Understanding Business intelligence system adoption

(1)

Understanding Business intelligence

system adoption

the Unified theory of acceptance and use of technology (UTAUT)

application

Qiyu Chen Student number: 11210389

University of Amsterdam, Faculty of Science

Thesis Master Information Studies: Business information system Final version: 05-July-2018

Supervisor: dhr. ir. A.M. (Loek) Stolwijk Examiner: dhr. dr. D. (Dick) Heinhuis.

Abstract

Background and Purpose - Organizations have invested enormous resources in the usage of Business Intelligence Systems(BIS). BIS facilitates organizations in operational and commercial decisions. The purpose of this paper is to investigate the key drivers for BIS adoptions.

Design, Method and approach - The research model was developed based on the unified theory of acceptance and use of technology (UTAUT) theory and the Amsterdam Information Management (AIM) model. The model was tested by 99 survey respondents from communities of BIS professionals. Results - The three constructs (Performance Expectancy, Effort Expectancy and social influence) of UTAUT model are successfully validated to explain Behavioral Intention. While the Role of information manager cannot be proven to be a moderating variable.

Conclusion - The proposed model helps consultancy companies and BIS software vendors to design marketing and sales strategy based on the drivers of behavioral intention.

Keywords - Business Intelligence Systems(BIS), the Unified theory of acceptance and use of technology (UTAUT) , Amsterdam Information Management model (AIM)

(2)

2

Introduction

In the last decades, companies are invested resources dramatically in Business intelligence systems (BIS), which enable business users to discover their rich, reliable and relevant data (Chang et al., 2014). The BIS system requires large amounts of enterprise resources, including License cost, hardware,

implementations, maintenance and support (Sommer and Sood 2014). However, there are a large number of BIS were not adopted successfully (Kulkarni, 2017). As a result, those companies cannot get the full benefit from the investment of systems.

In addition, the BIS market is gradually moving towards a user-oriented environment. With compare to the traditional BIS system, the modern BI tools empower users easier access to the enterprise

applications and external data source systems (Howson et al., 2018). For example, under the traditional BIS architecture, if additional analytical dimensions from an ERP module need to add to a user

dashboard in the front end, the data elements first need to be mapped from the ERP system to a

centralized analytical data repository (eg. data warehouse). This process requires several approvals from both business and IT sides before actual development kick-off, which is time-consuming (Howson et al., 2018). According to the recent Gartner report by Howson et al (2018), the business analysts are enabled by modern BIS to establish direct connections to source systems, including ERP, CRM, and many other business information systems. As a result, end users of BIS require less development and support resource from IT when discovering and analyzing their data. The increasing flexibility of end users makes individual adoption of BIS system critically important (Howson et al., 2018).

Prior research shows there is a limited number of success factors are used for management reference (Yeoh et al., 2010). Also, many scholars contributed to the field of summarizing the success factors into technological and managerial perspectives. Among those scholars, the Unified theory of acceptance and use of technology (UTAUT) proposed by Venkatesh et al. (2003) has been most cited for technology adoption and utilization. The model provided a comprehensive understanding of the critical success factors of technology adoption (Foshay & Kuziemsky, 2014). This research will validate the UTAUT theory in the context of BIS adoptions. In addition, there is a knowledge gap in understanding the different roles played by information managers in the BIS implementation (Popovic et al., 2012). Understanding the emphasis from the different role of information managers will facilitate BIS implementation managers to define implementation strategy.

To address the above knowledge gap, the following research question is raised: “To what extent do the independent variables explain the variation in the behavioral intention of using Business Intelligence System (BIS)?”, with the sub-question: “Is there a difference in explaining this variation between the different roles played by the information managers?”. The research model integrates the Unified theory of acceptance and use of technology (UTAUT) with Amsterdam Information Management model (AIM), developed by Abcouwer & Truijens (2006), emphasis on positioning the role of information managers. The remainder of this paper is structured as follow. The next chapter reviews the UTAUT and AIM model. Subsequently, the development of the research model will be explained. Then the assessment of the research model and results are presented, followed by the conclusion and future research.

