008 2010
European Data
Protection Act Sarbanes-Oxley
European
Financial Services Compliance
and Market Act Deadline
Acronym Key
CFR = Code of Federal Regulations
FERC = Federal Energy Regulatory Commission
Figure 2.4 Regulations introduced
IAS/IFRS (Source:
= International Gartner)Standards/International Financial Reporting Standards
Accounting
PURPA = Public Utility Regulatory Policy Act
Business value of BI
23
24
control. By requiring that organisations, and even zz The blueprint for the quantitative data required by BI to
managers personally, are liable for the information function successfully.
provided, the information must be auditable and traceable zz Assurance of the effectiveness of the risk &
from the administrative source to the final report. Business compliance framework.
Intelligence offers the capabilities to audit and trace the zz Assurance of visibility of targets, risk and compliance
information, in design- and run-time. Managing a company requirements.
can be compared with navigation by means of a compass zz Improved access, reliability and relevance of
and a map. In both situations you need to select the right management information used for decision-making.
route to the goal and need information while underway.
Applying this to BI, BI Solutions are the tools that provide
the information to understand where you are. What BI can
Organisational Model
provide and support is: Primary Measures
Major Bank needs to improve the management of both operational and credit risk across its global Major Bank achieved its objectives, including the
operations. The company requires information to create and validate group-wide risk rating models so it most important objective of being able to comply
could improve business performance and, more importantly, ensure compliance with the requirements of with Basel II in time, which had been a major
Basel II, in order to demonstrate that its systems are secure and resilient. concern for some time.
Chief Marketing Officer Needs to reduce churn and maximise market share. Needs a coherent pricing strategy, including bundles, offers,
margin maximisation etc.
Head of Customer Service Wants a single view of customer interaction. Needs to change from being a cost centre to a profit centre.
CEO, BU Director Has to increase revenue and profit. Needs visible, trusted immediate information for decision making.
With the increasing mobility of customers to choose their messages around this information - requires a carefully
suppliers, knowledge of customer needs is critical to orchestrated mix of business acumen and an analytical
retain customers, to attract new customers, or to offer framework that supports fact-based decision making.
other products to existing customers. Therefore, from a Without an analytical structure in place, even an
marketing and sales perspective it is essential to offer experienced analyst will have difficulty manually analysing
the right product at the right moment. The customer’s all of the complex information they may be collecting
behaviour is recorded within applications for CRM and on customers. And, while still a powerful resource, an
call-centre processes. Business Intelligence offers operational Customer Relationship Management (CRM)
solutions for marketing and sales organisations to analyse system alone will be unable to provide the deeper
their customer behaviour. customer understanding required to add value to every
interaction with each customer. Predictive Analytics is a
Predicting customer product preferences and purchasing general term, which describes a number of techniques
habits - and crafting the most relevant marketing used to identify pieces of information or decision-making
Business value of BI
25
26
in data. A common misconception is that it always involves allows for the gathering of all the relevant information
huge amounts of data through intelligent technologies about customers, as well as the organisation of it in a
to find patterns and give magical solutions to business consistent manner. In this way, a 360° view of customers
problems. This is not true. Instead a predictive model is achieved, which then forms the basis for wide-ranging
is generated following a highly interactive and iterative analytical methods that aid in the measurement and
process, based on CRISP-DM, as shown in figure 2.6. building of truly interactive, profitable propensity models.
Business expertise must be used jointly with advanced
technologies to identify underlying relationships and Once developed, predictive models can be used in
features in the data. conjunction with other models or business rules to
accurately predict customer behaviour. The real skill is
A seemingly useless pattern in data discovered through to find the best predictive model, and its corresponding
data mining can often be transformed into a valuable piece predictors, for a particular business problem. This is
of actionable information using business experience and difficult since there are so many options, algorithms,
expertise. formulas, rules and weights you can apply to determine
precisely how best to combine predictor variables. To
make sense of large volumes of data and how best to
model them, several Data Mining products have been
Business
Understanding
Data
Understanding
developed.
