econstor

Wieder, Bernhard; Ossimitz, Maria-Luise; Chamoni, Peter

Article
The Impact of Business Intelligence Tools on
Performance: A User Satisfaction Paradox?
‫اﻟﻌﺑﺎرة اﻟﻣوھﻣﺔ ﻟﻠﺻﺣﺔ‬
‫ﻣﺗﻧﺎﻗﺿﺔ‬

International Journal of Economic Sciences and Applied Research

Provided in Cooperation with:

Eastern Macedonia and Thrace Institute of Technology (EMaTTech),
Kavala, Greece

Suggested Citation: Wieder, Bernhard; Ossimitz, Maria-Luise; Chamoni, Peter (2012) :

The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?,
International Journal of Economic Sciences and Applied Research, ISSN 1791-3373, Eastern
Macedonia and Thrace Institute of Technology, Kavala, Vol. 5, Iss. 3, pp. 7-32

This Version is available at:

http://hdl.handle.net/10419/114571

‫اﻟﻣﻔﺎرﻗﺔ )ﺑﺎﻹﻧﺟﻠﯾزﯾﺔ‬: Paradox)

‫ھﻲ ﺑﯾﺎن ﺑﺎﻟرﻏم ﻣن أن اﻟﺗﻌﻠﯾل ﯾﺑدو‬
‫ظﺎھرﯾﺎ ً ﻣﺑﻧﻲ ﻋﻠﻰ ﻣﻘدﻣﺎت وﺣﺟﺞ‬
‫ إﻻ أﻧﮫ ﯾؤدي إﻟﻰ ﺗﻧﺎﻗض‬،‫ﻣﻧطﻘﯾﺔ‬
‫اﻟﺑﯾﺎن ﻧﻔﺳﮫ أو ﯾؤدي إﻟﻰ اﺳﺗﻧﺗﺎج‬
‫ﻋﺑﺎرة ﻏﯾر ﻣﻧطﻘﯾﺔ‬
 International Journal of Economic Sciences and Applied Research 5 (3): 7-32

The Impact of Business Intelligence Tools on Performance:

A User Satisfaction Paradox?

Bernhard Wieder1, Maria-Luise Ossimitz2 and Peter Chamoni3

Abstract

This paper reports the findings of an exploratory, cross-sectional field study investigating
the factors that define and drive benefits associated with the deployment of dedicated BI tools.
BI tools are software products primarily designed and deployed to support this
analytical process (e.g. data warehouse software, data mining software, digital dashboards
applications).

Building upon DeLoneand McLean’s (1992; 2002; 2003) information systems success
model, we develop, test and refine a BI quality and performance model adapted for the specific
purpose, application, user group and technology of BI tools. The ultimate performance
predictors in this model are user satisfaction and the impact of BI tools on managerial decision
quality, both of which are determined by data quality.
Partial Least Square (PLS) modeling is used to analyze data collected in a survey
administered to IT executives of large Australian Stock Exchange (ASX) listed companies.
The results confirm some of the theoretical relationships established in – especially the
original – DeLone-McLean model in the specific context of BI. More importantly, the results
also confirm the important role of explicit BI management as antecedent of benefits derived from
BI tools, and the key impact of data quality on managerial decision making and organizational
performance.
However, the results also reveal a ‘user satisfaction paradox’: In contrast to the
predictions derived from the DeLone-McLean model, organizational performance is negatively
associated with user satisfaction with BI tools. Financial performance data collected for
ex-post verification of this unexpected result confirm this paradox. We discuss BI-specific
interpretations of these unexpected findings and provide avenues for future research.

Keywords: Business Intelligence (BI), information systems success, data quality, user
satisfaction, IT impact analysis

7
 The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?
1. Introduction
BI software describes software products primarily designed to support this
analytical process (e.g. data warehouse software, data mining software, digital
dashboards software), BI tools are BI software products deployed in an organization,
and a BI system is a collective of BI tools and related technologies, applications and
processes used in support of BI.

More recently, studies emerged investigating the determinants of information and

systems quality in the context of DW (Nelson et al., 2005), the effects of DW on decision
performance (Yong-Tae, 2006), the measurement of BI (Lönnqvist and Pirttimäki, 2006),
the effects of BI on performance (Elbashir et al., 2008), ,and the impact of BI on
organizational decisions (Davenport, 2010)

The purpose of this research is to integrate and extend the findings of previous DW/
BI research by developing, testing and refining an information systems success model for
the specific purpose, application, target group and technology of BI.

3
Definitions of ‘information system’ typically refer to interaction between people, procedures and
technology in the process of capturing, transmitting, storing, retrieving, manipulating and displaying
data and information for a specific purpose.
4
Watson (2010) provides an analysis of the development of DSS in the context of data warehousing
and Clark et al. (2007) provide a comprehensive literature analysis of research on management
support systems (MSS), including BI.

9
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

2. Theory development
BI is a broad concept of managing and providing data for improved (managerial) decision
making5. This implies that BI success or BI quality is to be measured around the quality or
quality increase of data provided and the quality or quality increase of decisions made in
an organization.

Data/Information Quality and Quality of Managerial Decision Making

maintaining and providing high quality data is more important concern in BI
systems than in other business information systems, in particular OLTP systems (e.g.
ERPS), which typically have a very large non-managerial user base and often provide
high levels of transaction automation and control.
Many attempts have been made to operationalize data or information quality
substantially7. Nelson et al. (2005) provided a comprehensive literature review, they
aggregate the large number of quality attributes into the following four dimensions of
information quality: Accuracy (intrinsic), completeness, currency and format (all
extrinsic). Their measurement model validation using PLS modeling reveals, however, that
currency does not load signifi cantly on information quality. They provide a possible
explanation for the insignifi cance of currency, but our alternative interpretation is in
increasingly real time data environments, currency has become a minor data quality
concern. We build on their work with minor variations (see measurement model below).
Data quality is not an aim in itself, but rather a means to the key aim of BI: Providing
(better) support for decision making, resulting in faster, better informed and more accurate
decisions. This leads to our first hypothesis:
H1: BI data quality impacts positively on managerial decision quality.

Good decisions are decisions which create competitive advantage, be it in the form
of entrepreneurial rents (Schumpeter, 1950; Rumelt, 1987), or a sustained competitive
position. The widely accepted short-term operationalization of competitive position is
performance relative to rival firms (Arend, 2003). Accordingly, we predict as follows:
H2: Managerial decision quality impacts positively on relative organisational
performance.

