Description
The implementation of a business intelligence (BI) system is a complex undertaking requiring considerable resources.
Spring 2010 Journal of Computer Information Systems 23
CRITICAL SUCCESS FACTORS FOR
BUSINESS INTELLIGENCE SYSTEMS
WILLIAM YEOH ANDY KORONIOS
University of South Australia University of South Australia
SA 5095 Australia SA 5095 Australia
Received: June 11, 2009 Revised: August 6, 2009 Accepted: September 14, 2009
The implementation of a business intelligence (BI) system is a
complex undertaking requiring considerable resources. Yet there
is a limited authoritative set of critical success factors (CSFs) for
management reference because the BI market has been driven
mainly by the IT industry and vendors. This research seeks to
bridge the gap that exists between academia and practitioners by
investigating the CSFs in?uencing BI systems success. The study
followed a two-stage qualitative approach. Firstly, the authors
utilised the Delphi method to conduct three rounds of studies.
The study develops a CSFs framework crucial for BI systems
implementation. Next, the framework and the associated CSFs
are delineated through a series of case studies. The empirical
?ndings substantiate the construct and applicability of the
framework. More signi?cantly, the research further reveals that
those organisations which address the CSFs from a business
orientation approach will be more likely to achieve better results.
Keywords: Business intelligence system, Critical success
factors, Delphi method, Case study
INTRODUCTION
Recently Business Intelligence (BI) applications have been
dominating the technology priority list of many CIOs [11, 12].
According to Reinschmidt and Francoise [22], a BI system is “an
integrated set of tools, technologies and programmed products that
are used to collect, integrate, analyse and make data available”.
Stated simply, the main tasks of a BI system include “intelligent
exploration, integration, aggregation and a multidimensional
analysis of data originating from various information resources”
[21]. Implicit in this de?nition, data is treated as a highly valuable
corporate resource, and transformed from quantity to quality
[27]. As a result, massive data from many different sources of a
large enterprise can be integrated into a coherent body to provide
‘360 degrees’ view of its business [5, 27]. Hence, meaningful
information can be delivered at the right time, at the right location,
and in the right form [5, 20] to assist individuals, departments,
divisions or even larger units to facilitate improved decision-
making [15].
While the BI market appears vibrant and the importance of BI
systems is more widely accepted, few studies have investigated
the critical success factors that affect the implementation
success. Although there exist a plethora of guidelines from the IT
industry, most rely on anecdotal reports [15]. This is because the
study of BI systems is a relatively new area that has been driven
primarily by the IT industry and by vendors [15]. Therefore,
empirical research to shed more light on those critical success
factors (CSFs) in?uencing the implementation of BI systems is
desirable. An understanding of the CSFs enables BI stakeholders
to optimise their scarce resources and efforts by focussing on
those signi?cant factors that are most likely to aid successful
system implementation.
RESEARCH MOTIVATION
The implementation of a BI system is not a conventional
application-based IT project (such as an operational or
transactional system), which has been the focus of many CSF
studies [10]. Instead, it shares similar characteristics with other
infrastructural projects such as enterprise resource planning (ERP)
systems implementation. That is, implementing a BI system is not
a simple activity entailing merely the purchase of a combination
of software and hardware; rather, it is a complex undertaking
requiring appropriate infrastructure and resources over a lengthy
period [10, 19, 28].
Speci?cally, the key infrastructural foundation for most
enterprise level BI systems — a data warehouse — is a subject-
oriented, integrated, time-variant and non-volatile collection of
data that differ from conventional online transactional processing
(OLTP) databases [14]. A complex data structure must be maintained
in order to provide an integrated view of the organisation’s data
so users can query across departmental boundaries for dynamic
retrieval of relevant decision-support information. Furthermore,
the BI system’s architecture is highly complex owing to the back-
end systems originating from multiple data sources and to the vast
volume of data to be processed. In addition, the implementation
of a BI system is often associated with the following challenges:
underlying original back-end systems and processes which were
not adapted for BI applications; poor data quality derived from
source systems that can often go unnoticed until cross-systems
analysis is conducted; and the maintenance process that tends to
be vague and ill-de?ned [10, 19, 23].
Despite the complexities in implementing BI systems, there
has been little empirical research about the CSFs impacting the
implementation of BI systems. The gap in the literature is re?ected
in the low level of contributions to international conferences and
journals. More importantly, the value of previous CSF studies
will obviously decline with age [16]. The rapid advancement of
technological innovation and the pace at which new technologies
are being adopted will apparently in?uence the state of criticality
for research into CSFs [16]. Furthermore, CSFs applicable to
other types of information systems may not necessarily apply
to a contemporary BI system. Therefore, the increased rate of
adoption of BI systems, the complexities of implementing a
contemporary BI system, the scarcity of academic research, and
24 Journal of Computer Information Systems Spring 2010
the far-reaching business implications justify a more focused
examination of CSFs as well as the associated contextual issues
required for implementing BI systems. This paper describes an
exploration of an important and topical area of interest — the
CSFs impacting the implementation of BI systems. It is expected
that this research will make a contribution to both theory and
practice. In theoretical terms, this research:
• adds to knowledge and contributes to the literature of
an emerging area of interest — the implementation
of BI systems, in particular, the CSFs that affect the
implementation effort;
• identi?es the criteria which determine the success of BI
systems implementation; and
• validates current CSFs understandings, and extends our
knowledge of contemporary BI systems.
In practical terms, the research project:
• identi?es the CSFs that impact on BI systems imple-
mentation, so enabling stakeholders to better use their
scarce resources by focussing on those key areas that
are most likely to have a greater impact.
The remainder of this article has been structured as follows.
The next section describes the two-stage research methodology
in this study. The later section presents and discusses the research
?ndings. In the last section the authors conclude the overall study
and then state their contributions.
RESEARCH DESIGN
This research adopted an interpretivist paradigm and followed
a two-stage qualitative approach. Based on extensive literature
review, a research framework and associated interview questions
were developed for use in the Stage 1 exploratory Delphi study.
This research framework was re?ned with the data gathered from
15 Delphi participants. In the second stage, the preliminary CSFs
framework derived from the Delphi study was further veri?ed
in ?ve case studies of large and complex organisations. Each
case study involved interviews with the BI stakeholders, and the
collection of project documents. Cross-case analysis was done
to examine the CSFs. As a result, a re?ned CSF framework was
developed.
Stage One — Delphi Study
In the absence of much useful literature on BI system, this
stage one of research seeks to explore and identify a set of CSFs
that are jointly agreed by a group of BI system experts. The Delphi
method was deemed to be the most appropriate method for this
study because it allows the gathering of subjective judgements
which are moderated through group consensus [17]. Moreover, this
research assumes that expert opinion can be of signi?cant value in
situations where knowledge or theory is incomplete [17]. For this
study, a Delphi panel composed of ?fteen BI systems experts was
established. Ziglio and Adler [34] assert that useful results can be
obtained from small group of 10-15 experts. Beyond this number,
further increases in understandings are small and not worth the
cost or the time spent in additional interviewing. Thus, the size of
such a Delphi panel is deemed suitably representative.
The Delphi study comprised three rounds. During the ?rst
round the authors conducted face-to-face interviews with each
participant. After the interview, further clari?cations (if any)
were made by follow-up phone calls and email communications.
Subsequently, the data gathered from the ?rst round of interviews
were analysed thoroughly by content analysis technique, a
constant comparison technique, to identify major themes [18].
In other words, the qualitative data were examined thematically
and emergent themes were ranked by their frequency and later
categorised.
In the second round, the suggested factors of all the
participants were consolidated into a single list. The list was then
distributed among the participants to facilitate comparison of the
expert’s perceptual differences. During the third round, the list of
candidate CSFs was surveyed by the Delphi participants using a
structured questionnaire survey approach. Speci?cally, a 5-point
Likert scale (i.e. from 1 ‘not important’ to 5 ‘critically important’)
was applied to rate the importance of the candidate CSFs in the
process of seeking consensus from the BI experts. From the
survey feedback, only those factors with standard deviations
(SD) of 1.0 or less, and average ratings of 3.5 and above, were
short-listed as CSFs because 1 SD from the mean contains 68%
of all scores. These criteria (i.e. SD3.5/5.0) offer
a working de?nition of a threshold for stability, and hence the
resultant CSFs are considered legitimate. Therefore, the existence
of CSFs within this de?nition in round three is considered to be a
critical point for terminating the Delphi study. The details of the
results are discussed below.
Stage Two — Case Study
Due to limited academic literature, the Stage 1 Delphi study
was used to narrow the CSFs focus of this research. However,
reliance on the Delphi study alone was not suf?cient for the
collection of data needed to rigorously address the research
objective. Therefore, a case study methodology was used for
Stage 2 of the theory-building process. That is, this second stage
sought to corroborate the CSFs ?ndings of Stage 1. The case study
methodology provides better explanations and understandings
on the examined phenomenon which would otherwise be lost
in other quantitative designs [18, 33]. For this study, in contrast
to sampling logic, a case study is an empirical investigation
following replication logic that leads to analytic generalisation
[33]. Thus multiple case studies in this research should be
regarded as multiple experiments and not multiple respondents
in a survey [33]. That is, relevance rather than representativeness
is prioritised in case selection. Given that the objective of this
study was to build theory, a case study process with multiple-case
design was the appropriate approach, and the use of the case study
methodology is justi?ed on these grounds.
Data collection for this study entailed semi-structured
interviews with key stakeholders of BI projects. To facilitate
data triangulation, data were also gathered from a number of
sources including relevant documents. A cross case analysis
approach was used in this study to gain better understandings
and increase the generalisability of the ?ndings [18]. In searching
for patterns, the authors examined similarities and differences
about relationships within the data. Hence, varying the order in
which case data are arrayed enables patterns to become more
obvious [24]. This research did not produce quantitative data. In
all cases, the authors were examining the presence or absence
of a particular CSF (e.g., were adequate resources provided?),
while at the same time ascertaining whether that characteristic
Spring 2010 Journal of Computer Information Systems 25
was ful?lled in a meaningful way. To assess the importance of
the seven previously-identi?ed CSFs, the authors studied ?ve
organisations that had implemented BI systems, including rail
corporations, energy utilities, water utilities, and a ship-builder.