Data
Preparation More detail on the specifics of data- and text mining is
Deployment
Data
given in section 7.6 on best practices in data- and text
Modelling
mining.
Evaluation
These products use customer data to build specialised
predictive models to solve business problems, such as:
Figure 2.6 Predictive model generation (The CRISP DM Model)
%Gain
50
Percentile
monitor the effects of marketing campaigns in light of the Figure 2.7 Predicting bad debt customers
A mobile operator made a significant annual investment in 40 general Using predictive analytical models to offer the most appropriate product to
marketing campaigns. These failed either to reverse the decline in average each caller, our customer has been able to convert a customer service centre
revenue per user (ARPU), or reduce customer churn, which was running at into a highly successful inbound sales channel. As a result, in just 6 months the
48% for pre-pay and 26% for post-pay. The operator had a data warehouse operator achieved a double-digit reduction in percentage churn and generated
containing 12 months’ off call detail records (CDRs) and billing data, but was additional revenues of €1.3 million. The operator now runs more than 200
unable to analyse this or develop strategies to improve performance. highly targeted campaigns a year to stimulate usage and sell specific bundles.
Return on investment in marketing campaigns exceeds the best previous
results by over 230%. The new solution continues to identify further innovative
offerings to improve customer retention and to stimulate revenue growth.
Business value of BI
27
28
CxO Needs visibility of management information. Wants to implement closed loop processes for managing corporate performance.
CEO Needs information to measure against KPI’s. Wants to improve stakeholder information.
CIO Needs the flexibility to deliver a continually changing set of performance information to the enterprise.
Corporate Performance Management is a strategic Performance management certainly requires more than an
management approach aimed at structurally X-raying an IT-tool for the presentation of Key Performance Indicators.
organisation on all levels relative to the specified objectives CPM should be based on an unambiguous definition of
and budgets. Examples of performance management the essential motivation of an organisation. Therefore the
methodologies are the Balanced Scorecard, INK-model, organisational decision making process must be
and Activity Based Costing. In the text-box below the defined at the following levels:
more formal definition of CPM is given by the founders of
the Balanced Score Card, Kaplan and Norton. zz The mission and vision statements lead to business
goals and a strategy. Critical Success Factors (CSF)
define the prerequisites to reach the goals. The
Definition of CPM… strategy states how the goals should be reached.
‘….provides managers with a zz The business goals and imposed strategy lead
comprehensive framework that translates to objectives and a policy (business plan). Key
a company’s vision and strategy into a Performance Indicators (KPI) define how the objectives
coherent set of performance measures.’ will be measured. The imposed policy will be stated
(Source: Kaplan and Norton) with business rules.
Business value of BI
29
30
Identifying the relevant Critical Success Factors (CSF) All steps in the pyramid, as shown in figure 2.9, need
and defining accurate Key Performance Indicators (KPI’s) aligned execution to ensure the effectiveness of the
is a real challenge. When completed, key processes and performance management initiative.
activities with the highest business value, i.e. highest
impact on the KPI, should be identified and prioritised. It is Effective measurement, monitoring and analysis of key
also important to have a satisfactory mix between leading performance indicators is critical to support performance
and lagging indicators that reflects both the historical management processes in an organisation. The CPM
results and the upcoming trends and forecasts in different approach results in a comprehensive pyramid from
activities and processes. strategy to operational goals and gives a clear picture
of the alignment between strategy and the accumulated
For each key process, activity or initiative the decision business information where Business Intelligence can
points have to be analysed to identify the measures support the continuous measuring, monitoring and
relevant to the decision. This analysis results in a clear analysis of the performance indicators at all levels.
set of required business information. Finally, the relevant
sources of information requirements are analysed with
regards to availability, completeness and quality.