5
This notion is also reflected BI software vendor promises and selling lines.
6
While we are aware of the differences between data and information, research on those constructs
does not usually draw a clear line and research on the antecedents or determinants of data quality
and information quality overlap substantially (see Nelson et al. 2005, in particular their analysis of
prior literature). We therefore use the terms data and information as de facto synonyma for the
purpose of this research.
7
See Nelson et al. (2005), in particular their systematic analysis of prior literature.
10
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

User Satisfaction and BI System Use

In all versions of DeLone and McLean’s (1992; 2002; 2003) information systems
success model, user satisfaction and system use are key links between information quality
and individual impacts or net benefits respectively. Considering that BI systems are
discretionary ‘informational’ systems in a sense that they are not required for business
process execution or other forms of transaction processing, the particular importance of
addressing user satisfaction and (the relationship with)actual use of the system is obvious.
Further to that, BI systems are often deployed as alternatives to ‘islands of spreadsheets’,
with the latter often remaining in place as some form of shadow systems. We therefore
expect a large variation in BI system use across organizations, even if they deploy similar
BI solutions, and we concur with DeLone and McLean’s assessment of the important role
of IS use in terms of achieving benefits associated with the system. Finally, Cox’s (2010)
recent research confirms the positive association between frequent BI use and quality and
speed of decisions. We therefore predict as follows:

H3: The scope of use of a BI system impacts positively on managerial decision

quality.

While improving decision support is the main purpose of BI, there are also other
benefits associated with BI, including reductions in total cost of ownership (TOC),
efficiency and quality increases in information processing, improved customer satisfaction,
improving internal communication and collaboration (Hocevar and Jaklic, 2010; Imhoff
and White, 2010; Watson and Wixom, 2010). Those benefits can only be realized, if the BI
tools implemented are actually used. Accordingly, we predict a direct impact of BI use on
performance:

H4: The scope of use of a BI system impacts positively on organisational

performance.
In line with DeLone and McLean (1992; 2002; 2003) we argue that user satisfaction
with a system is likely to increase usage, even more so with discretionary systems.

H5: User satisfaction with a BI system impacts positively on the scope of use of
the system.

11
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Scope of BI System

As mentioned in the introduction, BI ‘is an umbrella term that is commonly used to

describe the technologies, applications and processes for gathering, storing, accessing and
analyzing data to help users make better decisions’ (Watson and Wixom, 2010). The range
of software products offered in support of BI is broad and varies significantly in terms
of purpose or role within a BI architecture, detailed functionality, functional scope and
level of sophistication. Examples of BI software include data warehouse (management)
software, extraction transformation and loading (ETL) tools, simple query tools, OLAP
engines, data mining software and visualization tools such as digital dashboards (Turban
and Volonino, 2011).
Many BI software products are either by functional design or by deployment subject
oriented, i.e. they have a functional focus (e.g. market analysis or sales forecasting).
Enterprise data warehouses, however, can potentially support a broad range of business
functions in an organization. Accordingly, BI systems deployed in organizations will have
a great level of variation in terms of functional scope, which will have a direct impact on
the scope of use of the BI system.

H6: The scope of a BI system impacts positively on the scope of use of the system

Quality of BI Management

The importance of proper management was already emphasized in early studies on

critical success factors of data warehouse projects (Wixom and Watson, 2001), and still
remains a critical dimension of BI maturity (TDWI-Research, 2008).
Wixom and Watson (2001) found that management support and resources help
to address organizational issues that arise during warehouse implementations, and that
adequate resources, user participation and highly-skilled project team members increase
the likelihood that warehousing projects finish on-time, on-budget and with the right
functionality. Standard development and implementation methodologies are also commonly
cited as critical success factors of BI projects (Hwang and Xu, 2008). Managing BI systems
‘scalable’ has been a major quality aspect of BI management from the earlier days of data
warehousing (Gardner, 1998) to the area of BI (Imhoff, 2005), and is still considered to be
an ongoing and future trend in BI (Watson, 2009).
Finally, TDWI (2008) emphasizes the importance of standards for developing,
testing, and deploying BI/DW functionality to be defined, documented, and implemented.

12
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

While the early literature focused on quality as aspects of DW/BI project management,
the more recent – especially practitioner oriented – literature increasingly deals with BI
management (and governance) as an ongoing process (BI management as a sub-function of
IT management).
Following from the above, we expect that high quality BI management has a positive
impact on various aspects of BI: Through end user involvement, timely completion of BI
projects and provision of adequate resources and support, we expect a positive impact
on user satisfaction and a steady increase in BI scope in an organization. Adherence to
standards and the provision of adequate resources and scalable solutions is expected to
result in higher levels of data quality, which is expected to indirectly contribute to user
satisfaction.

H7: BI management quality impacts positively on BI data quality.

H8: BI management quality impacts positively on the scope of the BI system.

H9a: BI management quality impacts positively on user satisfaction with a BI

system.

H9b: This relationship is mediated by data quality.

Figure 1 provides a graphical summary of our hypotheses (path-model).

Figure 1: Research Model

Data
Quality

BI User Decision
BI Use Quality
Management Satisfaction

BI Scope Performance

13
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

3. Research Design and Method

Sample Selection and Data Collection

A cross-sectional research design was employed with a survey administered to the

500 largest Australian Stock Exchange (ASX) listed companies in terms of capitalization.
Target respondents were the most senior IT managers (CIO or equivalent), because they
were considered best suited to answer questions about both their management domain and
firm performance. Contact details of the managers including email addresses were obtained
from a private data provider.
In June 2009, an email invitation was sent to the target respondents inviting them to
participate in the survey by completing a comprehensive web-based online questionnaire.
The initial invitation was followed up by an invitation letter accompanied by a hard-copy
of the survey. With many invitation emails not reaching their addressees (‘bounce backs’)
and an almost equal number of invitation letters being returned to sender because of the
addressees having left the company, a contact details review was performed in September
2009, and email and hard-copy invitations were sent out to the corrected contacts. Despite
this review and follow-up, 69 of the 500 firms could not be reached, reducing the effective
sample size to 431.
44 firms (10.21%) responded to the survey, but 11 had to be removed from the
sample, because they failed to either complete all questions in the questionnaire or meet
the minimum size criteria8 of $50 million AUD annual revenue and 50 full-time equivalent
(FTE) employees. A non-response bias was inherent to the study insofar as only firms
which deployed BI software (as defined above) were encouraged to participate. In the
absence of publicly available data on the use of BI software in the target group, the impact
of this exclusion cannot be determined.
Table 1 provides a breakdown of the 33 use able responses by industry and firm size.
The largest industry groups represented in the sample are mining and real estate with the
rest of the respondents representing a broad cross-sectional sample of Australia’s private
sector industry.

8
Some of the respondents of the top-500 ASX listed companies completed the survey for their
respective business units, and not all of them met the minimum size criteria for inclusion in the
survey.