(4)

2

Literature review

Unified theory of acceptance and use of technology (UTAUT)

The theory of the Unified theory of acceptance and use of technology (UTAUT) model summarized the eight main models of technology adoption and conducted a wide range of online survey (Venkatesh et al. 2003). The model explained there are four independent variables (social influence, facilitating condition, performance expectancy, effort expectancy) which predict two dependent variables (Behavioral

Intention and Usage behavioral) (Venkatesh et al. 2003). Additionally, there are four moderator variables (Age, Gender, Experience, and Volutarience of use) influence the relationships between dependent and Independent variables (Venkatesh et al. 2003). In contrast to the previous research, UTAUT model presented a more holistic view of critical factors in relations to technology adoption (Waehama, McGrath, Korthaus, & Fong, 2014). Moreover, the explained variance of behavioral intention is 70% which means 70% of the Behavioral Intention can be defined by the three constructs: Performance Expectancy, Effort Expectancy and Social Influence (Venkatesh et al. 2003). It’s also summarized that the explanatory power of the UTAUT model is relatively higher than other theories (Waehama, McGrath, Korthaus, & Fong, 2014).

Figure 1, Unified theory of acceptance and use of technology (UTAUT) (Venkatesh et al. 2003)

Amsterdam Information management

The Amsterdam Information management (AIM) model provides a bridge between organization and information (Abcouwer & Truijens, 2006). The AIM was built on strategic alignment model from Henderson and Venkatraman (1996). Moreover, both vertical and horizontal dimensions are extended to deliver more insights on the dynamics of information management (Abcouwer & Truijens, 2006). This model is widely cited in positioning information managers and Business IT alignment. As demonstrated in figure 2, the framework is divided into nine segments, which is also called “Negenvlaksmodel” (Abcouwer & Truijens, 2006). The vertical part of the framework represents the connection between

(5)

3

Business, Information/ Communication, and Technology. The horizontal section represents strategy, Structure, and Operations. From this matrix, nine information managers’ roles are defined.

Figure 2, Amsterdam information management model (Abcouwer & Truijens, 2006)

Research Method

Extended research model and Hypothesis

Based on the UTAUT and AIM model elaborated in the previous chapter, this research use four independent variables to measure the impact on behavioral intention. The variable and definition are listed in below table 3.

Variable Type Definition Performance

expectancy

Independent Performance expectancy is defined as the degree to which the employees believe that using this new system will achieve higher performance (Venkatesh et al., 2003). Effort expectancy Independent Effort expectancy is defined as the degree of effort, which employees believe in using a

new technology (Venkatesh et al., 2003).

Social influence Independent Social influence is defined to what extent the influence from other individuals impact on the decision of a person whether accept or reject a new technology (Venkatesh et al., 2003).

(6)

4

Roles of information managers (RIM)

Independent In the horizontalsegments of AIM model, the role of information manager is categorized as Strategy, Structure and Operation (Abcouwer & Truijens, 2006).

Behavioral intention

Dependent The indication of a person's readiness to perform a specific behaviour (Venkatesh et al., 2003).

Figure 3 variables used in this research

Based on the variables defined in the above paragraph, figure 4 shows the Hypothesis. This research validates the UTAUT model in the context of Business Intelligence systems (BIS) adoptions (H1, H2, and H3). Also, the research aims to discover if any of the relations between Independent variables and Behavioural intention is a significant difference between various roles of information managers (H4, H5, and H6). For instance, compared with participants of Strategy group, Operational information managers may state that social influence has a larger impact on Behavioral Intention.

Hypothesis Independent

variable

Dependent variable H1 There is no significant relationship between Performance expectancy (PE) and

Behavioral intention (BI) to use Business Intelligence systems.

PE BI

H2 There is no significant relationship between Effort expectancy (EE) and Behavioral intention (BI) to use Business Intelligence systems.

EE BI

H3 There is no significant relationship between Social influence (SI) and Behavioral intention (BI) to use Business Intelligence systems.

SI, RIM BI

H4 There is no significant difference among Roles of information managers (RIM) for relationships between Performance expectancy (PE) and behavioral intention (BI) to use Business Intelligence systems.