STAGE ONE CSF
FINDINGS AND DISCUSSIONS
This section presents the ?ndings and interpretation of the stage
one Delphi study. As illustrated in Figure 1, this CSFs framework
outlines how a set of critical factors contributes to the success
of a BI system implementation. Following Ariyachandra and
Watson [2], the implementation success criteria of this research
take into account two key dimensions: process performance (i.e.
how well the process of a BI system implementation went), and
infrastructure performance (i.e. the quality of the system and the
standard of output). The infrastructure performance has parallels
with the three major IS success variables described by Delone and
McLean [7, 8], namely system quality, information quality, and
system use, whereas process performance can be assessed in terms
of time-schedule and budgetary considerations [2]. Speci?cally,
system quality is concerned with the performance characteristics
of the information processing system itself, in which the system
should be ?exible, scalable and able to integrate data [2, 7, 8].
Information quality refers to accuracy, completeness, timeliness,
relevance, consistency, and usefulness of information generated
by the system [2, 7, 8]. System use is de?ned as “recipient
consumption of the output of an information system” [7, 8].
Subsequently, individual users and their respective organisations
would assess the bene?ts of the BI system implementation [13].
This perception of the bene?ts would then become part of an
interactive, business-driven evolutionary continuum to further
support evolving business needs for improved BI systems [3,
4]. That is, a BI system implementation is viewed as an organic
cycle that evolves over time. Based on constant evaluation of
the information, as well as user feedback, the system resembles
a loop that requires re-evaluation of existing BI solutions,
and subsequently the system will be modi?ed, optimised and
improved accordingly. In other words, completion of the system
implementation does not mean that all BI related problems are
resolved [21]. The system will succeed only when business users
keep identifying and modelling knowledge, as well as monitoring
and modifying data repositories on an ongoing basis [21]. Hence,
the entire process is cyclical, but with a series of interrelated steps
[25].
In brief, this framework treats the CSFs as necessary for
implementation success of a BI system, whereas the absence
of the CSFs would lead to failure of the system. The Delphi
participants provided detail and justi?cation to those critical
factors. The CSF framework details the CSFs identi?ed in the
?rst stage of this study, and they are presented according to the
major dimensions of interest proposed by Wixom and Watson
[32], namely organisation, process, and technology. For each CSF
description, the primary data came from the ?ndings of the third-
round consensus amongst the Delphi participants. The secondary
data came from the ?rst two rounds of qualitative interviews with
individual participants. Additional data from published literature
were also used to support the arguments of the participants.
FIGURE 1 — CSFs Framework for Implementation of BI Systems
26 Journal of Computer Information Systems Spring 2010
Organisational dimension
Committed management support and sponsorship.
Committed management support and sponsorship has been
widely acknowledged as the most important factor for BI system
implementation. All Delphi participants agreed that consistent
support and sponsorship from business executives make it easier
to secure the necessary operating resources such as funding,
human skills, and other requirements. One interviewee stated
?rmly, “If you don’t have top level sponsorship — it is doomed!”
Another participant explained the situation this way,
“Project Sponsorship has been shown to be the single most
important determinant of IT project success or failure. A
BI project is no different to any other IT project in this
respect . . . Maintaining the commitment and support of
the projects sponsor throughout the project — because
circumstances can change over the life of the project.”
Many participants also asserted that it is more bene?cial if
the sponsor is from the business side of the enterprise rather
than from the IT sector. Similarly, a study by Watson et al. [30]
indicates that the ideal BI sponsor should come from a business
function. Such a sponsor often has a strong stake in the success of
the BI initiative. Most importantly, some interviewees highlighted
the point that the sponsor should be in serious need of the BI
capabilities for a speci?c business purpose.
A BI system implementation is an adaptive information
improvement initiative for decision support [3, 4]. Some Delphi
interviewees further indicated that the typical application-based
funding model for implementation of transactional systems does
not apply to BI systems that are evolutionary in nature. That is,
a BI system evolves through an iterative process of development
in accordance with dynamic business requirements [19].
Therefore the BI initiative, especially for the enterprise-wide
scale, requires consistent funding and resource allocation directly
from senior management to overcome continual organisational
issues. Contrary to conventional OLTP-based systems, these
organisational challenges arise during the course of the cross-
functional implementation, as it often uncovers many issues in
such areas as business processing, data ownership, data quality
and stewardship, and organisational structure. Many functional
units tend to focus on tactical gains, ignoring the rippling effects
imposed on other business units, and one participating expert
observed that,
“The whole BI effort cut across many areas in the
organisation that’s making it very dif?cult, it hits a lot of
political barriers. For instance, for a system owner, they
are only interested in delivering day to day transaction, as
long as all that done . . . that’s what they care about.”
Therefore the commitment and involvement of senior
management is imperative, particularly in breaking down
the barriers to change and the ‘states of mind’ within the
organisation.
Clear vision and well-established business case. As a BI
initiative is driven by business, so a strategic business vision is
needed to direct the implementation. Many Delphi participants
indicated that a long-term vision, primarily in strategic and
organisational terms, is needed to establish a solid business case.
The business case must be aligned to the strategic vision, thereby
meeting the business objectives and needs. If the business vision
is not thoroughly understood, it would eventually impact the
adoption and outcome of the BI system. Speaking to this point,
an interviewee emphasised that,
“In order for BI initiatives to be taken seriously and to
be supported by corporate leadership, they need to be
integrated with the overall strategy. Otherwise they will
not receive the leadership support that is required to make
them successful. The vision is the tool that leadership
can quickly understand and identify the linkages to the
corporate strategy.”
Many participants argued that the overriding reason some BI
projects fail is not due to technical challenges, because many of
the technological issues have proven answers. Rather, the most
common cause for failure is that the BI initiative does not align
with the business vision and so fails to meet the core objectives
of the business. As a result, the BI system will not satisfy the
business needs and neither will it satisfy the customers. The
possession of a well established business case is important
for sustaining organisational commitment to a new BI system.
Most interviewees rejected the notion that if an excellent system
was established then people would want to use it. In fact, one
interviewee claimed that,
“A BI system that is not business driven is a failed system!
BI is a business-centric concept. Sending IT off to solve a
problem rarely results in a positive outcome. There must be
a business problem to solve.”
Many participants stressed that a solid business case that was
derived from a detailed analysis of business needs would increase
the chances of winning support from top management. As stated
?rmly by one expert,
“In order for the leadership to support, they must
understand; when they understand and can easily explain
and provide the support needed. Of course, the business
case is an extremely important tool for both leadership and
the implementation team.”
Thus, a substantial business case should identify the proposed
strategic bene?ts, resources, risks, costs, and timeline. More
signi?cantly, it is important to understand that a BI system
implementation is not a project, it is a process [4]. That is, BI
systems are organic in nature. They evolve dynamically and in
directions that are not necessarily ?nite and predictable. For
instance, the warehouse data size of most BI systems doubles
during the ?rst year of operation, and the number of users also
increases markedly [22].
Process dimension
Business-centric championship and balanced team
composition. Most participants believed that having the right
champion from the business side of the organisation is critical
for implementation success. They expressed the view that a
champion who has excellent business acumen is always important
since he/she will be able to foresee the organisational challenges
and change course accordingly. More importantly, this business-
Spring 2010 Journal of Computer Information Systems 27
centric champion would view the BI system primarily in strategic
and organisational perspectives, as opposed to one who might
over-focus on the technical issues. For example, one interviewee
commented that,
“The team needs a champion. By a champion, I do not
mean someone who knows the tools. I mean someone who
understands the business and the technology and is able to
translate the business requirements into a (high-level) BI
architecture for the system.”
In fact, a BI initiative often spans multiple functional units
and demands extensive data and resources from these business
units. In this respect, the champion is critical to ensure the careful
management of the organisational challenges that arise during the
course of the project. Unlike operational system projects, such
challenges include getting system owners to recognise the strategic
value of their data and to re?ect on how their data interacts with
data from other transactional systems. Therefore, the champion
needs to ensure collaboration between business units and between
the business and the BI project team.
Organisations tend to rely on their IT staff to be solely
responsible for most system implementation projects. However,
BI projects are fundamentally different from OLTP projects [10,
25]. The project team must design a robust and maintainable
architecture that can accommodate the emerging and changing
requirements, this work requiring highly competent team
members. Not surprisingly, all interviewees agreed that the
composition and skills of a BI team have a major in?uence on
the success of the systems implementation. They indicated that
the BI team should be cross-functional and composed of both
technical and business personnel, so-called “best of both worlds”.
A BI initiative is essentially a business-driven project and is
critical for the making of strategic decisions. From a technical
perspective, a BI project is comparable to a systems integration
project and requires the active involvement of the business side
of the enterprise [19]. Typically, the project team has to deal
with diverse platforms, multiple interfaces, connection to legacy
systems, an array of tools, and so forth. All these tasks call for
people with different skills and competencies, and so a suitable
mix of technical and business expertise is a key to success.
Most experts recommended that a BI team should identify and
include business domain experts, especially for such activities
as data standardisation, requirement engineering, data quality
analysis, and testing. This enables the system design to be driven
by the business and ensures that the BI needs derived from
business are a driver of the logical data architecture. To enable
business users to navigate and manipulate the data model, the
structure and model of the data warehouse must be closely related
to their perception of the business objectives and processes.
Business-driven and iterative development approach. The
next factor to be considered is the business-driven and iterative
development approach. According to most Delphi participants,
adequate business-oriented project scoping and planning allow the
BI team to concentrate on the best opportunities for improvement.
Scoping helps in the selection of clear parameters and develops
a common understanding among all business stakeholders as to
what is in scope and what is excluded [1]. For instance, a Delphi
participant gave an in-sight into his experience,
“The success of 90 percent of our project is determined
prior to the ?rst day. This success is based on having a
very clear and well-communicated scope, having realistic
expectations and timelines, and having the appropriate
budget set aside.”
Most interviewees agreed that thorough scoping and planning
facilitate ?exibility and adaptability to changing requirements
within the time frame and resources. Moreover, adequate scoping
enables the project team to focus on crucial milestones and
pertinent issues while shielding them from becoming trapped in
unnecessary events. As one participant remarked,
“The scope needs to be controlled because ‘scope creep’
can cause a project to not meet its targeted conclusion.
That does not mean that you cannot have a change control
procedure or practice in place; this is a form of control. I
have seen many projects miss their delivery and cost goals
because of scope creep.”
Many experts stated that it is advisable to start with small
changes and developments and then to adopt an incremental
delivery, a so-called ‘iterative’ approach. Large-scale change
efforts are always fraught with greater risks given the substantial
variables to be managed simultaneously [1]. Moreover, modern
businesses are changing very quickly anyway and are always
seeking to identify the immediate impacts of those changes, and
so an incremental delivery approach is more cautious and provides
the tools for delivery of short, measurable steps. Furthermore, an
incremental delivery approach allows for building a long-term
solution as opposed to a short term one [1, 4]. As explained by
this interviewee,
“Adopting incremental delivery manages risks, provides
tangible results visible to the client, improves the client’s
ability to take ownership, eases knowledge transfer,
supports effective change management, and allows for
long-term solution.”