14
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

Table 1: Respondents by Industry and Firm Size

Industry N %
Materials – Mining 5 15.2%
Real Estate 4 12.1%
Capital Goods 3 9.1%
Food, Beverage & Tobacco 3 9.1%
Transportation 3 9.1%
Diversified Financials – Banks 3 9.1%
Commercial & Professional Services 3 9.1%
Consumer Services 2 6.1%
Retailing and Wholesale 2 6.1%
Media 1 3.0%
Energy 1 3.0%
Insurance 1 3.0%
Pharmaceuticals, Biotechnology & Life Sciences 1 3.0%
Software & Services 1 3.0%
Total 33 100.0%

Annual Revenue
% Employees (FTE) %
(in Millions)
< 50 0.0% < 50 0.0%
50 < 100 9.1% 50 < 100 3.0%
100 < 500 36.4% 100 < 500 21.2%
500 < 2,500 21.2% 500 < 1,000 27.3%
2,500 < 10,000 24.2% 1,000 < 3,000 9.1%
10,000 < 50,000 9.1% 3,000 < 10,000 21.2%
> 50,000 0% > 10,000 18.2%

Table 2 depicts the positions/roles of the respondents. Two thirds of the respondents
were heads of IT, either in an explicit ‘CIO’ role or as heads of IT at the group or business
unit level. The other respondents were managers of IT in general or BI in particular.

Table 2: Respondents’ Positions/Roles

Position %
Head of IT – Group 33.3
Other IT Manager 24.2
CIO 21.2
Manager Business Intelligence 9.1
Head of IT Business Unit/Division 6.1
Other 6.1

15
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Measurement Model

BI Management Quality
In the absence of an established measurement for BI management quality, reflective
indicators were derived from the BI Maturity Model (BIMM), developed and used by the
TDWI (Chamoni and Gluchowski, 2004; TDWI-Research, 2008). After the exclusion
of some indicators used in this reference model but not loading on the construct of BI
Management Quality in our survey, the following four indicators were used for analysis:
(1) BI development standardization, (2) BI project management success (as evidenced by
BI projects being delivered in time and within budget), (3) BI resources (the availability of
resources in IT required for BI), and the scalability of BI solutions.
Respondents were asked to rate their firm’s performance in terms of achieving the
above-mentioned objectives on a five point Likert type scale (1 = not achieved at all; 5 =
‘fully achieved’).

Scope of BI
The measurement model for Scope of BI was also developed primarily based on
technical practitioner literatures on BI or data warehousing.
The first dimension of Scope of BI refers to the number of BI tools available in
an organisation. Respondents were asked to select or list commercial OLAP software,
querying and frontend reporting software, digital dashboards anddata mining software (see
Appendix) used in their organization. The count of software products deployed in each
organization was used to measure ‘BI tools available’.
The second dimension of Scope of BI was ‘BI functional scope’, which refers to
business functions or processes typically supported by BI solutions. The questionnaire items
were derived from practitioner literature combined with our own software functionality
analysis (see Appendix for items).We allowed for additional functions to be added as open
items. The count of business functions or processes supported by BI solutions in each
organization was used to measure BI functional scope.

Data Quality
The measurement model for data quality built on the extensive IS research on data
or information quality, in particular the fundamental research provided by Nelson et al.
(2005), who adapted general IS quality theory to the specific context of data warehousing.
We build on Nelson's et al. (2005) findings adopting elements of their broad definition
of completeness, but refine the concept insofar as we emphasize the importance of avoiding
information overload. The resulting dimension used in our study is adequacy of data
volume, with data relevance – which is included in their definition of data completeness –
measured separate in our study. We also adopt the dimension accuracy or correctness and
format (presentation) of data, but extend the concept of format by also explicitly addressing
transparency of data. The latter addition reflects concerns raised by traditional spreadsheet

16
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

users that data warehouse based BI tools are black boxes which lack transparency. Related
to those concerns is the question of trust in data, which has been raised in data quality
literature before and appears to be particularly relevant for BI.
Following exploratory factor analysis and initial PLS testing, the following items
scored significant loadings on the data quality construct used in our research: (1) Adequacy
of data volume9, (2) data relevance, (3) data transparency, and (4) trust in data. Accuracy/
correctness and format/presentation had to be excluded from the measurement model.
Respondents were asked to rate their firm’s performance in terms of achieving the
above-mentioned information-related objectives on a five point Likert type scale (1 = ‘not
achieved’ at all; 5 = ‘fully achieved’).

User Satisfaction
Considering the research method (survey) used in our study, we were unable to
measure user satisfaction directly at the user level, e.g. by interviewing or surveying a
sample of users within each organisation. Instead we asked our survey respondents
about their assessment of user satisfaction with the BI system. While this is a limitation
of our study, we argue that (most) senior IT managers in an organisation would have a
reasonably good understanding of how satisfied users are with the BI solutions deployed
in an organisation, even more so as BI solutions typically have a relatively small and more
senior user group than large scale operational systems such as ERPS.
Exploratory factor analysis and initial PLS testing revealed that the following four
(out of initially seven) items revealed highly significant loadings on the user satisfaction
construct: Users’ perceptions about (1)the effectiveness and efficiency of the BI system,
(2) the suitability/task relevance of information provided by the BI system, (3) the extent
to which the BI system meets user requirements and (4) general user satisfaction with BI
system.
Respondents were asked to rate the user satisfaction with the BI system on a five
point Likert type scale (1 = ‘very negative’; 3 = ‘neutral’; 5 = ‘very positive’).

BI System Use
In the absence of an established measurement model for BI system use, reflective
indicators were – once again – derived from the BI Maturity Model (BIMM), developed
and used by the TDWI (Chamoni and Gluchowski 2004, TDWI-Research 2008). Based
on this reference, two aspects of BI system use were captured: (a) the functional scope of
BI, and (b) level of sophistication of BI use (see Appendix). For the measurement of the
latter aspect, we distinguished between the following usage levels: Passive use, ad-hoc
reporting, OLAP use and analytics expert use. While we argue there is an implicit rank
in this measurement in terms of level of sophistication, we acknowledge that it does not
reflect an ordinal scale for statistical purposes. We therefore generated a separate score

9
‘Adequacy’ captures both the notion of having enough data and not experiencing data overload.

17
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

for each level of use across each function. Out of the resulting four diffusion scores, only
‘passive use’ and ‘ad-hoc reporting use’ were included in our analysis, because OLAP use
and analytic use were negligible across the sample and therefore did not load significantly
on our construct of BI system use.

Quality of Managerial Decision Making

Considering the research method (survey) used in our study, we were unable to
measure user satisfaction directly at the user level, e.g. by interviewing or surveying a
sample of users within each organisation. Instead we asked our survey respondents about
their assessment of user satisfaction with the BI system.
Like in the case of user satisfaction, the research method used only allowed us to
measure the impact of the deployment of BI solutions at a very aggregate level and only
indirectly by asking respondents about their perceptions about the said impact; another
limitation of our study.
The indicators used to measure this construct were derived from decision science
(Yong-Tae, 2006) and comprised five aspects of decision making quality, four of which
loaded significantly on our ‘impact’ construct: (1) Effectiveness of decision making, (2)
accuracy/correctness of decision making, (3) timeliness/speed of decision making, and
making rationale/informed decisions.
Respondents were asked to rate the impact of the BI system on the quality of
managerial decision making along the five aspects mentioned above on a five point Likert
type scale (1 = ‘very negative’; 3 = ‘neutral’; 5 = ‘very positive’).