PE, RIM BI

H5 There is no significant difference among Roles of information managers (RIM) for relationships between Effort expectancy (EE) and behavioral intention (BI) to use Business Intelligence systems.

EE, RIM BI

H6 There is no significant difference among Roles of information managers (RIM) for relationships between Social influence (SI) and behavioral intention (BI) to use Business Intelligence systems.

SI, RIM BI

(7)

5

Removed variables

There are several variables of the original model of UTAUT that are removed from this research to fit into the Context of Business Intelligence System adoptions.

Firstly, User behavior measures the actual usage of new technology after implementation (Venkatesh et al., 2003). This variable can only be measured after the introduction of Business intelligence system. Also, the managerial implication of this research is to facilitate organizations for successful Business intelligence system implementation. Therefore, it’s more relevant to measure the drivers of system adoption rather than the actual usage. Similarly, facilitating conditions, independent variable of user behavior, is also excluded from this research.

Age, Gender, voluntariness, and Experience are moderator factors in the original UTAUT model (Venkatesh et al. 2003). For Business intelligence system (BIS) implementation projects, experiences, gender, and age of end users cannot be changed easily in an organization. Therefore, these variables are not in the scope of the research. Also, the BIS implementations require an extensive amount of

investment, which is driven by high-level management (Popovic et al., 2012). Employees are required to use the system and collaborate with others. Therefore, voluntariness of use is also excluded in this research model. The research model is presented in Figure 5.

Figure 5, the research model

Research instrument

To ensure content validity, a survey is developed based on the theoretical framework. According to the Amsterdam Information Management model, the respondents will be grouped by a forced choice question into three horizontal categories of information management: Strategy, Structure, and

Operation (Abcouwer and Truijens, 2006). After the role of information managers is defined, the 7-point Likert-type scale is used for measurements of other research constructs. Some scholars argue that the

(8)

6

reliability of 7-point Likert scale is optimized with compare to 5-point Likert scale (Norris & Preston, 1997). More specifically, Likert type data collected is measured as interval data. Applying analysis to the (objectivist) data, which makes it possible to derive patterns from it, in order to accept or reject the hypotheses. This method ensures the detailed analysis of the relationship between independent and dependent variables.

Before rolling out the survey in BIS professional communities, the survey questions are validated by Business Intelligence system users. Seven users participated in the pilot survey. Their feedbacks and observation were used to adjust the wording of the survey questions.

Data collection and Analysis

In order to reach out as many participants as possible, the survey was posted in three Business intelligence Linkedin Groups, which in total there are over 600,000 users joined. Also, according to Gartner Magic Quadrant Business Intelligence and Analytics Platforms report, Tableau, Qlik, and Microsoft Power BI are ranked as Top three Leaders in the Business intelligence vendors market in the past three years (Howson et al., 2018). Therefore, the survey also posted in the user communities of 3 major BI vendors. After a month of data collection， ninety-nine response was collected. The sample size is aligned with the sample size used by previous scholars in the same type of research (Hair et al., 2011).

The respondents cover various geographical locations and Roles of information managers. Table 6 provides an overview of the characteristics of respondents. Regarding the size of employers, it is apparent that respondents are predominantly from employers with more than 1000 employees, which is 57% of the sample. In terms of occupation, more than 50% of the participants are from the

(9)

7 Figure 6, Demographics of sample collected

The data analysis was performed using Structural Equation Modeling (SEM), which the underlying algorithm (Partial Least Squares, PLS) has been widely used in information studies fields (Chin et al., 2003). This PLS method is selected for this research for two reasons. First, PLS is appropriate for

research models that are in earlier development stage and has not been validated extensively (Teo, Wei, and Benbasat, 2003). Secondly, the frequency distribution of the data collected is unknown, which also favors the use of this statistical model. For the statistical analysis, SmartPLS has been applied.