Therefore, the scope of a BI initiative should be selected
in such a way that a complete system for a speci?c business
sector can be delivered within a reasonable time, rather than
one ‘massive and complete big bang’ solution later on. Once the
users start working with the BI system, they will fully realise the
potential reporting and analysis possibilities. The preliminary BI
system is then further enhanced and developed in an evolutionary
and iterative approach. One interviewee elaborated that,
“You cannot roll out the whole BI system at once but people
want to see some key areas. You need to do data marts for
a couple of key areas and then maybe a small number of
other key reports in an attempt to keep all stakeholders
happy. Then when the ?rst release is done and you get
some feedback, you can work on other data mart areas and
enhance existing subject areas over time.”
Therefore, an incremental delivery approach allows an
organisation to concentrate on crucial issues, so enabling teams to
prove that the system implementation is feasible and productive
for the enterprise.
User-oriented change management. Having an adequate user-
oriented change management effort was deemed critical by most
28 Journal of Computer Information Systems Spring 2010
Delphi participants. They reported that better user participation in
the process of change can lead to better communication of their
needs, which in turn can help ensure successful introduction of
the system. Many Delphi participants shared the view that formal
user participation can help meet the demands and expectations of
various end users. No doubt, users know what they need better
than an architect or developer who lacks direct experience of the
product. This is mainly because business users will directly work
with the data models without an application layer that conceals
the complexity of the model (as is the case in conventional OLTP
systems) [22]. One Delphi participant commented that,
“Users should be an important partner in building and
delivering the right system. Without their consistent input,
we technicians cannot deliver the right system.”
This view was supported by another participing expert who
asked,
“How can the project team design and implement a BI system
to meet the users’ needs without their involvement?”
It is evident that key users must be involved throughout the
implementation cycle because they can provide valuable input
that the BI team may otherwise overlook. The data dimensions,
business rules, metadata, and data context that are needed by
business users should be incorporated into the system and validated
against the de?nition of deliverables [29]. Consequently, user
support will constantly evolve in response to organic business
requirement and supplementary BI applications [10].
Technological dimension
Business-driven, scalable and ?exible technical framework.
Turning now to technological issues, a key factor emphasised by
many Delphi respondents was that the technical framework of a
BI system must be able to accommodate scalability and ?exibility
requirements in line with dynamic business needs. That is, ?exible
and scalable infrastructure design allows for easy expansion of the
system to align it with evolving information needs [21]. So with a
strategic view embedded in the system design, this scalable system
framework could include additional data sources, attributes, and
dimensional areas for fact-based analysis, and it could incorporate
external data from suppliers, contractors, regulatory bodies, and
industrial benchmarks. It would then allow for the building of a
long-term solution to meet the incremental needs of business, as
explained by an interviewee,
“Scalability is always concerns to me. It seems that most
BI applications and systems always seem to grow to be
larger than expected or their throughput is greater than
anticipated. If the design is not scalable and ?exible,
it is more dif?cult to make changes to accommodate the
increase in size.”
In fact, a BI infrastructure involves all the tasks substantive
to path the technical layer for the entire BI environment. This
includes the implementation of new software and hardware,
the interoperability between the legacy systems and the new
BI environment on a network, as well as on a database level,
an administration subsystem and so on [19]. Establishing the
technical infrastructure for the initial BI solution is always time
consuming [29], but with the proper selection of scalable and
?exible hardware and software components, the effort would
be minimised for the next delivery cycle. As a consequence, the
system will be able to adapt to the emerging and ever-changing
business requirements.
Sustainable data quality and integrity. In regard to the
important factor of sustainable data quality and integrity, the
Delphi ?ndings indicate that the quality of data, particularly in
the source systems, is crucial if a BI system is to be implemented
successfully. According to most interviewees, a primary purpose
of a BI system is to integrate ‘silos’ of data for advanced analysis
so as to improve the decision-making process. Often, many data-
related issues within the back-end systems are not discovered until
that data are populated and queried within the BI system [31].
Thus data quality at sources will affect the quality of management
reports, which in turn in?uence the decision outcomes [9].
Corporate data can only be fully integrated and exploited for
greater business value once their quality and integrity are assured.
Speaking to this point, a BI expert asked, “If the data is corrupt
then what is the point?’ Another interviewee further exclaimed
that, “Without quality data the BI is not intelligence!” These
comments were echoed by another participant, who asserted,
“Garbage in garbage out. The user community doesn’t
care to understand why the information is wrong and once
you have a data integrity issue you are in trouble.”
Many Delphi participants believed that common measures and
de?nitions address the data quality dimensions of representational
consistency, interpretability and ease of understanding. This
allows all stakeholders to know that a term has a speci?c meaning
no matter where it is used across the source systems. It is typical
for a large organisation to have many terms with slightly different
meanings, because different business units tend to de?ne terms
in ways that best serve their purposes. Often, accurate data may
have been captured at the source level, but the record cannot be
used with other data sources due to inconsistent data identi?ers
[26]. This is because data values that should uniquely describe
entities are varied in different business units. A typical BI system
tends to be cross-functional and cross-departmental, so if only
one speci?c business section is scoped in the initial phase, the
business de?nitions and business rules must later be standardised
in order to be understood consistently on an enterprise level [19,
26]. This characteristic could have an impact on how the business
data are interpreted among different units. Once an organisation
has accumulated a large number of reports it becomes more
dif?cult to re-architect these areas. As a result, a cross-system
analysis is important to help pro?le a uniform master data set
which is in compliance with business rules. There needs to be
an organisational agreement on the de?nitions and measurements
that are part of the deliverables [26]. Hence, the development of
a master data set on which to base the logical data warehouse
construction for BI system will ease terminology problems. As a
result, the BI team would use common de?nitions to develop an
enterprise-wide dimensional model that is business orientated.
In short, this Delphi study was the ?rst step in exploring the
CSFs which can in?uence the implementation of BI systems. The
results show that there is a combination of multi-dimensional
CSFs peculiar to successful BI system implementation. More
importantly, the study has narrowed the research focus through
the identi?cation of a set of CSFs as presented. The next stage
Spring 2010 Journal of Computer Information Systems 29
comprised multiple case studies for the purpose of further
validating the CSFs ?ndings. The case studies examined whether
these critical factors — and/or any other factors — in?uence the
implementation success of BI systems.
STAGE TWO FINDINGS AND DISCUSSIONS
This section details and discusses the ?ndings of stage 2 case
studies. The BI system backgrounds of these case companies are
illustrated in Table 1.
As shown in Table 2, after analysis of the triangulated results
for all ?ve organisations, three instances of notable success
emerged (C1, C2 and C4), together with one moderately successful
case (C3) and one failure (C5). The three successful cases of BI
system implementation described their respective BI systems as
stable, easy to use, fully functional, ?exible and responsive within
anticipated times. Furthermore, the information generated was
considered accurate, timely, complete, consistent, and relevant to
most participants. In addition to the encouraging trend of system
use among end-users, the project leaders of these organisations
con?rmed that their implementation projects were completed on
time and within budget. However, the moderately successful case
was experiencing uncontrollable external factors in its BI system
implementation. The key application of its BI system was not
identical to those of conventional commercial enterprises. Due
to its unique form of business and the peculiar bonus system with
its major client, it was more concerned with ensuring on-time
delivery of assets and meeting quality and safety standards rather
than reducing costs or staf?ng. The BI system thus enables them
to analyse and investigate underlying business activities with
ease. Also, auditable reporting can be generated from the system
to assist the business meet its strict regulatory requirements.
On the other hand, the ?rm that experienced BI failure did
so because it encountered business issues at the early phase of
its implementation process. The business needs and requirements
for BI system had not been clearly de?ned, yet there existed silo
information systems with multiple versions of the truth. In that
?rm the BI initiative was driven mainly by the information system
manager alone and was viewed as a technological issue, and as
a result the management had to suspend the BI initiative. This
instance of failure served as a useful contrast case for comparative
analysis in this research.
Next, to demonstrate how the implementation success
compared against the management of the CSFs of the ?ve case
organisations, an analysis of the CSFs 1 to 7 was conducted
through a cross-case analysis. Table 3 summarises the relevant
TABLE 1 — Case Company and Its BI System Background
30 Journal of Computer Information Systems Spring 2010
TABLE 2 — Implementation Success for the Five Cases
CSFs performance in matrices recommended by Miles and
Huberman [18], and these were used as an initial step in analysing
patterns in the data. For each case, management of each CSF is
rated through a summary rating of P (for a CSF that was fully-
addressed), P (for a CSF that was partially addressed), or X (for a
CSF that was ignored).
Essentially, the evidence from these studies clearly
substantiated the construct and applicability of the multi-
dimensional framework. More importantly, the studies further
reveal the signi?cance of addressing those CSFs through the
business orientation approach. That is, without a speci?c business
purpose, the BI initiatives rarely produce a substantial impact
on business. As a result, the implementation of a BI system
has a much greater likelihood for success when business needs
are identi?ed at the outset and used as the driver behind the
implementation effort. Thus, the entire system implementation
must be business-driven and organisation-focussed. It should
also have interactive business-side involvement, and be adapted
to meet evolving business requirements throughout the lifecycle.
Invariably, a ‘build it and they will come’ approach which
overlooks business-focused strategies in system implementation
TABLE 3 — Evaluation of Critical Success Factors in Multiple Organisations
proves to be unsatisfactory and very expensive. In other words,
this particular meta-factor (i.e. a business orientation approach)
dictates the commandment of the proposed CSFs.
Notably, the three successful cases (C1, C2 and C4) seemed
to emphasise the business-oriented approach when addressing
the CSFs, while the partially successful case (C3) appeared to
comprise a mixture of business and customer-centric approaches.
The instance of failure (C5) was not totally business-driven but
instead was technology oriented. The three successful cases
shifted their focus from the technological view and instead
adopted an approach that put their respective business needs ?rst.
On the basis of these case studies, it is apparent that the manner
in which an organisation addresses those CSFs, whether through
a business-oriented, technology-oriented, or customer-oriented
approach, will have a substantial impact on the implementation
outcome. Having a clearly-de?ned set of CSFs is important, but it
is even more critical to address the CSFs from the right approach.
In the case of BI systems implementation, the triangulated data of
case studies clearly demonstrates that by placing business needs
ahead of other issues an enterprise has a higher likelihood of
achieving a useful BI system.