Relative Performance (Competitive Advantage)

Respondents were asked to rate their firm’s performance relative to their main
competitors. One of the advantages of using relative measures it that they control for
differences in performance that are due to industry, environment, and strategy effects
(Govindarajan and Fisher, 1990; Garg et al., 2003).

Profitability, revenue growth and market share are well established indicators of
financial performance (e.g. Kaplan and Norton, 1996; Slater and Olson, 2000) and were
therefore adopted in our study. Following a balance scorecard approach (Kaplan and
Norton, 1996), leading performance indicator closest10 to financial performance were also
included in the form of relative customer satisfaction and customer loyalty.
Considering the mix of leading and lagging performance indicators used, the
measurement model was also specified as reflective, following Tippins and Sohi (2003)
and Johansson and Yip (1994).

10
The other antecedents of firm performance (business process performance and learning and
growth) were not included acknowledging the static nature of our research.

18
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

Financial Performance (ROA)

To overcome some of the weaknesses inherent to perception based measures of firm
performance and to increase the reliability of our performance measures, we collected
publicly available financial data of the firms in our sample to determine the return-on-assets
(ROA) in the financial year prior to the completion of the survey.
This additional financial performance variable was not considered a main testing
variable for two reasons: First, we could only derive the ROA at the company level (as
listed at the ASX), but not at the business unit level (some of the respondents referred to).
The second reason is that ROA is a directly observable and well established performance
construct in itself, and therefore strictly speaking not a latent variable. In spite of these
limitations, the inclusion of an ‘ROA-based firm performance indicator’ allowed us
to establish a link between perceived performance and traditional archival financial
performance indicators.

Partial Least Square Modeling (PLS)

Structural equation models (SEM) are strongly suited to testing both theories and
measurement models (Bagozzi, 1980). The partial least squares (PLS) procedure was
used, because it is most appropriate for the non-normal datasets and small sample sizes in
the current research (Wold, 1982; Chin, 1998). PLS uses very general soft distributional
assumptions and non-parametric prediction-orientated model evaluation measures (Wold,
1982; Chin, 1998).
The next section herein evaluates the measurement models, and then the following
section assesses the structural model to determine the results. Chin and Dibbern’s (2010)
guidelines for reporting on PLS analyses were followed11.

Evaluation of the Measurement Model

The adequacy of reflective measurement models is examined via; (1) individual
item reliability, (2) convergent validity, and (3) discriminant validity (Chin, 1998; Hulland,
1999). First, individual item reliability is assessed by examining the item’s loading on
its construct as opposed to the other latent variable constructs in the model. As shown in
Table 3, all construct-specific loadings are above 0.60, with many in the 0.80 and 0.90
‘high’ range (Chin, 1998; Hulland, 1999). Table 3 also confirms that each indicator’s load
is highest for the relevant latent variable construct.

11
SmartPLS version 2.00 M3 was used (Ringle et al., 2005).

19
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Table 3: Measurement Model – Discriminant Validity

BI Data User Decision Perf. Perf.
BI Mgt Scope Quality Satisf. BI Use Quality Ind. ROA
BI development standardisation 0.8553 0.2358 0.7238 0.6212 0.0392 0.2558 -0.2210 0.0840
BI projects on time 0.8668 0.3796 0.6673 0.5111 0.1356 0.2717 -0.2301 -0.3579
BI resources 0.7964 0.3514 0.5586 0.5354 0.0000 0.3485 -0.1179 -0.1895
BI scalability 0.9209 0.2831 0.7115 0.6471 0.0453 0.2541 -0.2772 -0.1754
BI tools available 0.2027 0.9140 0.0616 -0.2077 0.6837 0.1184 0.3353 0.0225
BI functional scope 0.4607 0.8912 0.2999 0.2842 0.4538 0.3219 0.1809 -0.0827
Data volume adequacy 0.7124 0.2129 0.9138 0.6512 0.0418 0.3567 -0.1556 0.1333
Data relevance 0.6528 0.1499 0.8916 0.6528 0.0268 0.5137 -0.0976 0.0225
Data transparency 0.6714 0.1145 0.8875 0.5902 -0.0610 0.4626 -0.1828 -0.0392
Data trusted 0.7517 0.2134 0.9021 0.5897 0.0655 0.5047 -0.0539 -0.1337
Effectiveness & efficiency of
0.6940 0.0442 0.6469 0.9486 -0.1788 0.3423 -0.4717 -0.2979
BI system
Suitability/task relevance of
0.5007 0.0113 0.6165 0.9265 -0.1327 0.3350 -0.3764 -0.2721
BI info.
BI system meeting user
0.6583 0.0481 0.6670 0.9395 -0.0878 0.3754 -0.3713 -0.1287
requirements
General end-user satisfaction 0.6674 -0.0020 0.6676 0.9512 -0.2606 0.3052 -0.4405 -0.2634
Scope of passive use 0.0356 0.4896 -0.0468 -0.3134 0.8573 -0.0204 0.2287 0.0467
Scope of ad-hoc reporting 0.0749 0.6136 0.0784 -0.0105 0.8829 0.1491 0.2589 0.0016
Decision effectiveness 0.3459 0.3532 0.4996 0.4232 0.1085 0.8432 0.0787 -0.0465
Accuracy/correctness of dec.
0.2914 0.1622 0.4389 0.1768 0.0796 0.7724 0.2193 -0.0693
making
Timeliness/speed of decision
0.2285 0.0711 0.3618 0.3221 -0.1060 0.8099 0.1364 -0.2080
making
Making rationale/informed
0.1403 0.1290 0.3067 0.2249 0.1453 0.7818 0.2210 0.0212
decisions
Customer loyalty -0.1736 0.0294 -0.2169 -0.3481 0.2243 -0.0305 0.6053 0.1063
Market share -0.2129 0.1892 -0.2327 -0.4102 0.1054 -0.0199 0.6094 0.4983
Profitability -0.2822 0.1747 -0.1539 -0.4411 0.3443 0.1801 0.8278 0.3272
Quality management -0.0050 0.3503 0.1670 -0.0759 0.2999 0.3909 0.7272 0.1865
Revenue growth -0.2196 0.2206 -0.1074 -0.3308 0.0082 0.1105 0.8106 0.3793
ROA -0.1803 -0.0301 -0.0063 -0.2552 0.0265 -0.0902 0.4447 1.0000

Table 4 reports the measurement indicators’ means and standard deviations along
with other standard measurement model quality indicators, as well as the bootstrapped
error terms, t-statistics and significance levels.
All composite reliability measures (0.84 to 0.97) comfortably exceed the
recommended threshold of 0.70 (Fornell and Larcker 1981, Chin 1998). Cronbach’s Alphas
() are slightly lower, but still greater or very close to 0.70 (Nunnally 1978, Chin 1998),
indicating strong reliability of the measurement model. All average variances extracted
(AVE) are higher than 0.50 (Fornell and Larcker, 1981), ranging from 0.52 to 0.89. Hence,
there is no concern with convergent validity either.