Vision
Strategy
Critical Sources
Factors
Key Performance Indicators
Finance Customer Internal Learning
performance knowledge processes & growth
Scorecards/Dashboards
Analysis & Reporting
Data Warehouse
A major passenger rail operator was constantly struggling with weak financial The project has supported this major passenger rail operator drive for unified
performance. Management had a need for consistent information in general, corporate performance management and has rationalised the management
and to define a set of information that could be used for steering the company’s decision process. The company is prosperous and the days of weak financial
strategy. The company was also concerned that the information currently performance are over. The organisation now has effective decision-support
available was not supporting the developing organisation. Management data information at all levels. Staff use the KPI’s to control and comment on their
was on different systems and the process of capturing the information needed area’s performance on a monthly basis, and do so directly in the application,
streamlining. rather than on separate reports. Management uses the application in its
meetings. and documents decisions in the application, avoiding parallel
reporting mechanisms and reducing the management overhead.
Business value of BI
31
ht place, r ight fo
e, rig rm
t tim iness goals Manage tactical initiatives at, r
ig h e r m b u s h o u s e Dat a m i g
M ig
, r p long t r a o n i ht
o n D eve l o u a l i t y D a t a
wa re
t i o n tor
p e il y b e nt B
vie i
w t Data q
s r c e t i o n p ro c e s s L i fe c y c l e r i c a l m at
Co
da manag e
nt
rso usi
n e
n
histo infor
nM
m
ge
na
on
ma
3
m
i g ht
i to r
ta
da
n
per
ess
fo r m
C re a t i n g R
u
ance
a
s
o
I d e nt i f y o p
r
p
i
e
n
e
s t
r
e
a
a
t
M
s
i o
d data
s
n
n g
s
l li
ct u re
e Data m o de
r
I n fo r
e n consis t u n
str u
mat
e n c re a
p
ng
io
Usi
o
po
s
r
a ue
t
t e
i n nf
nc
l
al
u
r
g
y
I
or Bu g
V
llin
rn
s
de
iit esmationsiness analy tics Business moove usin
Pre traceability Dat a g late b
d i c u
t per for mance Sim
34
Business focus Develop long-term business goals Manage tactical initiatives to Manage and monitor daily
achieve business goals business operations
Primary users Executives and Business analysts Business analysts and LOB managers Front-line employees and
and LOB managers operational managers
Data Historical and projected data Historical data Real-time, low-latency and
historical data
3.1 BI Foundation
There are two critical success factors in almost any BI
Executive
initiative. The first is consistency of information and the
Figure 3.1 defines a generic reporting structure as seen in A BI solution is usually deployed to support such reporting
many companies. Data is administered in various back- and analysis needs of an organisation. A BI solution brings
office systems, often proprietary software, supporting together data from different source systems. Therefore BI
the operational processes. Service centres run the daily solutions often reveal inconsistencies in the back-office
operations and use the back-office systems to run the data stores. This is because data passes multiple levels of
business. Performance of the service centres and the enrichment and aggregation. Enriching and aggregating
business processes they run is reported on various levels data to support the different control levels within an
and with various orientations. Reporting is based on organisation should not lead to a lack of traceability of
legal entities, commercial labels, internal organisational information. The degree in which the developed BI solution
structure and location. On the highest level of the resolves these issues is key to the success of the overall
organisation the measures must add up and still provide BI initiative.
a consistent and traceable view of the performance of the
company as a whole. Key in achieving accountable, consistent and traceable
information in a Business Intelligence solution is the
presence of data management processes. Business data
definitions, data quality measurements, data ownership
and data governance processes should already be part
of any organisation. In many cases the first iterations of a To support a successful BI initiative we primarily need a
Business Intelligence initiative still fail due to lack of these solid BI foundation. Data Management, including data
capabilities. modelling, data quality and data governance capabilities,
is required to understand the definition and value of
The BI foundation model as shown in figure 3.2, defines data. Data Integration, including Data Warehousing,
the prerequisites for a successful BI initiative and relates Data Migration and Content Management, is required
the various subjects in the overall flow from enterprise data to process data into actionable information. Business
and content to valuable information. Intelligence, including Business Reporting, Business
Business
Reporting
Data Data Ware Business
Modeling Housing Reporting
Corporate
Enterprise Governance
data & Data Data Business
Content Quality Migration Analytics
Customer
Intimacy
Data Content Business
Governance Management Modeling
Product
Leadership
Analytics and Business Modelling, is then applied to towards creation of the BI foundation is critical. We do
measure and improve business performance. not need a fully detailed corporate data model to start
building information products for our business users.