Spring 2010 Journal of Computer Information Systems 31
What is more, in order to meet the need for systems which
provide management with dynamic analytics and business
reporting, the ?ndings of these case studies indicate that
business stakeholders should involve interactively throughout
the implementation process. In other words, it necessitates the
participation of business stakeholders in the development of a
reporting that usually demands practical business experience.
Moreover, due to evolving business needs and ever-changing
information requirements, it was found that the respective BI
teams had to provide continual high-level maintenance and
support not only on tools application, but also at broader data
modelling and system scalability issues. The designing of data
models and system architecture frameworks needs consistent
input from those most familiar with the business needs of the
enterprise.
In summary, the three successful cases clearly demonstrated
that addressing the CSFs from a business perspective was the
cornerstone on which they successfully based the implementation
of their BI systems. Conversely, the unsuccessful case failed
because it focused primarily on the technology and neglected the
core requirements of its business. In order to better address the
CSFs it is essential for an organisation to emphasise the business
orientation approach, and in so doing it will gain an advantage
over competitors. Indeed, this view was supported by Gartner
Research [6] who stated that, “best in class organisations focus on
business objectives and use a business-driven approach to de?ne
and scope their people, process, application, technology and/or
services strategy.”
CONCLUDING REMARKS
AND CONTRIBUTIONS
Understanding CSFs is a key for successful implementation of
a BI system. This study examined the CSFs impacting BI systems
implementation. A set of multi-dimensional CSFs was identi?ed
during the course of three rounds of a Delphi study with 15 BI
system experts. The ?ndings from the Delphi study were then
examined empirically in case studies of ?ve large organisations.
The evidence from these studies clearly con?rmed the construct
and applicability of the CSFs framework.
An analysis of the ?ndings further indicates that non-technical
factors, including organisational and process-related factors, are
more in?uential and important than technological and data-related
factors. Furthermore, the present study also gives evidence that
the contextual issues of the CSFs are quite different from the
implementation of other systems. Therefore, these CSFs cannot
be applied to BI systems without giving careful consideration to
the relevant contextual issues.
It appears that there is a macro-level pattern for interpreting
the CSFs related to such infrastructure-based projects. Both the
organisational and process dimensions are probably generic
and vary somewhat among BI systems and other infrastructural
systems implementation. But it is apparent that there is a new
understanding of factors associated with the technological
dimension due to the technical challenges that vary with the nature
of the infrastructure system. Nonetheless, this research suggests
that organisations are in a better position to successfully address
those CSFs through the business-orientation approach. That
is, without a clear business-driven objective, the BI initiatives
rarely produce substantial impact on business. As a result, the
implementation of a BI system has a much greater likelihood of
success when speci?c business needs are identi?ed at the outset,
and when those needs are used to direct the nature and scope of the
implementation effort. Therefore, this business orientation meta-
CSF should be regarded as the most critical factor in determining
the implementation success of BI systems.
This research has made a theoretical contribution to
our understanding of the CSFs that impact on BI systems
implementation. The literature review reveals relatively little
previous work on this subject. This study helps to ?ll the gap by
building the theory of the ways in which CSFs impact BI systems
implementation. This study represents the ?rst rigorously-
conducted enquiry which will develop our understanding of the
factors that affecting the implementation of BI systems. The
?ndings and outcomes extend current theory and allow ?rms to
identify and focus their scarce resources in those CSFs areas.
Besides that, academic researchers are often criticised for failing
to address issues of concern to practitioners. The collection
and analysis of empirical data in this study responds to those
criticisms and supplements the current limited understanding
of the factors that affect the successful implementation of BI
systems. The result of this work highlights those factors that need
to be addressed, and it also points out those that are not so critical.
Hence, it focuses attention on those important areas that might
otherwise be neglected or taken for granted but are signi?cant for
the implementation success.
Not only does this research contribute to the academic
literature on this topic but it bene?ts organisations in several ways
as well. First, large and complex organisations that are planning
to implement enterprise level BI systems will be better able to
identify those factors that will enhance the likelihood of success.
The ?ndings will help them to determine those factors on which
they should give particular attention to ensure that they receive
continuous management scrutiny. For senior management,
this research ?nding can certainly assist them by optimising
their scarce resources on those key areas that will improve the
implementation process. Also, management can concentrate on
monitoring, controlling and supporting only those critical areas.
The ?ndings with regard to the CSFs represent best practices
for ?rms that have successfully implemented BI systems. The
evidence that was revealed provides reference for BI stakeholders
that can increase the chances of implementation success.
REFERENCES
[1] Ang, J. & Teo, T. S. H. “Management Issues in Data
Warehousing: Insights from the Housing and Development
Board,” Journal of Decision Support Systems, 29(1), 2000,
11-20.
[2] Ariyachandra, T. & Watson, H. “Which Data Warehouse
Architecture Is Most Successful?” Business Intelligence
Journal, 11(1), 2006.
[3] Arnott, D. “Decision Support Systems Evolution:
Framework, Case Study and Research Agenda,” European
Journal of Information Systems, 13(4), 2004, 247-259.
[4] Arnott, D. & Pervan, G. “A Critical Analysis of Decision
Support Systems Research,” Journal of Information Tech-
nology, 20(2), 2005, 67-87.
[5] Bose, R. “Advanced Analytics: Opportunities and
Challenges,” Industrial Management & Data Systems,
109(2), 2009, 155-172.
[6] Burton, B., Geishecker, L., & Hostmann, B. “Organizational
Structure: Business Intelligence and Information Manage-
ment,” Gartner Research, 2006.
32 Journal of Computer Information Systems Spring 2010
[7] Delone, W., & McLean, E. “Information Systems Success:
The Quest for the Dependent Variable,” Journal of
Information System Research, 3(1), 1992, 60-95.
[8] Delone, W., & McLean, E. “The DeLone and McLean Model
of Information Systems Success: A Ten-Year Update,”
Journal of Management Information Systems, 19(4), 2003,
9-30.
[9] Friedman, T. “Gartner Says More Than 50 Percent of Data
Warehouse Projects Will Have Limited Acceptance or Will
Be Failures Through 2007,” 2005. Retrieved 21 Feb 2009,
fromhttp://www.gartner.com/it/page.jsp?id=492112
[10] Fuchs, G. “The Vital BI Maintenance Process”, in Business
Intelligence Implementation: Issues and Perspectives,” In
B. Sujatha (Ed), ICFAI University Press, Hyderabad, 2006,
116-123.
[11] Gartner Research, “Gartner EXP Worldwide Survey of
1,500 CIOs Shows 85 Percent of CIOs Expect ‘Signi?cant
Change’ Over Next Three Years,” 2008. Retrieved 21 Feb
2009, fromhttp://www.gartner.com/it/page.jsp?id=587309
[12] Gartner Research, “Gartner EXP Worldwide Survey of
More than 1,500 CIOs Shows IT Spending to Be Flat in
2009”, 2009. Retrieved 21 Feb 2009, fromhttp://www.
gartner.com/it/page.jsp?id=855612
[13] Hwang, M. & Xu, H. “A Structural Model of Data
Warehousing Success,” Journal of Computer Information
Systems, Fall 2008, 48-56
[14] Inmon, W.H. Building the Data Warehouse 4 Edn, John
Wiley & Sons, Indianapolis, 2005.
[15] Jagielska, I., Darke, P., & Zagari, G. “Business Intelligence
Systems for Decision Support: Concepts, Processes and
Practice,” Proceedings of the 7th International Conference
of the International Society for Decision Support Systems,
2003.
[16] Little, R.G. Identi?cation of Factors Affecting the Imple-
mentation of Data Warehousing, Thesis, Auburn University,
1998.
[17] Linstone, H. A., & Turoff, M. The Delphi Method:
Techniques and Applications. Addison-Wesley, Reading,
MA, 2002.
[18] Miles, M. & Huberman, A.M. Qualitative Data Analysis: An
Expanded Sourcebook, Sage, Thousand Oaks, CA. 1994.
[19] Moss, L. & Atre, S. Business Intelligence Roadmap: The
Complete Lifecycle for Decision-Support Applications.
Addison-Wesley, Boston, MA. 2003.
[20] Negash, S. “Business Intelligence,” Communications of the
Association for Information Systems, 13, 2004, 177-195.
[21] Olszak, C & Ziemba, E. “Approach to Building and
Implementing Business Intelligence Systems,” Interdis-
ciplinary Journal of Information, Knowledge, and Man-
agement, 2, 2007, 135-148.
[22] Reinschmidt, J. & Francoise, A. Business Intelligence
Certi?cation Guide, IBM International Technical Support
Organization, San Jose, CA, 2000.
[23] Shin, B. “An Exploratory Investigation of System Success
Factors in Data Warehousing,” Communications of the
Association for Information Systems, 6(4), 2003, 141-170.
[24] Stuart I., McCutcheon D. & Hand?eld R. “Effective
Case Research in Operations Management: A Process
Perspective,” Journal of Operations Management, 20(5),
2002, 419-433.
[25] Turban, E., Sharda, R., Aronson, J. & King, D. Business
Intelligence, Prentice Hall, New Jersey, 2007.
[26] Wadehra, A. “The ABCs of Master Data Management:
Architecture, Business Case, and Customer,” Business
Intelligence Journal, 12(1), 2007.
[27] Wang, H. & Wang S. “A Knowledge Management Approach
to Data Mining Process for Business Intelligence,” Industrial
Management & Data Systems, 108(5), 2008, 622-634.
[28] Watson, H., Abraham, D., Chen, D. “Data Warehousing
ROI: Justifying and Assessing a Data Warehouse,” Business
Intelligence Journal, 2004, 6-17.
[29] Watson, H., & Haley, B. “Managerial considerations,”
Communications of the ACM, 41(9), 1998, 32-37.
[30] Watson, H., Annino, D. A., Wixom, B. H. “Current Practices
in Data Warehousing,” Journal of Information Systems
Management, 18(1), 2001, 1-9.
[31] Watson, H., Fuller, C., & Ariyachandra, T. “Data Warehouse
Governance: Best Practices at Blue Cross and Blue Shield
of North Carolina,” Journal of Decision Support Systems,
38(3), 2004, 435-450.
[32] Wixom, B., & Watson, H. “An Empirical Investigation of
the Factors Affecting Data-Warehousing Success,” MIS
Quarterly, 25(1), 2001, 17-41.
[33] Yin, R. Case Study Research: Design and Methods 3 Edn,
Sage, Thousand Oaks, CA, 2003.
[34] Ziglio, E., & Adler, M. Gazing into the Oracle: The Delphi
Method and its Application to Social Policy and Public
Health, Jessica Kingsley Publishers, London, 1996.
doc_794498715.pdf
The implementation of a business intelligence (BI) system is a complex undertaking requiring considerable resources.