20
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

Table 4: Measurement Model – Descriptive Statistics and Quality Indicators

Composite Cronbach's Bootstrapping
Constructs and Indicators: Mean Std. dev. Loadings reliability Alpha AVE
rc  SE t-statistic

A. BI Management 0.92 0.88 0.74

BI development standardization 3.03 1.287 0.86*** 0.05 15.66
BI projects on time 2.94 1.298 0.87*** 0.04 22.10
BI resources 3.00 1.000 0.80*** 0.10 8.03
BI scalability 3.36 1.168 0.92*** 0.03 27.92
B. Scope of BI 0.90 0.77 0.81
BI tools available 3.82 1.911 0.91*** 0.03 36.00
BI functional scope 4.91 2.708 0.89*** 0.09 10.18
C. Data Quality 0.94 0.92 0.81
Data volume adequacy 3.42 0.936 0.91*** 0.04 22.71
Data relevance 3.42 0.867 0.89*** 0.05 19.13
Data transparency 3.39 0.966 0.89*** 0.06 14.66
Data trusted 3.36 0.929 0.90*** 0.04 22.21
D. User Satisfaction 0.97 0.96 0.89
Effectiveness & efficiency of
3.27 1.008 0.95*** 0.02 45.98
BI system
Suitability/task relevance of
3.64 0.859 0.93*** 0.03 35.92
BI info.
BI system meeting user
3.33 1.051 0.94*** 0.03 31.22
requirements
General end-user satisfaction
3.21 0.927 0.95*** 0.02 52.28
with BI system
E. BI Use 0.86 0.68 0.76
Scope of passive use 2.88 2.147 0.86*** 0.16 5.50
Scope of ad-hoc reporting 2.36 1.765 0.88*** 0.05 18.16
F. Decision Quality 0.88 0.82 0.64
Decision effectiveness 3.70 0.684 0.84*** 0.13 6.67
Accuracy/correctness of
3.53 0.671 0.77*** 0.14 5.67
decision making
Timeliness/speed of decision
3.67 0.777 0.81*** 0.15 5.31
making
Making rationale/informed
3.52 0.667 0.78*** 0.19 4.09
decisions
G1. Performance (indicators) 0.84 0.77 0.52
Customer loyalty 3.73 0.839 0.61*** 0.19 3.11
Market share 3.48 0.972 0.61** 0.24 2.52
Profitability 3.79 0.893 0.83*** 0.12 6.67
Quality management 3.76 1.062 0.73*** 0.21 3.48
Revenue growth 3.79 0.857 0.81*** 0.11 7.43
G2. Performance (ROA) 1.00 1.00 1.00
ROA 0.051 0.0458 1.00 - -

*** significant at p < 0.01; ** significant at p < 0.05 (two-tailed)

21
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Table 4 also reports on the bootstrapping results (SE, t-statistic and p-values) for the
indicator variables. With the exception of performance indicator ‘market share’, which is
significant at p < 0.05, all other indicator loadings are highly significant at p < 0.01.
As for the assessing of discriminant validity, Chin (1998) outlined two procedures
for: (1) cross-loadings (see Table 3 above) and (2) the AVE-PHI matrix. The diagonal
elements in the Table 5 show the square roots of the AVE of each construct, whereas the
off-diagonal elements show the PHI matrix of latent variable (LV) correlations. The cross-
loading test requirements are fully met: No indicator has a higher correlation on a LV other
than the one it is intended to measure, and each block of indicators does not load higher on
its respective LV than indicators for other LVs (Fornell and Larcker, 1981; Chin, 1998)12.

Table 5: Measurement Model: Discriminant Validity

A B C D E F G1 G2
A. BI Management 0.861
B. Scope of BI 0.359 0.903
C. Data Quality 0.776 0.193 0.899
D. User Satisfaction 0.675 0.027 0.691 0.941
E. BI Use 0.065 0.637 0.022 -0.178 0.870
F. Decision Quality 0.325 0.237 0.512 0.360 0.078 0.802
G1.Performance (indic.) -0.249 0.291 -0.135 -0.443 0.281 0.201 0.722
G2.Performance (ROA) -0.180 -0.030 -0.006 -0.255 0.027 -0.090 0.445 1.000

An interesting detail shown in Table 5 is the negative correlations, in particular

between the performance constructs and BI management, data quality and user satisfaction.
This observation will be further explored below.

Evaluation of the Structural Model

The results of the structural model are summarised in Table 6 and Figure 2.
The model provides strong support for hypotheses 1, 6, 7 and 8, and some (week)
support for hypotheses 4, 9a and 9b. Hypotheses 2, 3 and 5 were rejected, and most notably,
the relationship between user satisfaction and BI use was negative and significant. Possible
explanations and implications are discussed below.
The amount of variance in the endogenous constructs explained by the model (R2) is
indicative of the predictive power of the exogenous latent variables. The explained variance
of most of the variables was substantial to moderate (Chin, 1998). As expected, BI related
variables can only explain a small percentage of (firm) performance.

12
The reporting of the PLS modeling results follows Chin’s (2010) guidelines.

22
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

Table 6: Structural Model Results

Bootstrap
Hypothesis/Path Analysis: Coefficient t-statistic
H1: Data Quality  Decision Quality 0.511*** 3.054
H2: Decision Quality  Performance (indicators) 0.181 0.753
H3: BI Use  Decision Quality 0.067 0.456
H4: BI Use  Performance (indicators) 0.267* 1.433
H5: User Satisfaction  BI Use -0.196** (!) 1.845
H6: Scope of BI  BI Use 0.642*** 6.431
H7: BI Management  Data Quality 0.776*** 9.961
H8: BI Management  Scope of BI 0.359** 2.134
H9a: BI Management  User Satisfaction 0.348* 1.483
H9b: BI Management  Data Quality  User Satisf. 0.327** (Table 7)
Other paths:
Data Quality  User Satisfaction 0.421** 1.868
Performance ind.  Performance (ROA) 0.445*** 2.728
R-squares:
 Scope of BI 12.9%  Decision Quality 26.7%
 Data Quality 60.2%  Performance (indicators) 11.1%
 User Satisfaction 52.5%  Performance (ROA) 19.8%
 BI Use 44.4%
*** significant at p < 0.01; ** significant at p < 0.05; * significant at p < 0.10 (one-tailed)

Figure 2: Structural Model Test Results

Data Quality
(60.2%)

0.511***
0.776*** 0.421**

0.348* - 0.196** Decision

BI User Satisfaction BI Use
0.067 Quality
Management (52.5%) (44.4%)
(26.7%)
?