The first increments of a BI initiative establishes the BI Within the scope of the BI initiative and the information
foundation, which forms the bases for producing valuable products needed, we however have to pay sufficient
BI products for our business users. To ensure timely attention to data management and integration aspects to
delivery of business value an incremental approach be successful in the long run.
2.
Articulate
strategy
1.
Analyse 3.
Objectives
Strategy Prioritise
& CSF
4.
5. Balance
Define 9.
KPI’s
Functional Discover
scope & Explore
10.
Access,
6. Monitor &
Identify Effect analyse
Data & change
sources
Construct Consume
11.
8. develop
7. 12. Decision
Develop
Evaluate Share & alternatives
Implement
& select collaborate
Train
tools
zz Outside-in
In this section an outside-in maturity model is introduced
Measuring the perception of Business Intelligence by the first, based on the BI maturity model of The Data
organisation. The focus in this perspective is how the Warehouse Institute (TDWI). Next an inside-out maturity
various participants in BI experience their BI solution. model is introduced.
value
Cost
Enterprise
Management Spreadmart data Marts Data data Analytical
Architecture Reporting Warehouses warehouseing Services
PREDICTIVE
MODELS
CAUSE
VALUE
EFFECT Action
Understand why
Plan the changes ‘What if...’
Understand what happened SCORECARD
CAPABILITY / MATURITY
DATA: Data Extract Transform Load • Data Ouality Improvement • Data Warehouse/Mart
Figure 3.5 REPORT:
BI MaturityAutomated
by Logica Standard Reporting • Automated Alerting
li v tio c e
o n e t h c e T i m e t o m a r k e t B e n c h m a r t i n g t of
dm
et y C
n
en
er
o p e n t re
s
co
po
a
4
e
tal
c
r tu n
M
g l
Simplified To
o
r k in
p
V i s i o n a n d s t r a te g y B
o l o g y
ation
d
n a li s
o
vi
is a M a l i t y of s e r
I
ratio
p nte
l e
ess
o
u
c
n g ip Q
t ol
P r o
m
h
e
t
io r ic s s
er i
t i
wn nd xp E
ra m of o f u a l rt e e
mati sti
t i o an Total cost ntrol tion on
n a a g e m e nt C o re r a
l i s a t i o n I nf r a s t r u ct u
48
4 Managing BI
Due to the specific nature of BI initiatives a closer look to
the project management and organisational aspects is
required. In most organisations various BI applications
and initiatives already exist. The challenge is bringing
those initiatives together in an integrated BI programme,
identifying additional business value or ways to reduce
the Total Cost of Ownership of Business Intelligence.
According to recent research in the marketplace the cost
of BI is going up whereas the benefits of BI are not always
that obvious.
Managing BI
49
50
O
Competence
pt
Center
im
as
zz infrastructure rationalisation;
a
tio
Data Quality Off
n
Improvement Shore
BI Application
Maintenance Service
zz processes and organisation optimisation.
Infr
ast
BI Service
ruc
Simplified Center
For all categories the cost- and other benefits are given,
ture
reporting
Ratio
Optimisation of BI
Project portfolio
Virtualisation BI as a together with the expected timelines to reach them.
nalisa
Service
BI
Framework
t io n
Assessment BI Platform
of rationalisation
BI TCO
3 months 6 months 12 months 24 months
Simplified reporting Review the pipeline of existing reports and delete Run cost Report consistency 3 to 6 months
the ones that either are no longer used or can be Maintenance cost
replaced with other existing reports
Common Language Enable a shared definition of KPI’s across the Design and build cost Consistency of KPI’s 6 to 12 months
organisation. It involves reviewing all existing Maintenance cost Cross-functional
KPI’s and dimensions, then finding a company- Administration cost data analysis
wide common ground. Reduce delivery time
Data Quality Improvement Asses main characteristics of data sources and Build cost Information quality 6 to 12 months
cleansing opportunities to meet BI expectations, Maintenance cost
then clean data.