Spring 2010 Journal of Computer Information Systems 23
CRITICAL SUCCESS FACTORS FOR
BUSINESS INTELLIGENCE SYSTEMS
WILLIAM YEOH ANDY KORONIOS
University of South Australia University of South Australia
SA 5095 Australia SA 5095 Australia
Received: June 11, 2009 Revised: August 6, 2009 Accepted: September 14, 2009
The implementation of a business intelligence (BI) system is a
complex undertaking requiring considerable resources. Yet there
is a limited authoritative set of critical success factors (CSFs) for
management reference because the BI market has been driven
mainly by the IT industry and vendors. This research seeks to
bridge the gap that exists between academia and practitioners by
investigating the CSFs in?uencing BI systems success. The study
followed a two-stage qualitative approach. Firstly, the authors
utilised the Delphi method to conduct three rounds of studies.
The study develops a CSFs framework crucial for BI systems
implementation. Next, the framework and the associated CSFs
are delineated through a series of case studies. The empirical
?ndings substantiate the construct and applicability of the
framework. More signi?cantly, the research further reveals that
those organisations which address the CSFs from a business
orientation approach will be more likely to achieve better results.
Keywords: Business intelligence system, Critical success
factors, Delphi method, Case study
INTRODUCTION
Recently Business Intelligence (BI) applications have been
dominating the technology priority list of many CIOs [11, 12].
According to Reinschmidt and Francoise [22], a BI system is “an
integrated set of tools, technologies and programmed products that
are used to collect, integrate, analyse and make data available”.
Stated simply, the main tasks of a BI system include “intelligent
exploration, integration, aggregation and a multidimensional
analysis of data originating from various information resources”
[21]. Implicit in this de?nition, data is treated as a highly valuable
corporate resource, and transformed from quantity to quality
[27]. As a result, massive data from many different sources of a
large enterprise can be integrated into a coherent body to provide
‘360 degrees’ view of its business [5, 27]. Hence, meaningful
information can be delivered at the right time, at the right location,
and in the right form [5, 20] to assist individuals, departments,
divisions or even larger units to facilitate improved decision-
making [15].
While the BI market appears vibrant and the importance of BI
systems is more widely accepted, few studies have investigated
the critical success factors that affect the implementation
success. Although there exist a plethora of guidelines from the IT
industry, most rely on anecdotal reports [15]. This is because the
study of BI systems is a relatively new area that has been driven
primarily by the IT industry and by vendors [15]. Therefore,
empirical research to shed more light on those critical success
factors (CSFs) in?uencing the implementation of BI systems is
desirable. An understanding of the CSFs enables BI stakeholders
to optimise their scarce resources and efforts by focussing on
those signi?cant factors that are most likely to aid successful
system implementation.
RESEARCH MOTIVATION
The implementation of a BI system is not a conventional
application-based IT project (such as an operational or
transactional system), which has been the focus of many CSF
studies [10]. Instead, it shares similar characteristics with other
infrastructural projects such as enterprise resource planning (ERP)
systems implementation. That is, implementing a BI system is not
a simple activity entailing merely the purchase of a combination
of software and hardware; rather, it is a complex undertaking
requiring appropriate infrastructure and resources over a lengthy
period [10, 19, 28].
Speci?cally, the key infrastructural foundation for most
enterprise level BI systems — a data warehouse — is a subject-
oriented, integrated, time-variant and non-volatile collection of
data that differ from conventional online transactional processing
(OLTP) databases [14]. A complex data structure must be maintained
in order to provide an integrated view of the organisation’s data
so users can query across departmental boundaries for dynamic
retrieval of relevant decision-support information. Furthermore,
the BI system’s architecture is highly complex owing to the back-
end systems originating from multiple data sources and to the vast
volume of data to be processed. In addition, the implementation
of a BI system is often associated with the following challenges:
underlying original back-end systems and processes which were
not adapted for BI applications; poor data quality derived from
source systems that can often go unnoticed until cross-systems
analysis is conducted; and the maintenance process that tends to
be vague and ill-de?ned [10, 19, 23].
Despite the complexities in implementing BI systems, there
has been little empirical research about the CSFs impacting the
implementation of BI systems. The gap in the literature is re?ected
in the low level of contributions to international conferences and
journals. More importantly, the value of previous CSF studies
will obviously decline with age [16]. The rapid advancement of
technological innovation and the pace at which new technologies
are being adopted will apparently in?uence the state of criticality
for research into CSFs [16]. Furthermore, CSFs applicable to
other types of information systems may not necessarily apply
to a contemporary BI system. Therefore, the increased rate of
adoption of BI systems, the complexities of implementing a
contemporary BI system, the scarcity of academic research, and
24 Journal of Computer Information Systems Spring 2010
the far-reaching business implications justify a more focused
examination of CSFs as well as the associated contextual issues
required for implementing BI systems. This paper describes an
exploration of an important and topical area of interest — the
CSFs impacting the implementation of BI systems. It is expected
that this research will make a contribution to both theory and
practice. In theoretical terms, this research:
• adds to knowledge and contributes to the literature of
an emerging area of interest — the implementation
of BI systems, in particular, the CSFs that affect the
implementation effort;
• identi?es the criteria which determine the success of BI
systems implementation; and
• validates current CSFs understandings, and extends our
knowledge of contemporary BI systems.
In practical terms, the research project:
• identi?es the CSFs that impact on BI systems imple-
mentation, so enabling stakeholders to better use their
scarce resources by focussing on those key areas that
are most likely to have a greater impact.
The remainder of this article has been structured as follows.
The next section describes the two-stage research methodology
in this study. The later section presents and discusses the research
?ndings. In the last section the authors conclude the overall study
and then state their contributions.
RESEARCH DESIGN
This research adopted an interpretivist paradigm and followed
a two-stage qualitative approach. Based on extensive literature
review, a research framework and associated interview questions
were developed for use in the Stage 1 exploratory Delphi study.
This research framework was re?ned with the data gathered from
15 Delphi participants. In the second stage, the preliminary CSFs
framework derived from the Delphi study was further veri?ed
in ?ve case studies of large and complex organisations. Each
case study involved interviews with the BI stakeholders, and the
collection of project documents. Cross-case analysis was done
to examine the CSFs. As a result, a re?ned CSF framework was
developed.
Stage One — Delphi Study
In the absence of much useful literature on BI system, this
stage one of research seeks to explore and identify a set of CSFs
that are jointly agreed by a group of BI system experts. The Delphi
method was deemed to be the most appropriate method for this
study because it allows the gathering of subjective judgements
which are moderated through group consensus [17]. Moreover, this
research assumes that expert opinion can be of signi?cant value in
situations where knowledge or theory is incomplete [17]. For this
study, a Delphi panel composed of ?fteen BI systems experts was
established. Ziglio and Adler [34] assert that useful results can be
obtained from small group of 10-15 experts. Beyond this number,
further increases in understandings are small and not worth the
cost or the time spent in additional interviewing. Thus, the size of
such a Delphi panel is deemed suitably representative.
The Delphi study comprised three rounds. During the ?rst
round the authors conducted face-to-face interviews with each
participant. After the interview, further clari?cations (if any)
were made by follow-up phone calls and email communications.
Subsequently, the data gathered from the ?rst round of interviews
were analysed thoroughly by content analysis technique, a
constant comparison technique, to identify major themes [18].
In other words, the qualitative data were examined thematically
and emergent themes were ranked by their frequency and later
categorised.
In the second round, the suggested factors of all the
participants were consolidated into a single list. The list was then
distributed among the participants to facilitate comparison of the
expert’s perceptual differences. During the third round, the list of
candidate CSFs was surveyed by the Delphi participants using a
structured questionnaire survey approach. Speci?cally, a 5-point
Likert scale (i.e. from 1 ‘not important’ to 5 ‘critically important’)
was applied to rate the importance of the candidate CSFs in the
process of seeking consensus from the BI experts. From the
survey feedback, only those factors with standard deviations
(SD) of 1.0 or less, and average ratings of 3.5 and above, were
short-listed as CSFs because 1 SD from the mean contains 68%
of all scores. These criteria (i.e. SD3.5/5.0) offer
a working de?nition of a threshold for stability, and hence the
resultant CSFs are considered legitimate. Therefore, the existence
of CSFs within this de?nition in round three is considered to be a
critical point for terminating the Delphi study. The details of the
results are discussed below.
Stage Two — Case Study
Due to limited academic literature, the Stage 1 Delphi study
was used to narrow the CSFs focus of this research. However,
reliance on the Delphi study alone was not suf?cient for the
collection of data needed to rigorously address the research
objective. Therefore, a case study methodology was used for
Stage 2 of the theory-building process. That is, this second stage
sought to corroborate the CSFs ?ndings of Stage 1. The case study
methodology provides better explanations and understandings
on the examined phenomenon which would otherwise be lost
in other quantitative designs [18, 33]. For this study, in contrast
to sampling logic, a case study is an empirical investigation
following replication logic that leads to analytic generalisation
[33]. Thus multiple case studies in this research should be
regarded as multiple experiments and not multiple respondents
in a survey [33]. That is, relevance rather than representativeness
is prioritised in case selection. Given that the objective of this
study was to build theory, a case study process with multiple-case
design was the appropriate approach, and the use of the case study
methodology is justi?ed on these grounds.
Data collection for this study entailed semi-structured
interviews with key stakeholders of BI projects. To facilitate
data triangulation, data were also gathered from a number of
sources including relevant documents. A cross case analysis
approach was used in this study to gain better understandings
and increase the generalisability of the ?ndings [18]. In searching
for patterns, the authors examined similarities and differences
about relationships within the data. Hence, varying the order in
which case data are arrayed enables patterns to become more
obvious [24]. This research did not produce quantitative data. In
all cases, the authors were examining the presence or absence
of a particular CSF (e.g., were adequate resources provided?),
while at the same time ascertaining whether that characteristic
Spring 2010 Journal of Computer Information Systems 25
was ful?lled in a meaningful way. To assess the importance of
the seven previously-identi?ed CSFs, the authors studied ?ve
organisations that had implemented BI systems, including rail
corporations, energy utilities, water utilities, and a ship-builder.