0.181
0.359**
0.642***
BI Scope 0.267*
(12.9%) Performance
Indicators
(11.1%)


.445***
p < 0.01
p<0.05 Perf. (ROA)
(19.8%)
p < 0.10
p<0.05mediating

23
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Analysis of Mediation and Moderation

To test for mediating effects of data quality on the relationship of BI management
and user satisfaction, we used the Hertel et al. (2008) multivariate adaptation of the Bollen
and Stine (1990) bootstrap percentiles approach. Shrout and Bolger (2002) demonstrate
and prove the superiority of Bollen and Stine’s (1990) bootstrapping approach over the
conventional mediation test statistics, and Hertel et al. (2008) adapted it innovatively to
PLS. The essence of Bollen and Stine’s bootstrapping approach is that the distribution of
the mediation scores (a × b) (Baron and Kenny 1986) is bootstrapped and that the resulting
scores are examined to determine the (α/2) × 100% and (1 – α/2) × 100% percentiles of the
distribution (α = confidence interval)13. If both percentile scores are either below or above
zero, a significant deviation from the expected distribution within the confidence interval is
confirmed.
Adapted to our model, we determine the product of the bootstrapping coefficients
of the paths BI Management  Data Quality and Data Quality  User Satisfaction and
determine the percentiles for α = 5% and α = 10%. The results are shown in Table 7.

Table 7: Mediation of Effect of BI Management on User Satisfaction

Effect Estimate Bootstrap Percentile (one-tailed)
95% 90%
Upper Lower Upper Lower
Effect on User Satisfaction
Direct via BI Management 0.348 0.707 -0.046 0.628 0.010
Indirect via Data Quality (a × b) 0.327 -0.014 -0.265 -0.043 -0.248

None of the other potential mediation paths in the model showed significant
mediation.
We also tested for a potential moderating effect of BI use on decision quality, but
found no such effect.

4. Discussion, Limitations and Outlook

Discussion

The purpose of this research was to integrate and extend the findings of previous
DW/BI research by developing, testing and refining an information systems success
model for the specific purpose, application, target group and technology of BI. Many of
the predicted relationships in our model were confirmed. For others we could not find
empirical support and contrary to established theory (DeLone and McLean, 1992; 2002;

13
For one-tailed analysis, the formula is (α) × 100% and (1 –α) × 100%.

24
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

2003), user satisfaction with BI systems was negatively associated with the scope of use of
BI systems. The following sections discuss both the expected and unexpected findings.

Confirmed Predictions
Our results confirm the quality of managerial decision making is strongly influenced
by the quality of data available in BI systems, and that the quality of managing BI within an
organization is an important antecedent of data quality and therefore also decision making quality
(total effect: 0.403). The findings substantiate the many calls for data quality management
initiatives expressed in the practitioner literature (e.g. Swartz, 2007; Sandler, 2008).
We also confirmed the expected strong relationship between the scope of BI tools
available and the actual use of BI, but most importantly, broader use of BI tools appears
to be positively – although weekly – associated with performance. The strong total effect
of BI scope on performance in the initial model (0.179) was confirmed in sensitivity
analysis which revealed a significant direct relationship between BI scope and performance
(0.272**). The strong relationship between the quality of BI management and BI scope
reinforces the importance of properly managing BI to achieve tangible benefits.
Better BI management also leads to higher user satisfaction, both directly and
mediated by data quality, but considering the controversial role of user satisfaction in the
model, there is doubt about the implications of this relationship.

Unconfirmed Predictions
The expected positive relationship between managerial decision quality and
performance remained unconfirmed. Possible explanations include a time lag between
managerial decisions and performance, the dominance of other performance drivers not
included in the model and limitations in the measurement of managerial decision quality.

Unexpected Findings
As an exogenous variable in the model, user satisfaction – and to a lesser extend BI
system use – not only failed to meet the expectations (‘no findings’), but had a significant
negative association with BI system use and a negative association with decision quality
and firm performance, and ex-post modeling also revealed a significant direct negative
association with both decision quality and performance.
While we acknowledge that the object of investigation in the Delone-McLean model
(1992) is the individual rather than the firm, the results are still surprising, even more so
as a large range of ex-post modeling and testing confirmed the relationships revealed in
our initial analysis. Limitations in the measurement model of user satisfaction may have
contributed to deviations from the expected findings, but could not fully explain this ‘user
satisfaction paradox’.
We are not aware of any established theory capable of explaining this paradox
directly. In search for our own explanation of the phenomenon, we arrived at the following
potential explanation:

25
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Frequent and advanced (business) users of IS are more likely to explore the
‘boundaries’ of systems, ask more challenging questions, are more likely to detect errors,
and are therefore more likely to be dissatisfied with the system and challenging for the IT
department than ‘basic’ users. BI systems are typically configured to provide relatively
easy to use or fully automated standard reports, but in order to explore the real potential
of these systems, ad-hoc reporting skills, advanced analytical skills and even configuration
skills are required. Moving beyond the pre-configured standard functionality is likely to
be associated with frustration about lack of user friendliness of the system, capabilities of
the system and lack of knowledge about the system logic, functionality and the underlying
models.
On the other hand, users who do not go beyond the base functionality of the system
and who do not ask critical questions are more likely to be ‘happy users’. But are ‘happy’
users ‘good’ users?
In the BI context, most likely they are not. More likely, they are evidence of lack
of (adequate) system use or lack of BI ‘mentality’ of BI culture, and potentially a leading
indicator of lack of performance.

Limitations and Outlook

Some of the limitations of our research have already been addressed above: The
small sample size, simplified measurement of user satisfaction and managerial decision
quality and reference to theory which evolved from individual user experiences with IS
rather than organizational experiences.
However, many of our predictions were confirmed, and the unexpected findings
provide a wide avenue for future research.

5. References

Arend, R. J., 2003, ‘Revisiting the logical and research considerations of competitive
advantage’, Strategic Management Journal, 24, 3, pp. 279-284.
Baars, H. and Kemper, H.-G., 2008, ‘Management support with structured and unstructured
data - an integrated business intelligence framework’, Information Systems Management,
25, 2, pp. 132-148.
Bagozzi, R. P., 1980, Causal Methods in Marketing, New York, John Wiley and Sons.
Baron, R. M. and Kenny, D. A., 1986, ‘The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations’,
Journal of Personality and Social Psychology, 51, pp. 1173-1182.
Bollen, K. A. and Stine, R., 1990, ‘Direct and Indirect Effects: Classical and Bootstrap
Estimates of Variability’, Sociological Methodology, 20, pp. 115-140.
Chamoni, P. and Gluchowski, P., 2004, ‘Integrationstrends bei Business-Intelligence-
Systemen’, Wirtschaftsinformatik, 46, 2, pp. 119-128.