Master Data Management Processes and tools to manage master data Design and build cost Information consistency 12 to 24 months
across the organisation. Maintenance cost
BI platform rationalisation Reduce the number of BI software providers License cost Improved traceability 3 to 12 months
and invest on the BI platform that covers the Maintenance fees and audit-ability
widest range of business requirements Training cost
Run cost
Virtualisation Run multiple software stacks on the Hardware cost Improved maintenance 6 to 12 months
same hardware
BI as a service Plan and organise the rent of BI platform and Software cost Replace fixed- with 12 to 24 months
services (hardware and software) from a BI Hardware cost variable cost
service provider. In this case the organisation no Build cost
longer owns its BI platform, but pays BI services Run cost
as it uses those services
BI Framework Structured methodology and approach Design and build cost Increase BI maturity 3 to 6 months
to BI solution engineering
Optimisation of BI Portfolio Review of scope of pending BI projects in Design and build cost Improve 3 to 6 months
order to define the best path to meet business Run cost time-to-market
requirements while reducing development and
delivery cost.
BI Application BI maintenance based on clear service level Maintenance cost Predictable service levels 6 to 12 months
Management Service agreements, leveraging blended deliver models
(near- and offshore)
BI Service Centre Consolidate BI applications development External cost Quality of service 6 to 12 months
providers as a mean to reduce external cost Improve the management
and improve quality of service. of providers
BI Competence Center Centralised BI organisation that defines Design and build cost Quality of service 12 to 24 months
and implements best practices and manages Maintenance cost Operational excellence
company-wide BI activities Environment cost
Managing BI
51
52
• Analytic skills
To design the BI solution based on the information
needs. Design and develop analytical applications to The core activities performed in the BI CC are:
support the business. Perform analytical processing of
the availble information in the BI solution. • Define BI vision and strategy
• IT Skills The definition of the BI strategy based on the business
To develop and maintain the BI system, aligned strategy of the organisation. Continuously align the
with the IT architecture and infrastructure of the BI vision with the tactical goals and objectives on all
organisation. levels of the organisation.
• Manage programs
Working together as one BI community, either virtually or Keep track of all BI projects running in the
physically implemented in the organisation, is essential to organisation. Make sure all projects are aligned
achieve sustainable added value out of BI investments. to the defined BI strategy and really add value to
Figure 4.2 presents the required skills combined with the the business. Deliver within budget, quality and
core activities of a BI CC. time constraints the agreed functionality. Manage
data resources, information delivery and running
environments
• Develop user skills The combination of different skills and capabilities of a BI
Create awareness of the value of BI in the business. CC often introduces a discussion about how to position
Make sure business users use BI solutions as the BI CC in the organisations hierarchy. In many cases
intended, by conducting evaluation and training the BI CC is not positioned in the organisations hierarchy,
sessions on a regular basis. Measure business user but positioned as a long term programme or virtual BI CC.
satisfaction with delivered BI solutions. Table 4.4 presents the advantages and disadvantages of
• Organise methodology leadership positioning a BI CC within the hierarchy of the organisation
Manage the BI methodologies used to develop BI or as a virtual BI CC.
solutions. Align existing best practices in a consistent
framework. Continuously evaluate projects on using
the framework to improve future projects and/or
improve the existing framework.
• Build technology blueprint
Define and establish the required technology blueprint,
based on the BI strategy and the collective set of
requirements of running BI projects. Envision a to-be
architecture and manage projects to move towards
that architecture.
• Establish standards
Standardisation is a key enabler for re-use.
Re-use of existing skills, practices and components.
Harvest developed standards in running BI projects
and leverage future projects with them. The level of
standardisation is also one of the key success factors
in outsourcing of BI development and maintenance.
• Control funding
Position the BI CC as a decision making unit in the
funding process of BI. Of course business will decide
on funding in the end, but BI CC must have a clear say
in what to invest in, and what not. This is essential to
prevent BI initiatives running in the organisation without
alignment with the BI CC.