STAGE ONE CSF
FINDINGS AND DISCUSSIONS
This section presents the ?ndings and interpretation of the stage
one Delphi study. As illustrated in Figure 1, this CSFs framework
outlines how a set of critical factors contributes to the success
of a BI system implementation. Following Ariyachandra and
Watson [2], the implementation success criteria of this research
take into account two key dimensions: process performance (i.e.
how well the process of a BI system implementation went), and
infrastructure performance (i.e. the quality of the system and the
standard of output). The infrastructure performance has parallels
with the three major IS success variables described by Delone and
McLean [7, 8], namely system quality, information quality, and
system use, whereas process performance can be assessed in terms
of time-schedule and budgetary considerations [2]. Speci?cally,
system quality is concerned with the performance characteristics
of the information processing system itself, in which the system
should be ?exible, scalable and able to integrate data [2, 7, 8].
Information quality refers to accuracy, completeness, timeliness,
relevance, consistency, and usefulness of information generated
by the system [2, 7, 8]. System use is de?ned as “recipient
consumption of the output of an information system” [7, 8].
Subsequently, individual users and their respective organisations
would assess the bene?ts of the BI system implementation [13].
This perception of the bene?ts would then become part of an
interactive, business-driven evolutionary continuum to further
support evolving business needs for improved BI systems [3,
4]. That is, a BI system implementation is viewed as an organic
cycle that evolves over time. Based on constant evaluation of
the information, as well as user feedback, the system resembles
a loop that requires re-evaluation of existing BI solutions,
and subsequently the system will be modi?ed, optimised and
improved accordingly. In other words, completion of the system
implementation does not mean that all BI related problems are
resolved [21]. The system will succeed only when business users
keep identifying and modelling knowledge, as well as monitoring
and modifying data repositories on an ongoing basis [21]. Hence,
the entire process is cyclical, but with a series of interrelated steps
[25].
In brief, this framework treats the CSFs as necessary for
implementation success of a BI system, whereas the absence
of the CSFs would lead to failure of the system. The Delphi
participants provided detail and justi?cation to those critical
factors. The CSF framework details the CSFs identi?ed in the
?rst stage of this study, and they are presented according to the
major dimensions of interest proposed by Wixom and Watson
[32], namely organisation, process, and technology. For each CSF
description, the primary data came from the ?ndings of the third-
round consensus amongst the Delphi participants. The secondary
data came from the ?rst two rounds of qualitative interviews with
individual participants. Additional data from published literature
were also used to support the arguments of the participants.
FIGURE 1 — CSFs Framework for Implementation of BI Systems
26 Journal of Computer Information Systems Spring 2010
Organisational dimension
Committed management support and sponsorship.
Committed management support and sponsorship has been
widely acknowledged as the most important factor for BI system
implementation. All Delphi participants agreed that consistent
support and sponsorship from business executives make it easier
to secure the necessary operating resources such as funding,
human skills, and other requirements. One interviewee stated
?rmly, “If you don’t have top level sponsorship — it is doomed!”
Another participant explained the situation this way,
“Project Sponsorship has been shown to be the single most
important determinant of IT project success or failure. A
BI project is no different to any other IT project in this
respect . . . Maintaining the commitment and support of
the projects sponsor throughout the project — because
circumstances can change over the life of the project.”
Many participants also asserted that it is more bene?cial if
the sponsor is from the business side of the enterprise rather
than from the IT sector. Similarly, a study by Watson et al. [30]
indicates that the ideal BI sponsor should come from a business
function. Such a sponsor often has a strong stake in the success of
the BI initiative. Most importantly, some interviewees highlighted
the point that the sponsor should be in serious need of the BI
capabilities for a speci?c business purpose.
A BI system implementation is an adaptive information
improvement initiative for decision support [3, 4]. Some Delphi
interviewees further indicated that the typical application-based
funding model for implementation of transactional systems does
not apply to BI systems that are evolutionary in nature. That is,
a BI system evolves through an iterative process of development
in accordance with dynamic business requirements [19].
Therefore the BI initiative, especially for the enterprise-wide
scale, requires consistent funding and resource allocation directly
from senior management to overcome continual organisational
issues. Contrary to conventional OLTP-based systems, these
organisational challenges arise during the course of the cross-
functional implementation, as it often uncovers many issues in
such areas as business processing, data ownership, data quality
and stewardship, and organisational structure. Many functional
units tend to focus on tactical gains, ignoring the rippling effects
imposed on other business units, and one participating expert
observed that,
“The whole BI effort cut across many areas in the
organisation that’s making it very dif?cult, it hits a lot of
political barriers. For instance, for a system owner, they
are only interested in delivering day to day transaction, as
long as all that done . . . that’s what they care about.”
Therefore the commitment and involvement of senior
management is imperative, particularly in breaking down
the barriers to change and the ‘states of mind’ within the
organisation.
Clear vision and well-established business case. As a BI
initiative is driven by business, so a strategic business vision is
needed to direct the implementation. Many Delphi participants
indicated that a long-term vision, primarily in strategic and
organisational terms, is needed to establish a solid business case.
The business case must be aligned to the strategic vision, thereby
meeting the business objectives and needs. If the business vision
is not thoroughly understood, it would eventually impact the
adoption and outcome of the BI system. Speaking to this point,
an interviewee emphasised that,
“In order for BI initiatives to be taken seriously and to
be supported by corporate leadership, they need to be
integrated with the overall strategy. Otherwise they will
not receive the leadership support that is required to make
them successful. The vision is the tool that leadership
can quickly understand and identify the linkages to the
corporate strategy.”
Many participants argued that the overriding reason some BI
projects fail is not due to technical challenges, because many of
the technological issues have proven answers. Rather, the most
common cause for failure is that the BI initiative does not align
with the business vision and so fails to meet the core objectives
of the business. As a result, the BI system will not satisfy the
business needs and neither will it satisfy the customers. The
possession of a well established business case is important
for sustaining organisational commitment to a new BI system.
Most interviewees rejected the notion that if an excellent system
was established then people would want to use it. In fact, one
interviewee claimed that,
“A BI system that is not business driven is a failed system!
BI is a business-centric concept. Sending IT off to solve a
problem rarely results in a positive outcome. There must be
a business problem to solve.”
Many participants stressed that a solid business case that was
derived from a detailed analysis of business needs would increase
the chances of winning support from top management. As stated
?rmly by one expert,
“In order for the leadership to support, they must
understand; when they understand and can easily explain
and provide the support needed. Of course, the business
case is an extremely important tool for both leadership and
the implementation team.”
Thus, a substantial business case should identify the proposed
strategic bene?ts, resources, risks, costs, and timeline. More
signi?cantly, it is important to understand that a BI system
implementation is not a project, it is a process [4]. That is, BI
systems are organic in nature. They evolve dynamically and in
directions that are not necessarily ?nite and predictable. For
instance, the warehouse data size of most BI systems doubles
during the ?rst year of operation, and the number of users also
increases markedly [22].
Process dimension
Business-centric championship and balanced team
composition. Most participants believed that having the right
champion from the business side of the organisation is critical
for implementation success. They expressed the view that a
champion who has excellent business acumen is always important
since he/she will be able to foresee the organisational challenges
and change course accordingly. More importantly, this business-
Spring 2010 Journal of Computer Information Systems 27
centric champion would view the BI system primarily in strategic
and organisational perspectives, as opposed to one who might
over-focus on the technical issues. For example, one interviewee
commented that,
“The team needs a champion. By a champion, I do not
mean someone who knows the tools. I mean someone who
understands the business and the technology and is able to
translate the business requirements into a (high-level) BI
architecture for the system.”
In fact, a BI initiative often spans multiple functional units
and demands extensive data and resources from these business
units. In this respect, the champion is critical to ensure the careful
management of the organisational challenges that arise during the
course of the project. Unlike operational system projects, such
challenges include getting system owners to recognise the strategic
value of their data and to re?ect on how their data interacts with
data from other transactional systems. Therefore, the champion
needs to ensure collaboration between business units and between
the business and the BI project team.
Organisations tend to rely on their IT staff to be solely
responsible for most system implementation projects. However,
BI projects are fundamentally different from OLTP projects [10,
25]. The project team must design a robust and maintainable
architecture that can accommodate the emerging and changing
requirements, this work requiring highly competent team
members. Not surprisingly, all interviewees agreed that the
composition and skills of a BI team have a major in?uence on
the success of the systems implementation. They indicated that
the BI team should be cross-functional and composed of both
technical and business personnel, so-called “best of both worlds”.
A BI initiative is essentially a business-driven project and is
critical for the making of strategic decisions. From a technical
perspective, a BI project is comparable to a systems integration
project and requires the active involvement of the business side
of the enterprise [19]. Typically, the project team has to deal
with diverse platforms, multiple interfaces, connection to legacy
systems, an array of tools, and so forth. All these tasks call for
people with different skills and competencies, and so a suitable
mix of technical and business expertise is a key to success.
Most experts recommended that a BI team should identify and
include business domain experts, especially for such activities
as data standardisation, requirement engineering, data quality
analysis, and testing. This enables the system design to be driven
by the business and ensures that the BI needs derived from
business are a driver of the logical data architecture. To enable
business users to navigate and manipulate the data model, the
structure and model of the data warehouse must be closely related
to their perception of the business objectives and processes.
Business-driven and iterative development approach. The
next factor to be considered is the business-driven and iterative
development approach. According to most Delphi participants,
adequate business-oriented project scoping and planning allow the
BI team to concentrate on the best opportunities for improvement.
Scoping helps in the selection of clear parameters and develops
a common understanding among all business stakeholders as to
what is in scope and what is excluded [1]. For instance, a Delphi
participant gave an in-sight into his experience,
“The success of 90 percent of our project is determined
prior to the ?rst day. This success is based on having a
very clear and well-communicated scope, having realistic
expectations and timelines, and having the appropriate
budget set aside.”
Most interviewees agreed that thorough scoping and planning
facilitate ?exibility and adaptability to changing requirements
within the time frame and resources. Moreover, adequate scoping
enables the project team to focus on crucial milestones and
pertinent issues while shielding them from becoming trapped in
unnecessary events. As one participant remarked,
“The scope needs to be controlled because ‘scope creep’
can cause a project to not meet its targeted conclusion.
That does not mean that you cannot have a change control
procedure or practice in place; this is a form of control. I
have seen many projects miss their delivery and cost goals
because of scope creep.”
Many experts stated that it is advisable to start with small
changes and developments and then to adopt an incremental
delivery, a so-called ‘iterative’ approach. Large-scale change
efforts are always fraught with greater risks given the substantial
variables to be managed simultaneously [1]. Moreover, modern
businesses are changing very quickly anyway and are always
seeking to identify the immediate impacts of those changes, and
so an incremental delivery approach is more cautious and provides
the tools for delivery of short, measurable steps. Furthermore, an
incremental delivery approach allows for building a long-term
solution as opposed to a short term one [1, 4]. As explained by
this interviewee,
“Adopting incremental delivery manages risks, provides
tangible results visible to the client, improves the client’s
ability to take ownership, eases knowledge transfer,
supports effective change management, and allows for
long-term solution.”