26
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

Chen, L.-d., Soliman, K. S., Mao, E. and Frolick, M. N., 2000, ‘Measuring user satisfaction
with data warehouses: an exploratory study’, Information & Management, 37, 3, pp.
103-110.
Chin, W. W., 1998, ‘The Partial Least Squares Approach to Structural Equation Modeling’,
Modern Methods for Business Research, G. A. Marcoulides, Ed.^Eds., Mahwah, NJ,
Lawrence Erlbaum Associates, pp. 195-336.
Chin, W. W. and Dibbern, J., 2010, ‘An Introduction to a permutation based procedure for
Multi-Group PLS Analysis: Results of Tests of differences on simulated data an a cross
cultural analysis of the sourcing of Information System Services between Germnay and
the USA’, Handbook of Partial Least Square Concepts, Methods and Applications, V.
E. Vinzi, W. W. Chin, J. Henseler and H. Wang, Ed.^Eds., Berlin Heidelberg, Springer-
Verlag.
Clark, J., T.D, Jones, M. C. and Armstrong, C. P., 2007, ‘The dynamic structure of
management support systems: theory development, research focus and direction’, MIS
Quarterly, 31, 3, pp. 579-615.
Cox, C., 2010, ‘Balancing decision speed and decision quality: Assessing the impact of
Business Intelligence Systems in high velocity environments’, Faculty of the College of
Business Administration, United States - California, TUI University, Ph.D.
Davenport, T. H., 2010, ‘Business Intelligence and Organizational Decisions’, International
Journal of Business Intelligence Research, 1, 1, pp. 1-12.
DeLone, W. H. and McLean, E. R., 1992, ‘Information systems success: The quest for the
dependent variable’, Information Systems Research, 3, 1, pp. 60-95.
DeLone, W. H. and McLean, E. R., 2002, ‘Information systems success revisited’,
System Sciences, 2002. HICSS. Proceedings of the 35th Annual Hawaii International
Conference on System Sciences - 2002.
DeLone, W. H. and McLean, E. R., 2003, ‘The DeLone and McLean Model of Information
Systems Success: A Ten-Year Update’, Journal of Management Information Systems,
19, 4, pp. 9-30.
Dodson, G., Arnott, D. and Pervan, G., 2008, ‘The use of Business Intelligence Systems in
Australia’, ACIS 2008 Proceedings, Christchurch, New Zealand.
Elbashir, M. Z., Collier, P. A. and Davern, M. J., 2008, ‘Measuring the effects of business
intelligence systems: The relationship between business process and organizational
performance’, International Journal of Accounting Information Systems, 9, 3, pp. 135-
153.
Ellis, D., 2004, ‘Data Mining and Business Intelligence: Where will it lead us?’, Infotech
Update, 13, 6, pp 1-3.
Foley, E. and Manon, G., 2010, ‘What is Business Intelligence?’, International Journal of
Business Intelligence Research, 1, 4, pp. 1-28.
Fornell, C. and Larcker, D. F., 1981, ‘Evaluating Structural Equation Models with
Unobservable Variables and Measurement Error’, Journal of Marketing Research, 18,
1, pp. 39-50.

27
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Foster, S., Hawking, P. and Stein, A., 2005, ‘Business intelligence solution evolution:
adoption and use’, Business Intelligence Journal, 10, 4, pp. 44-54.
Gardner, S. R., 1998, ‘Building the Data Warehouse’, Association for Computing Machinery.
Communications of the ACM, 41, 9, pp. 52-60.
Garg, V. K., Walters, B. A. and Priem, R. L., 2003, ‘Chief executive scanning emphases,
environmental dynamism, and manufacturing performance’, Strategic Management
Journal, 24, pp. 725-744.
Gonzales, M. L., 2011, ‘Success factors for business intelligence and data warehousing
maturity and competitive advantage’, Business Intelligence Journal, 16, 1, pp. 22-29.
Govindarajan, V. J. and Fisher, J., 1990, ‘Strategy, control systems and resource sharing:
effects on business-unit performance’, Academy of Management Journal, 33, pp. 259-
285.
Hertel, G., Schroer, J., Batinic, B. and Naumann, S., 2008, ‘Do Shy People Prefer to Send
E-Mail?’, Social Psychology, 39, 4, pp. 231-243.
Hocevar, B. and Jaklic, J., 2010, ‘Assessing benefits of business intelligence systems - a
case study’, Management, 15, 1, pp. 87-120.
Hugh, J. W., Dorothea, L. A., Daniel, C., David, P. and Dominic, T., 2004, ‘Data warehousing
ROI: justifying and assessing a data warehouse’, Business Intelligence Journal, 9, 2, pp.
6-17.
Hulland, J., 1999, ‘The use of partial least square (PLS) in strategic management research:
a review of four recent studies’, Strategic Management Journal, 20, 2, pp. 195-204.
Hwang, H. G., Ku, C.-Y., Yen, D. C. and Cheng, C.-C., 2004, ‘Critical factors influencing
the adoption of data warehouse technology: a study of the banking industry in Taiwan’,
Decision Support Systems, 37, 1, pp. 1-21.
Hwang, M. I. and Xu, H., 2008, ‘A Structural Model of Data Warehousing Success’,
Journal of Computer Information Systems, Iss Fall 2008, pp. 48-56.
Imhoff, C., 2005, ‘What Do Customers Really Want?’, DM Review, 15, 5, pp. 12-68.
Imhoff, C. and White, C., 2010, ‘Business Intelligence and Collaboration: A Natural
Marriage’, Business Intelligence Journal, 15, 3, pp. 44-48.
Inmon, W. H., 2000, ‘The data warehouse environment: quantifying cost justification and
Return on Investment’, Microsoft Corporation and Billinmon.com llc.
Inmon, W. H., 2004, ‘The logical data warehouse: delving into the mysteries of the logical
and physical worlds’, DM Review Online, June 2004.
Johansson, J. K. and Yip, G. S., 1994, ‘Exploiting globalization potential: U.S. and Japanese
strategies’, Strategic Management Journal, 15, pp. 579-601.
Kaplan, R. S. and Norton, D. P., 1996, The Balanced Scorecard, Boston, MA.
Lönnqvist, A. and Pirttimäki, V., 2006, ‘The measurement of business intelligence’,
Information Systems Management, 23, 1, pp. 32-40.
Luhn, H. P., 1958, ‘A Business Intelligence System’, IBM Journal, 2, 4, pp. 314-319.
March, S. T. and Hevner, A. R., 2007, ‘Integrated decision support systems: A data
warehousing perspective’, Decision Support Systems, 43, 3, pp. 1031-1043.