Managing BI
53
54
BI CC in the Roles and responsibilities are clearly defined Political & budgetary constraints
organisation hierarchy Clear reporting lines, BI CC manager
BI CC high visibility in the organization
When a BI CC is positioned in the organisations hierarchy most common options are presented in the table below,
we of course still have to decide where to position it. The including their respective advantages and disadvantages.
CEO BI CC operates for the whole organisation Loosely aligned with business user organisation units
Closely aligned with overall business strategy Not on top-of-mind of CEO
CFO Aligned with most demanding user organisation Non financial BI requirements less supported by the
Usually clear and easily funded business cases to support BI CC due to prioritisation issues
Loosely aligned with other organisation units
CIO No bias for specific business organisation units Lack of alignment with business organisation units
Alignment with overall IT strategy, processes and project BI projects lack business orientation
portfolio
Managing BI
55
56
MMI
C
The Capability Maturity Model ® Integration
(CMMI ®) provides a comprehensive
approach for process improvement. In
addition to the Maturity Levels the CMMI
also specifies what the achievable goals
are and where and when what priorities
should be placed during the different
stages of maturity. Finally, the CMMI model
offers an extensive support for the design
and realisation of a successful process
improvement programme.
The risk of this is deemed too high and unacceptable. In SAS 70 (Statement on Auditing Standards
practice, however, an increasing amount of BI solutions No 70) is an internationally recognised
is developed outside the organisation. Here test data are research standard and was developed
used that often closely resemble data from the customer’s by the American Institute of Certified
operational environment. There are two measures that can Public Accountants (AICPA). It is process
bring this risk to an acceptable level: focused and examines whether enough
management measures are taken by the
zz By establishing clear rules of conduct and linking service organisation and also whether they
contractual agreements to these rules, the risks can are effective and respected. As is the case
be reduced to a level comparable to or even safer with the annual financial statement, an
than internal development and maintenance. With the independent auditor or auditors can supply
current international regulations, such as SAS-70, in annual reports on this matter. To be able to
the field of process control of service organisations, service clients listed on a stock exchange in
it is now also possible to make such contractual the United States it is even a requirement.
agreements at an international level. An SAS 70 report turns out to be the only
reporting standard that meets the strict
zz In the past, the testing of data warehouse solutions Sarbanes Oxley regulations.
was often performed using sample production data.
When the results sounded appealing to the users it
was quickly concluded that this was good enough.
With the increasing use of data warehouse solutions
for financial reporting and risk analysis this approach
to testing is not sufficient any longer. It is increasingly
common to see controlled test sets and formal test
projects used. Tools and methods also offer remedies
to mask production data in an efficient manner so that
confidentiality is guaranteed and an effective test set
Managing BI
57
58
4.3.4 Technology
With modern technology, most of the basic code is with a background in programming and data modelling.
generated. The products available on the market today Programming is more about structuring than coding.
are used by software engineers and knowledge workers
and it is said that the products require no programming The added value offered by recent ETL products are
knowledge and experience. This could lead to the standardised reusable components and the possibility
argument that blended delivery is not even necessary to create and maintain reusable components yourself.
because, in the future IT expertise is no longer required for Examples are standard transformations for the various
the use of these products. types of dimensions in a data warehouse, or standard
connectors for different types of interfaces. It is exactly
This is, however, the way software manufacturers position this feature that is useful when contemplating offshore
their products in the market. And it is certainly true that development of ETL components. Each standard building
the accessibility and user-friendliness of such products are block ensures clearer specifications in subsequent
considerably better than in the past. Practice has shown projects.
that operating such products is still reserved to IT experts
Near Shore
In every stage of a BI initiative an analysis is required to
weigh complexity and required customer interaction. Low
Based on this result the most appropriate delivery Low Customer interaction/ High
intimacy
approach can be applied.