Therefore, the scope of a BI initiative should be selected
in such a way that a complete system for a speci?c business
sector can be delivered within a reasonable time, rather than
one ‘massive and complete big bang’ solution later on. Once the
users start working with the BI system, they will fully realise the
potential reporting and analysis possibilities. The preliminary BI
system is then further enhanced and developed in an evolutionary
and iterative approach. One interviewee elaborated that,
“You cannot roll out the whole BI system at once but people
want to see some key areas. You need to do data marts for
a couple of key areas and then maybe a small number of
other key reports in an attempt to keep all stakeholders
happy. Then when the ?rst release is done and you get
some feedback, you can work on other data mart areas and
enhance existing subject areas over time.”
Therefore, an incremental delivery approach allows an
organisation to concentrate on crucial issues, so enabling teams to
prove that the system implementation is feasible and productive
for the enterprise.
User-oriented change management. Having an adequate user-
oriented change management effort was deemed critical by most
28 Journal of Computer Information Systems Spring 2010
Delphi participants. They reported that better user participation in
the process of change can lead to better communication of their
needs, which in turn can help ensure successful introduction of
the system. Many Delphi participants shared the view that formal
user participation can help meet the demands and expectations of
various end users. No doubt, users know what they need better
than an architect or developer who lacks direct experience of the
product. This is mainly because business users will directly work
with the data models without an application layer that conceals
the complexity of the model (as is the case in conventional OLTP
systems) [22]. One Delphi participant commented that,
“Users should be an important partner in building and
delivering the right system. Without their consistent input,
we technicians cannot deliver the right system.”
This view was supported by another participing expert who
asked,
“How can the project team design and implement a BI system
to meet the users’ needs without their involvement?”
It is evident that key users must be involved throughout the
implementation cycle because they can provide valuable input
that the BI team may otherwise overlook. The data dimensions,
business rules, metadata, and data context that are needed by
business users should be incorporated into the system and validated
against the de?nition of deliverables [29]. Consequently, user
support will constantly evolve in response to organic business
requirement and supplementary BI applications [10].
Technological dimension
Business-driven, scalable and ?exible technical framework.
Turning now to technological issues, a key factor emphasised by
many Delphi respondents was that the technical framework of a
BI system must be able to accommodate scalability and ?exibility
requirements in line with dynamic business needs. That is, ?exible
and scalable infrastructure design allows for easy expansion of the
system to align it with evolving information needs [21]. So with a
strategic view embedded in the system design, this scalable system
framework could include additional data sources, attributes, and
dimensional areas for fact-based analysis, and it could incorporate
external data from suppliers, contractors, regulatory bodies, and
industrial benchmarks. It would then allow for the building of a
long-term solution to meet the incremental needs of business, as
explained by an interviewee,
“Scalability is always concerns to me. It seems that most
BI applications and systems always seem to grow to be
larger than expected or their throughput is greater than
anticipated. If the design is not scalable and ?exible,
it is more dif?cult to make changes to accommodate the
increase in size.”
In fact, a BI infrastructure involves all the tasks substantive
to path the technical layer for the entire BI environment. This
includes the implementation of new software and hardware,
the interoperability between the legacy systems and the new
BI environment on a network, as well as on a database level,
an administration subsystem and so on [19]. Establishing the
technical infrastructure for the initial BI solution is always time
consuming [29], but with the proper selection of scalable and
?exible hardware and software components, the effort would
be minimised for the next delivery cycle. As a consequence, the
system will be able to adapt to the emerging and ever-changing
business requirements.
Sustainable data quality and integrity. In regard to the
important factor of sustainable data quality and integrity, the
Delphi ?ndings indicate that the quality of data, particularly in
the source systems, is crucial if a BI system is to be implemented
successfully. According to most interviewees, a primary purpose
of a BI system is to integrate ‘silos’ of data for advanced analysis
so as to improve the decision-making process. Often, many data-
related issues within the back-end systems are not discovered until
that data are populated and queried within the BI system [31].
Thus data quality at sources will affect the quality of management
reports, which in turn in?uence the decision outcomes [9].
Corporate data can only be fully integrated and exploited for
greater business value once their quality and integrity are assured.
Speaking to this point, a BI expert asked, “If the data is corrupt
then what is the point?’ Another interviewee further exclaimed
that, “Without quality data the BI is not intelligence!” These
comments were echoed by another participant, who asserted,
“Garbage in garbage out. The user community doesn’t
care to understand why the information is wrong and once
you have a data integrity issue you are in trouble.”
Many Delphi participants believed that common measures and
de?nitions address the data quality dimensions of representational
consistency, interpretability and ease of understanding. This
allows all stakeholders to know that a term has a speci?c meaning
no matter where it is used across the source systems. It is typical
for a large organisation to have many terms with slightly different
meanings, because different business units tend to de?ne terms
in ways that best serve their purposes. Often, accurate data may
have been captured at the source level, but the record cannot be
used with other data sources due to inconsistent data identi?ers
[26]. This is because data values that should uniquely describe
entities are varied in different business units. A typical BI system
tends to be cross-functional and cross-departmental, so if only
one speci?c business section is scoped in the initial phase, the
business de?nitions and business rules must later be standardised
in order to be understood consistently on an enterprise level [19,
26]. This characteristic could have an impact on how the business
data are interpreted among different units. Once an organisation
has accumulated a large number of reports it becomes more
dif?cult to re-architect these areas. As a result, a cross-system
analysis is important to help pro?le a uniform master data set
which is in compliance with business rules. There needs to be
an organisational agreement on the de?nitions and measurements
that are part of the deliverables [26]. Hence, the development of
a master data set on which to base the logical data warehouse
construction for BI system will ease terminology problems. As a
result, the BI team would use common de?nitions to develop an
enterprise-wide dimensional model that is business orientated.
In short, this Delphi study was the ?rst step in exploring the
CSFs which can in?uence the implementation of BI systems. The
results show that there is a combination of multi-dimensional
CSFs peculiar to successful BI system implementation. More
importantly, the study has narrowed the research focus through
the identi?cation of a set of CSFs as presented. The next stage
Spring 2010 Journal of Computer Information Systems 29
comprised multiple case studies for the purpose of further
validating the CSFs ?ndings. The case studies examined whether
these critical factors — and/or any other factors — in?uence the
implementation success of BI systems.
STAGE TWO FINDINGS AND DISCUSSIONS
This section details and discusses the ?ndings of stage 2 case
studies. The BI system backgrounds of these case companies are
illustrated in Table 1.
As shown in Table 2, after analysis of the triangulated results
for all ?ve organisations, three instances of notable success
emerged (C1, C2 and C4), together with one moderately successful
case (C3) and one failure (C5). The three successful cases of BI
system implementation described their respective BI systems as
stable, easy to use, fully functional, ?exible and responsive within
anticipated times. Furthermore, the information generated was
considered accurate, timely, complete, consistent, and relevant to
most participants. In addition to the encouraging trend of system
use among end-users, the project leaders of these organisations
con?rmed that their implementation projects were completed on
time and within budget. However, the moderately successful case
was experiencing uncontrollable external factors in its BI system
implementation. The key application of its BI system was not
identical to those of conventional commercial enterprises. Due
to its unique form of business and the peculiar bonus system with
its major client, it was more concerned with ensuring on-time
delivery of assets and meeting quality and safety standards rather
than reducing costs or staf?ng. The BI system thus enables them
to analyse and investigate underlying business activities with
ease. Also, auditable reporting can be generated from the system
to assist the business meet its strict regulatory requirements.
On the other hand, the ?rm that experienced BI failure did
so because it encountered business issues at the early phase of
its implementation process. The business needs and requirements
for BI system had not been clearly de?ned, yet there existed silo
information systems with multiple versions of the truth. In that
?rm the BI initiative was driven mainly by the information system
manager alone and was viewed as a technological issue, and as
a result the management had to suspend the BI initiative. This
instance of failure served as a useful contrast case for comparative
analysis in this research.
Next, to demonstrate how the implementation success
compared against the management of the CSFs of the ?ve case
organisations, an analysis of the CSFs 1 to 7 was conducted
through a cross-case analysis. Table 3 summarises the relevant
TABLE 1 — Case Company and Its BI System Background
30 Journal of Computer Information Systems Spring 2010
TABLE 2 — Implementation Success for the Five Cases
CSFs performance in matrices recommended by Miles and
Huberman [18], and these were used as an initial step in analysing
patterns in the data. For each case, management of each CSF is
rated through a summary rating of P (for a CSF that was fully-
addressed), P (for a CSF that was partially addressed), or X (for a
CSF that was ignored).
Essentially, the evidence from these studies clearly
substantiated the construct and applicability of the multi-
dimensional framework. More importantly, the studies further
reveal the signi?cance of addressing those CSFs through the
business orientation approach. That is, without a speci?c business
purpose, the BI initiatives rarely produce a substantial impact
on business. As a result, the implementation of a BI system
has a much greater likelihood for success when business needs
are identi?ed at the outset and used as the driver behind the
implementation effort. Thus, the entire system implementation
must be business-driven and organisation-focussed. It should
also have interactive business-side involvement, and be adapted
to meet evolving business requirements throughout the lifecycle.
Invariably, a ‘build it and they will come’ approach which
overlooks business-focused strategies in system implementation
TABLE 3 — Evaluation of Critical Success Factors in Multiple Organisations
proves to be unsatisfactory and very expensive. In other words,
this particular meta-factor (i.e. a business orientation approach)
dictates the commandment of the proposed CSFs.
Notably, the three successful cases (C1, C2 and C4) seemed
to emphasise the business-oriented approach when addressing
the CSFs, while the partially successful case (C3) appeared to
comprise a mixture of business and customer-centric approaches.
The instance of failure (C5) was not totally business-driven but
instead was technology oriented. The three successful cases
shifted their focus from the technological view and instead
adopted an approach that put their respective business needs ?rst.
On the basis of these case studies, it is apparent that the manner
in which an organisation addresses those CSFs, whether through
a business-oriented, technology-oriented, or customer-oriented
approach, will have a substantial impact on the implementation
outcome. Having a clearly-de?ned set of CSFs is important, but it
is even more critical to address the CSFs from the right approach.
In the case of BI systems implementation, the triangulated data of
case studies clearly demonstrates that by placing business needs
ahead of other issues an enterprise has a higher likelihood of
achieving a useful BI system.