Results

Validity

To test the internal consistency for this research, the reliability analysis was performed for each construct. Wieland et al. (2017) highlighted that Cronbach’s Alpha is widely used in the contemporary research community to assess reliability. More specifically, Cronbach’s Alpha present how closely related a set of items are grouped under a latent variable (Shi, 1992). As illustrated in below table 7, the Cronbach’s Alpha values calculated by SmartPLS. Similar research suggests the satisfactory level of Cronbach’s Alpha is 0.7 (Lance et al., 2006). Therefore, are all four constructs achieved high internal consistency.

(10)

8 Figure 7, Reliability assessment using Cronbach’s Alpha

Partial Least Squares (PLS) Path Analysis

Partial Least Squares (PLS) is the Algorithm that finds cardinal directions in latent variables that explain the variance of dependent variables (Wold, 1982). Since 1997, this method has been widely used in similar research for assessing the loading between dependent and independent variables (Popovic et al., 2012).

According to Hair et al. (2017), Partial Least Squares (PLS) does not assume that data is normally distributed. In other words, the path coefficients calculated in PLS Path analysis cannot prove to be significant by using parametric significance testing methods (Hair et al., 2017). As suggested by Hair et al. (2017), Bootstrapping, a non-parametric procedure, is applied to resolve this problem. In this research, the bootstrapping analysis randomly extracted 5000 samples (with replacement). It’s

presented in below figure 8 that the three path coefficients are statistically significant, which p is lower than the predetermined level of test (0.05). Therefore, the first three hypothesis (H1, H2, and H3) are rejected.

Figure 8, Bootstrapping results

As illustrated in figure 9, the findings provided some interesting insights into the relationships between behavioral intention and the independent variables. The path coefficient between the relation of Effort Expectancy (EE) and Behavioral Intention is 0.474, which means if the Effort Expectancy increases by one standard deviation, the Behavioral Intention will increase by 0.474 times the standard deviation of Behavioral Intention. Similarly, the loadings of 1) Performance Expectancy (PE) to Behavioral intention and 2) Social Influence (SI) to Behavioral are 0.464 and 0.435 respectively. Therefore, it can be

concluded that Performance Expectancy, Effort Expectancy and Social Influence have a positive impact on Behavioral Intention. More importantly, the explained variance (R2) of this model is 0.676, which indicate 67.6% of the Behavioral Intention can be explained by the three variables (PE, EE, and SI).

(11)

9 Figure 9, Path Analysis for combined 3 groups

Multi-Group Analysis (MGA)

After validating the first 3 hypothesis, the moderating effect of information managers’ role is explained in this chapter. Similar to above step, the PLS Path analysis is also performed at groups level (Strategy, Structure, and Operation). As illustrated in below figure 11, the path coefficients are different across the three roles. Especially for the relationship between Performance Expectancy and Behavioral Intention, the loading from Group Strategy is 0.6; for the same path, the loading of group Operation much lower (0.368).

(12)

10 Figure 10, PLS Path analysis at group level.

To validate if the loadings are significantly different among the groups, the author analyzed the sample data in 3 pairs (Strategy - Operation, Strategy - Structure, and Structure - Operation). There are various methods used in IS research to identify if the PLS model significantly differs from groups, including parametric test and PLS-MGA. According to Sarstedt et al. (2011), the parametric analysis is suitable for continuous data and rely on distributional assumptions. On the other hands, PLS-MGA is a

non-parametric approach (Henseler, 2010). More specifically, the multi-group analysis allows researchers to test if pre-defined data groups have a significant difference in their group-specific estimates, in this case, Path coefficients (Hair et al., 2018). Besides, the ‘10 times rule’ method is widely used as minimum sample size estimation, which means the minimum sample size should be ten times bigger than the number of latent variables (Hair et al., 2011). Due to the unknown distribution of population and sample size is on the edge of minimum requirement (30 per group), the PLS-MGA approach is selected for this research.

The software SmartPLS is used for the PLS-MGA analysis, which 5000 subsamples are extracted for bootstrapping and the significant level is set as 0.05 with a 2-tailed test. The three groups comparison results of the PLS-MGA are presented in the following Figure 11. For the relationship between Social Influence and behavioral intention, the p-value of loading difference between group Operation and Structure is 0.703. In other words, the group-specific path is not significantly different with regards to

(13)

11

the significant level of test (Sarstedt et al., 2011). Similarly, the p-values for other groups are neither smaller than 0.025 nor bigger than 0.975. Therefore, H4, H5, and H6 are rejected based on the multigroup analysis.