28
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

Nelson, R. R., Todd, P. A. and Wixom, B. H., 2005, ‘Antecedents of Information and
System Quality: An Empirical Examination within the context of data warehousing’,
Journal of Management Information Systems, 21, 4, pp. 199-235.
Nunnally, J. C., 1978, Psychometric Theory, New York, McGraw-Hill.
Petter, S. and McLean, E. R., 2009, ‘A meta-analytic assessment of the DeLone and McLean
IS success model: An examination of IS success at the individual level’, Information &
Management, 46, 3, pp. 159-166.
Power, D. J., 2003, ‘A Brief History of Decision Support Systems’, Retrieved May 31,
2003, from DSSResources.COM/history/dsshistory2.8.html.
Rekom, P. E., 2000, Data Warehouse: A Case study of the factors effecting user satisfaction.
Faculty of the Rossier School of Education. California, University of Southern California.
Doctor of Education: 287.
Research, G., 2011, Gartner Quarterly IT Spending Forecast.
Ringle, C. M., Wende, S. and Will, S., 2005, SmartPLS 2.0 (M3), Hamburg.
Rumelt, R. P., 1987, ‘Theory, strategy and entreprenuership’, The competitive challenge, D.
J. Teece, Ed.^Eds., Cambridge, Ballinger, pp. 137-158.
Sammon, D. and Adam, F., 2005, ‘Towards a model of organisational prerequisites for
enterprise-wide systems integration: Examining ERP and data warehousing’, Journal of
Enterprise Information Management, 18, 4, pp. 458.
Sammon, D., Adam, F. and Carton, F., 2003, ‘Benefit realisation through ERP: The re-
emergence of data warehousing’, Electronic Journal of Information Systems Evaluation,
6, 2, pp. 155-163
Sandler, D., 2008, ‘Four elements of successful data quality programs’, Business Intelligence
Journal, 13, 4, pp. 22-29.
Schumpeter, J. A., 1950, Capitalism, socialism and democracy, New York, Harper and
Row.
Shankaranarayanan, G. and Even, A., 2004, ‘Managing metadata in data warehouses:
Pitfalls and possibilities’, Communications of AIS, 2004, 14, pp. 247-274.
Shrout, P. E. and Bolger, N., 2002, ‘Mediation in Experimental an Nonexperimental Studies:
New Procedures and Recommendations’, Psychological Methods, 7, 4, pp. 422-445.
Slater, S. F. and Olson, E. M., 2000, ‘Strategy type and performance: the influence of sales
force management’, Strategic Management Journal, 21, pp. 813-829.
Solomon, M. D., 2005, ‘Ensuring a successful data warehouse initiative’, Information
Systems Management, 22, 1, pp. 26-36.
Swartz, N., 2007, ‘Gartner Warns Firms of 'Dirty Data'’, Information Management Journal,
41, 3, pp. 6-6.
TDWI-Research, 2008, 2008 TDWI BI Benchmark Report - Organizational and
Performance Metrics for BI Teams.
Tippins, M. J. and Sohi, R., 2003, ‘IT competency and firm performance: is organizational
learning a missing link?’, Strategic Management Journal, 24, pp. 745-761.

29
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

Tseng, F. S. C. and Chou, A. Y. H., 2006, ‘The concept of document warehousing for
multi-dimensional modeling of textual-based business intelligence’, Decision Support
Systems, 42, 2, pp. 727-744.
Turban, E. and Volonino, L., 2011, Information Technology for Management, Wiley.
Watson, H. J., 2001, ‘Current practices in data warehousing’, Information Systems
Management, 18, 1, pp. 47-55.
Watson, H. J., 2009, ‘Tutorial: Business Intelligence-Past, Present and Future’,
Communications of the Association for Information Systems, 25, pp. 487-510.
Watson, H. J., 2010, ‘The More Things Change, The More They Remain the Same’,
Business Intelligence Journal, 15, 3, pp. 4-6.
Watson, H. J., Goodhue, D. L. and Wixom, B. H., 2002, ‘The benefits of data warehousing:
why some organizations realize exceptional payoffs’, Information & Management, 39,
6, pp. 491-502.
Watson, H. J. and Wixom, B. H., 2010, ‘The BI-based Organization’, International Journal
of Business Intelligence Research, 1, 1, pp. 13-28.
Williams, S., 2004, ‘Delivering strategic business value’, Strategic Finance, 86, 2, pp. 41-
48.
Wixom, B. H., 2004, ‘Business Intelligence Software for the classroom: microstrategy
resource on the Teradata University Network’, Communications of AIS, 2004, 14, pp.
234-246.
Wixom, B. H., Watson, H., Reynolds, A. and Hoffer, J., 2008, ‘Continental Airlines
continues to soar with Business Intelligence’, Information Systems Management, 25,
2, pp. 102-112.
Wixom, B. H. and Watson, H. J., 2001, ‘An empirical investigation of the factors affecting
data warehousing success’, MIS Quarterly, 25, 1, pp. 17-41.
Wold, H., 1982, ‘Soft modeling: the basic design and some extensions’, Systems under
indirect observations: causality, structure, prediction, K. G. Joreskog and H. Wold,
Ed.^Eds., Amsterdam, North-Hollamd.1-54.
Yong-Tae, P., 2006, ‘An empirical investigation of the effects of data warehousing on
decision performance’, Information & Management, 43, 1, pp. 51-61.
Zeng, Y. C., Roger, H. L. and Yen, D. C., 2003, ‘Enterprise integration with advanced
information technologies: ERP and data warehousing’, Information Management &
Computer Security, 11, pp. 115.

30
The Impact of Business Intelligence Tools on Performance: A User Satisfaction Paradox?

Appendix: Measurement Details (Extract)

Scope of BI:

a) BI tools available:
‘What types of BI products/tools are in use in your company? (Multiple answer option)’
OLAP Tools: Data Mining Tools
 Cognos (now IBM)  SAS - Enterprise Miner
 Hyperion Solutions (now Oracle)  SPSS - Clementine, AnswerTree, Neural
 Microsoft Connect.
 SAP Business Objects  IBM - Intelligent Miner
 Microstrategy  Oracle - Darwin
 SAP BI  CSI - Advisor Toolkit
 Cartesis SA  Angoss Software - Knowledge Studio/Seeker
 Applix  Trajecta - dbProphet
 Oracle (other than Hyperion)  Partek
 Infor  Megaputer Intelligence - PolyAnalyst
 Others (list here):  Silicon Graphics - MineSet
 Clopinet
 Unica
Querying and Frontend Reporting Tools
 Eudaptics Software - Viscovery
 List here:
 HYPERparallel - Discovery
 Others (list here):
Digital Dashboards
 List here:

b) BI Functional Scope:
Which basic business functions or processes are directly supported by your BI solution? (Multiple
answer option)
 Regular financial/tax reporting (external reporting)Suppliers
 Assurance and special compliance support (e.g. SOX)
 Group consolidation
 Cost analysis
 Operational planning and budgeting
 Other internal financial reporting
 Strategic planning
 Market/Sales planning/analysis
 Campaign management
 Production planning and control*)
 Supply-Chain analysis
 Supplier analysis
 HR analysis
 Other (list here)

*) excluded from analysis to avoid industry bias.

31
 Bernhard Wieder, Maria-Luise Ossimitz and Peter Chamoni

BI Use

What functional areas does the BI solution support, and how is the BI solution used in
these areas?
Passive users are report receivers only.
Ad-hoc 'reporters' are producers of ad-hoc reports (rather than re-usable reports).
OLAP users are authors of re-usable reports, analysts or power users in general (but not analytic
experts).
Analytic experts use 'business analytics
Passive Ad-hoc Analytics
methods', e.g. data mining techniques or OLAP users
users ‘reporters’ experts
artificial intelligence.
 Executives/Directors
 Accounting/Finance
 Purchasing
 Production/SCM*)
 Marketing/Sales
 Customer Support
 Human Resource Management
 IT/ORG
 Legal Department
 R&D (incl. Product Development) *)
 Other (please specify):

*) excluded from analysis to avoid industry bias.

32