Figure 4.3 Blended delivery model
4.4 Estimating investment in BI
In general there are two methods of estimating the down approach to estimation is most appropriate. Details
necessary investment in BI. The first is top-down to perform an expert estimate are not available in these
estimating, utilizing industry benchmarks. The second early stages of a BI initiative. When one arrives in the BI
one is bottom-up estimating, utilizing expert estimation, development stages an expert estimate, supported by
supported with collected metrics of previous projects. metrics, should be applied to verify and detail the earlier
Both methods should be applied in the estimation of your made top-down estimates. Both methods of estimation
BI initiative. During the first stages of a BI initiative a top- are explained in detail in the following sections.
Managing BI
59
60
Managing BI
61
62
References
Abonyi, J., Feil, B., and Abraham, A. ‘Computational Intelligence in Data Mining’, Informatica, 29, 3-12.
Cios, K. J., Pedrycz, W., Swiniarski, ‘Data Mining, A Knowledge Discovery Approach’, Springer, New York.
R. W., and Kurgan, L. A.
Dan Linstedt ‘The Business of Data Vault Modelling’, This book presents the business of next generation data warehousing,
including the Data Vault model and approach or methodology. (www.danlinstedt.com)
Dy, J. G., and Brodley, C. E. ‘Feature Selection for Unsupervised Learning.’, Journal of Machine Learning Research, 5, 845-889
Fayyad, U., Piatetsky-Shapiro, G., ‘From Data Mining to Knowledge Discovery in Databases.’, American Association for Artificial Intelligence.
and Smyth, P.
Gartner Gartner, Inc. (NYSE: IT) is the world’s leading information technology research and advisory company. (www.gartner.com)
Inmon Bill Inmon, world-renowned expert, speaker and author on data warehousing, is widely recognized as the ‘father of data
warehousing.’ He is creator of the Corporate Information Factory and more recently, creator of the Government Information
Factory. (www.cif.com)
Imhoff Claudia Imhoff, Ph.D., is the President and Founder of Intelligent Solutions, a leading consultancy on data warehousing
and business intelligence technologies and strategies. (www.intelsols.com)
Kaplan and Norton Founders of the balanced score card approach (www.valuebasedmanagement.net)
Kimball Worldwide known innovator, writer, educator, speaker and consultant in the field of data warehousing. He has remained
steadfast in his long-term conviction that data warehouses must be designed to be understandable and fast. His books
on dimensional design techniques have become the all time best sellers in data warehousing. (www.kimballgroup.com)
Mimno Mr. Mimno has accumulated extensive practical experience in identifying mistakes that are commonly made in the
development of data warehousing applications, and assists MMH clients in building successful data warehouses
incrementally.(www.mimno.com)
Miradi, M. ‘Knowledge discovery and pavement performance, data mining’, Wohrmann Print Service, The Netherlands.
Moss Ms. Moss has written a number of books, white papers, and articles on business intelligence, project management,
information asset management, development methodologies, data quality, and organizational realignments. In 1991 she
self-published her first methodology RSDM 2000, Relational System Development Methodology, Volumes I & II. Since
then, she has co-authored the books Data Warehouse Project Management, Impossible Data Warehouse Situations, and
Business Intelligence Roadmap, and Data Strategy
NESMA NESMA (Netherlands Software Metrics Association), the second largest Functional Sizing Measurement Organisation in
the world.(www.nesma.nl)
OVUM At Ovum we fully understand convergence across telecoms, IT services and software. We invest heavily in researching
what is happening in a market that is dynamic and full of risk and reward. We analyse the changes and identify the threats
and opportunities ahead for our clients. (www.ovum.com)
TDWI TDWI (The Data Warehousing Institute™) provides education, training, certification, news, and research for executives
and information technology (IT) professionals worldwide. Founded in 1995, TDWI is the premier educational institute for
business intelligence and data warehousing. (www.tdwi.org)
Treacy and Wiersema Founders of the value discipline model (www.valuebasedmanagement.net)
Zachman John A. Zachman is the originator of the ‘Framework for Enterprise Architecture’ which has received broad acceptance
around the world as an integrative framework, or ‘periodic table’ of descriptive representations for Enterprises.