Figure 11, P values of PLS-MGA analysis

Conclusion and Discussion

This research confirms the effect of four direct constructs (Effort Expectancy, Performance Expectancy, Social Influence, and Behavioral intention) from the original UTAUT model over Business intelligence system (BIS) adoptions. Knowledge gained in this research will enhance the ability of organizations in understanding the key drivers of Business Intelligence System (BIS) adoptions. On the other hands, the moderating effect of information managers’ roles cannot be proved.

There are some assumptions applied. First of all, it was assumed that participants provided their answers honestly in the survey. To facilitate the honest reply, participants were informed about anonymity and confidentiality before participants of the study. Also, Likert Scale is used for this

research and used as interval data, which assume the difference between the Likert scales are the same. For example, the interval between Strongly disagree-Disagree and Agree-Strongly Agree is the same. While some scholars argue that Likert type data should be classified as ordinal data since the increments on the scale are not equal (Norman, 2011). Therefore, it’s required for future researchers to be aware of the different treatment on the Likert-type data.

This paper can be considered as the pioneer of similar research. The role of information managers is classified using the three horizontal dimensions of Amsterdam Information Management model. For future research, it is worthwhile to include the three horizontal dimensions (Business, Information/ Communication, and Technology) as new constructs and examine if there is the three horizontal dimensions have the moderating effect on Business intelligence system adoptions that create new extensions of the UTAUT model. Secondly, we are not able to prove the roles of information managers have a significant impact on the independent variables (PE, EE, SI) towards the dependent variable (BI). On the other hands, for group comparison between Strategy and Operation, the p-value (the

relationship between Performance Expectancy and Behavioural intention) is 0.935, which is close to the significance level of 0.975; The p-value for the relation between Social influence and Behavioural

(14)

12

intention is 0.059, which is close to the significance level of 0.025. This results may vary because the sample size of this research is just above the minimum requirements (‘10 times rule’) of the Multi-group Analysis. For future research, if a larger sample size can be collected to represent a better

(15)

13

Reference

Aleš Popovic, Hackney, R., Coelho, P. S., & Jaklič, J. (2012). Towards business intelligence systems success: Effects of maturity and culture on analytical decision making. Decision Support Systems, 54(1), 729-739. doi:10.1016/j.dss.2012.08.017

Aoun, C., Vatanasakdakul, S., & LI, Y. (2010). AIS in Australia: UTAUT application & cultural implication. AIS in Australia,21. Retrieved June 17, 2018.

Chang, Y., Hsu, P., & Wu, Z. (2014). Exploring managers intention to use business intelligence: the role of motivations. Behaviour & Information Technology, 34(3), 273-285. doi:10.1080/0144929x.2014.968208 Colman, A. M., Norris, C. E., & Preston, C. C. (1997). Comparing rating scales of different lengths: Equivalence of scores from 5-point and 7-point scales. Psychological Reports, 80, 355-362.

Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1992). Extrinsic and intrinsic motivation to use computers in the workplace. Journal of Applied Social Psychology, (22:14), pp 1111-1132.

Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: a comparison of two theoretical models. Management Science, (35:8), pp 982-1003.

Foster, K., Smith, G., Ariyachandra, T., & Frolick, M. N. (2015). Business Intelligence Competency Center: Improving Data and Decisions. Information Systems Management, 32(3), 229-233.

doi:10.1080/10580530.2015.1044343

Foshay, N., Taylor, A., & Mukherjee, A. (2014). Winning the Hearts and Minds of Business Intelligence Users: The Role of Metadata. Information Systems Management, 31(2), 167-180.

doi:10.1080/10580530.2014.890444

Foshay, N., & Kuziemsky, C. (2014). Towards an implementation framework for business intelligence in healthcare. International Journal of Information Management, 34(1), 20-27.

doi:10.1016/j.ijinfomgt.2013.09.003

Hair, J.F., Ringle, C.M. & Sarstedt, M. (2011) PLS-SEM: indeed a silver bullet. The Journal of Marketing Theory and Practice, 19(2), 139–152.