Spring 2010 Journal of Computer Information Systems 31
What is more, in order to meet the need for systems which
provide management with dynamic analytics and business
reporting, the ?ndings of these case studies indicate that
business stakeholders should involve interactively throughout
the implementation process. In other words, it necessitates the
participation of business stakeholders in the development of a
reporting that usually demands practical business experience.
Moreover, due to evolving business needs and ever-changing
information requirements, it was found that the respective BI
teams had to provide continual high-level maintenance and
support not only on tools application, but also at broader data
modelling and system scalability issues. The designing of data
models and system architecture frameworks needs consistent
input from those most familiar with the business needs of the
enterprise.
In summary, the three successful cases clearly demonstrated
that addressing the CSFs from a business perspective was the
cornerstone on which they successfully based the implementation
of their BI systems. Conversely, the unsuccessful case failed
because it focused primarily on the technology and neglected the
core requirements of its business. In order to better address the
CSFs it is essential for an organisation to emphasise the business
orientation approach, and in so doing it will gain an advantage
over competitors. Indeed, this view was supported by Gartner
Research [6] who stated that, “best in class organisations focus on
business objectives and use a business-driven approach to de?ne
and scope their people, process, application, technology and/or
services strategy.”
CONCLUDING REMARKS
AND CONTRIBUTIONS
Understanding CSFs is a key for successful implementation of
a BI system. This study examined the CSFs impacting BI systems
implementation. A set of multi-dimensional CSFs was identi?ed
during the course of three rounds of a Delphi study with 15 BI
system experts. The ?ndings from the Delphi study were then
examined empirically in case studies of ?ve large organisations.
The evidence from these studies clearly con?rmed the construct
and applicability of the CSFs framework.
An analysis of the ?ndings further indicates that non-technical
factors, including organisational and process-related factors, are
more in?uential and important than technological and data-related
factors. Furthermore, the present study also gives evidence that
the contextual issues of the CSFs are quite different from the
implementation of other systems. Therefore, these CSFs cannot
be applied to BI systems without giving careful consideration to
the relevant contextual issues.
It appears that there is a macro-level pattern for interpreting
the CSFs related to such infrastructure-based projects. Both the
organisational and process dimensions are probably generic
and vary somewhat among BI systems and other infrastructural
systems implementation. But it is apparent that there is a new
understanding of factors associated with the technological
dimension due to the technical challenges that vary with the nature
of the infrastructure system. Nonetheless, this research suggests
that organisations are in a better position to successfully address
those CSFs through the business-orientation approach. That
is, without a clear business-driven objective, the BI initiatives
rarely produce substantial impact on business. As a result, the
implementation of a BI system has a much greater likelihood of
success when speci?c business needs are identi?ed at the outset,
and when those needs are used to direct the nature and scope of the
implementation effort. Therefore, this business orientation meta-
CSF should be regarded as the most critical factor in determining
the implementation success of BI systems.
This research has made a theoretical contribution to
our understanding of the CSFs that impact on BI systems
implementation. The literature review reveals relatively little
previous work on this subject. This study helps to ?ll the gap by
building the theory of the ways in which CSFs impact BI systems
implementation. This study represents the ?rst rigorously-
conducted enquiry which will develop our understanding of the
factors that affecting the implementation of BI systems. The
?ndings and outcomes extend current theory and allow ?rms to
identify and focus their scarce resources in those CSFs areas.
Besides that, academic researchers are often criticised for failing
to address issues of concern to practitioners. The collection
and analysis of empirical data in this study responds to those
criticisms and supplements the current limited understanding
of the factors that affect the successful implementation of BI
systems. The result of this work highlights those factors that need
to be addressed, and it also points out those that are not so critical.
Hence, it focuses attention on those important areas that might
otherwise be neglected or taken for granted but are signi?cant for
the implementation success.
Not only does this research contribute to the academic
literature on this topic but it bene?ts organisations in several ways
as well. First, large and complex organisations that are planning
to implement enterprise level BI systems will be better able to
identify those factors that will enhance the likelihood of success.
The ?ndings will help them to determine those factors on which
they should give particular attention to ensure that they receive
continuous management scrutiny. For senior management,
this research ?nding can certainly assist them by optimising
their scarce resources on those key areas that will improve the
implementation process. Also, management can concentrate on
monitoring, controlling and supporting only those critical areas.
The ?ndings with regard to the CSFs represent best practices
for ?rms that have successfully implemented BI systems. The
evidence that was revealed provides reference for BI stakeholders
that can increase the chances of implementation success.
REFERENCES
[1] Ang, J. & Teo, T. S. H. “Management Issues in Data
Warehousing: Insights from the Housing and Development
Board,” Journal of Decision Support Systems, 29(1), 2000,
11-20.
[2] Ariyachandra, T. & Watson, H. “Which Data Warehouse
Architecture Is Most Successful?” Business Intelligence
Journal, 11(1), 2006.
[3] Arnott, D. “Decision Support Systems Evolution:
Framework, Case Study and Research Agenda,” European
Journal of Information Systems, 13(4), 2004, 247-259.
[4] Arnott, D. & Pervan, G. “A Critical Analysis of Decision
Support Systems Research,” Journal of Information Tech-
nology, 20(2), 2005, 67-87.
[5] Bose, R. “Advanced Analytics: Opportunities and
Challenges,” Industrial Management & Data Systems,
109(2), 2009, 155-172.
[6] Burton, B., Geishecker, L., & Hostmann, B. “Organizational
Structure: Business Intelligence and Information Manage-
ment,” Gartner Research, 2006.
32 Journal of Computer Information Systems Spring 2010
[7] Delone, W., & McLean, E. “Information Systems Success:
The Quest for the Dependent Variable,” Journal of
Information System Research, 3(1), 1992, 60-95.
[8] Delone, W., & McLean, E. “The DeLone and McLean Model
of Information Systems Success: A Ten-Year Update,”
Journal of Management Information Systems, 19(4), 2003,
9-30.
[9] Friedman, T. “Gartner Says More Than 50 Percent of Data
Warehouse Projects Will Have Limited Acceptance or Will
Be Failures Through 2007,” 2005. Retrieved 21 Feb 2009,
fromhttp://www.gartner.com/it/page.jsp?id=492112
[10] Fuchs, G. “The Vital BI Maintenance Process”, in Business
Intelligence Implementation: Issues and Perspectives,” In
B. Sujatha (Ed), ICFAI University Press, Hyderabad, 2006,
116-123.
[11] Gartner Research, “Gartner EXP Worldwide Survey of
1,500 CIOs Shows 85 Percent of CIOs Expect ‘Signi?cant
Change’ Over Next Three Years,” 2008. Retrieved 21 Feb
2009, fromhttp://www.gartner.com/it/page.jsp?id=587309
[12] Gartner Research, “Gartner EXP Worldwide Survey of
More than 1,500 CIOs Shows IT Spending to Be Flat in
2009”, 2009. Retrieved 21 Feb 2009, fromhttp://www.
gartner.com/it/page.jsp?id=855612
[13] Hwang, M. & Xu, H. “A Structural Model of Data
Warehousing Success,” Journal of Computer Information
Systems, Fall 2008, 48-56
[14] Inmon, W.H. Building the Data Warehouse 4 Edn, John
Wiley & Sons, Indianapolis, 2005.
[15] Jagielska, I., Darke, P., & Zagari, G. “Business Intelligence
Systems for Decision Support: Concepts, Processes and
Practice,” Proceedings of the 7th International Conference
of the International Society for Decision Support Systems,
2003.
[16] Little, R.G. Identi?cation of Factors Affecting the Imple-
mentation of Data Warehousing, Thesis, Auburn University,
1998.
[17] Linstone, H. A., & Turoff, M. The Delphi Method:
Techniques and Applications. Addison-Wesley, Reading,
MA, 2002.
[18] Miles, M. & Huberman, A.M. Qualitative Data Analysis: An
Expanded Sourcebook, Sage, Thousand Oaks, CA. 1994.
[19] Moss, L. & Atre, S. Business Intelligence Roadmap: The
Complete Lifecycle for Decision-Support Applications.
Addison-Wesley, Boston, MA. 2003.
[20] Negash, S. “Business Intelligence,” Communications of the
Association for Information Systems, 13, 2004, 177-195.
[21] Olszak, C & Ziemba, E. “Approach to Building and
Implementing Business Intelligence Systems,” Interdis-
ciplinary Journal of Information, Knowledge, and Man-
agement, 2, 2007, 135-148.
[22] Reinschmidt, J. & Francoise, A. Business Intelligence
Certi?cation Guide, IBM International Technical Support
Organization, San Jose, CA, 2000.
[23] Shin, B. “An Exploratory Investigation of System Success
Factors in Data Warehousing,” Communications of the
Association for Information Systems, 6(4), 2003, 141-170.
[24] Stuart I., McCutcheon D. & Hand?eld R. “Effective
Case Research in Operations Management: A Process
Perspective,” Journal of Operations Management, 20(5),
2002, 419-433.
[25] Turban, E., Sharda, R., Aronson, J. & King, D. Business
Intelligence, Prentice Hall, New Jersey, 2007.
[26] Wadehra, A. “The ABCs of Master Data Management:
Architecture, Business Case, and Customer,” Business
Intelligence Journal, 12(1), 2007.
[27] Wang, H. & Wang S. “A Knowledge Management Approach
to Data Mining Process for Business Intelligence,” Industrial
Management & Data Systems, 108(5), 2008, 622-634.
[28] Watson, H., Abraham, D., Chen, D. “Data Warehousing
ROI: Justifying and Assessing a Data Warehouse,” Business
Intelligence Journal, 2004, 6-17.
[29] Watson, H., & Haley, B. “Managerial considerations,”
Communications of the ACM, 41(9), 1998, 32-37.
[30] Watson, H., Annino, D. A., Wixom, B. H. “Current Practices
in Data Warehousing,” Journal of Information Systems
Management, 18(1), 2001, 1-9.
[31] Watson, H., Fuller, C., & Ariyachandra, T. “Data Warehouse
Governance: Best Practices at Blue Cross and Blue Shield
of North Carolina,” Journal of Decision Support Systems,
38(3), 2004, 435-450.
[32] Wixom, B., & Watson, H. “An Empirical Investigation of
the Factors Affecting Data-Warehousing Success,” MIS
Quarterly, 25(1), 2001, 17-41.
[33] Yin, R. Case Study Research: Design and Methods 3 Edn,
Sage, Thousand Oaks, CA, 2003.
[34] Ziglio, E., & Adler, M. Gazing into the Oracle: The Delphi
Method and its Application to Social Policy and Public
Health, Jessica Kingsley Publishers, London, 1996.
doc_794498715.pdf