(www.zachmaninternational.com)
185
186
Index of Terms
accountable ........................................................................................................................................................................................................ 22, 36, 72
appliances ..................................................................................................................................................................................................................... 83
BI strategy ........................................................................ 52, 53, 66, 67, 68, 69, 70, 71, 72, 73, 85, 86, 94, 98, 100, 102, 103, 109, 110, 122, 171, 172
Data Integration ..................................................................................................................................................... 37, 39, 85, 87, 118, 121, 130, 131, 159, 186
Data mining ................................................................................................................. 26, 44, 90, 153, 154, 155, 156, 158, 159, 160, 161, 163, 164, 165
data modelling ............................................................................................................................................................ 37, 55, 58, 88, 116, 134, 135, 137, 146
data models ................................................................................................................................................................... 74, 83, 88, 90, 105, 129, 134, 147
Data Vault ................................................................................................................ 88, 105, 111, 116, 124, 134, 135, 136, 138, 140, 141, 142, 146, 176
DBMS .................................................................................................................................................................................................................... 88
definition of BI .................................................................................................................................................................................................................... 34
dimension ..................................................................................................................................... 60, 74, 76, 77, 124, 147, 149, 150, 151, 152, 153, 182
Dimensional modelling .............................................................................................................................................................. 88, 105, 116, 146, 147, 153, 176, 182
ETL ............................................................................................................................................................ 56, 58, 83, 87, 116, 123, 124, 125, 126
187
188
facts ...................................................................................................................................................... 44, 60, 74, 76, 81, 146, 147, 149, 150, 182
FPA .................................................................................................................................................................................................................... 61
funding .................................................................................................................................................................................................................... 53
granularity ............................................................................................................................................................ 76, 77, 88, 90, 140, 147, 148, 159, 182
increments ..................................................................................................................................................... 38, 56, 67, 73, 76, 79, 83, 85, 86, 176, 183
investment ................................................................................................................................................................ 11, 20, 21 42, 43, 57, 59, 60, 61, 85
maturity .............................................................................................................................................................. 10, 41, 42, 43, 44, 45, 56, 66, 68, 85
measures ................................................................................................................................................ 28, 30, 36, 38, 57, 77, 146, 147, 148, 149, 182
MIFID .................................................................................................................................................................................................................... 23
OLAP ................................................................................................................................................. 44, 88, 90, 153, 154, 158, 159, 160, 163, 177
SAS-70 .................................................................................................................................................................................................................... 57
Solvency II .................................................................................................................................................................................................................... 23
SOX .................................................................................................................................................................................................................... 23
stakeholders ............................................................................................................................................................... 18, 22, 23, 69, 78, 85, 101, 104, 105
Testframe .................................................................................................................................................................................................................... 93
text mining ................................................................................................................................. 26, 38, 90, 116, 153, 154, 160, 161, 162, 163, 164, 165
V-model .................................................................................................................................................................................................................... 93
189
190
Index of Figures
Figure 2.1 Value disciplines ..................................................................................................................................................................................................................... 19
191
192
Index of Tables
Table 1.1 Target audience ..................................................................................................................................................................................................................... 14
OVUM .................................................................................................................................................................................................................... 83
193
Logica
250 Brook Drive
Green Park
Reading, RG2 6UA
United Kingdom
T: +44 20 7637 9111
F: +44 20 7468 7006
E: hga.bi.ww@logica.com
I: www.logica.com/BI
About Logica
Logica is a leading IT and business services company, employing 40,000 people. It provides business consulting, systems
integration, and IT and business process outsourcing services. Logica works closely with its customers to release their
potential - enabling change that increases their efficiency, accelerates growth and manages risk. It applies its deep
industry knowledge, technical excellence and global delivery expertise to help its customers build leadership positions in
their markets. Logica is listed on both the London Stock Exchange and Euronext (Amsterdam) (LSE: LOG; Euronext: LOG).
More information is available at www.logica.com.
Havey
oumi
ssedoneofoureBooks
?
Feel
freet
ocont
actusathga.
bi.
ww@l
ogi
ca.
com