Hair, J. F., Sarstedt, M., Ringle, C. M., & Gudergan, S. P. (2018). Advanced Issues in Partial Least Squares Structural Equation Modeling (PLS-SEM), Thousand Oaks, CA: Sage.

H.H. Teo, K.K. Wei, I. Benbasat, Predicting intention to adopt interorganizational linkages: an institutional perspective, MIS Quarterly 27 (1) (2003) 19–49.

Howson, C., Sallam, R., Richardson, J., & Tapadinhas, J. (2018, February 26). Magic Quadrant for Analytics and Business Intelligence Platforms. Retrieved June 13, 2018, from

https://www.gartner.com/doc/reprints?id=1-4RXB56A&ct=180227&st=sb&signin=fad99c5c102ff0dd4389e0d8a4ea081e

Lance CE, Butts MM, Michels LC. The Sources of Four Commonly Reported Cutoff Criteria: What Did They Really Say? Organizational Research Methods. 2006 Apr 1;9(2):202–20.

(16)

14

L. T., & Lin, G. C. (2004). Critical success factor framework for the implementation of integrated-enterprise systems in the manufacturing environment. International Journal of Production Research, 42(17), 3731-3742. doi:10.1080/00207540410001721781

Kulkarni, U. R., Robles-Flores, J. A., & A. P. (2017). Business Intelligence Capability: The Effect of Top Management and the Mediating Roles of User Participation and Analytical Decision Making Orientation. Journal of the Association for Information Systems, 18(7), 516-541

Marko Sarstedt, , Jörg Henseler, , Christian M. Ringle, (2011), Multigroup Analysis in Partial Least Squares (PLS) Path Modeling: Alternative Methods and Empirical Results, in Marko Sarstedt, Manfred Schwaiger, Charles R. Taylor (ed.) Measurement and Research Methods in International Marketing (Advances in International Marketing, Volume 22) Emerald Group Publishing Limited, pp.195 - 218 McAfee, A., Brynjolfsson, E., Davenport, T. H., Patil, D. J., & Barton, D. (2012). Big data. The management revolution. Harvard Business Review, 90(10), 61-67.

Moore, G. C., & Benbasat, I. (1991). Development of an instrument to measure the perceptions of adopting an information technology innovation. Information Systems Research, (2:3), pp 192-222. Norman, G. (2011). Likert scales, levels of measurement and the ‘‘laws’’ of statistics. METHODOLOGIST’S CORNER. doi:10.1007/s10459-010-9222-y

Sangar, A. B., & A.Iahad, N. B. (2013). Critical Factors That Affect The Success Of Business Intelligence Systems (BIS) Implementation In An Organization. NTERNATIONAL JOURNAL OF SCIENTIFIC &

TECHNOLOGY RESEARCH , 2(2), 176-180.

Shi, S. G. (1992). Accurate and efficient double-bootstrap confidence limit method. Computational Statistics and Data Analaysis, 13(1), 21–32.

Sommer, D. and B. Sood. 2014. Market Share Analysis: BI and Analytics Software 2013. http://www.gartner.com/doc/2723017 (accessed 8 Jun, 2018).

Venkatesh, V. (2006). Where to go from here? Thoughts on future directions for research on

individuallevel technology adoption with a focus on decision making. Decision Sciences, 37(4), 497-518. Venkatesh, V., & Morris, M. G. (2000). Why don't men ever stop to ask for directions? Gender, social influence, and their role in technology acceptance and usage behavior. MIS quarterly, 24(1), 115-139. Venkatesh, V., Morris, M. G., & Ackerman, P. L. (2000). A longitudinal field investigation of gender differences in individual technology adoption decision-making processes. Organizational behavior and human decision processes, 83(1), 33-60.

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS quarterly, 27(3), 425-478.

Waehama, W., McGrath, M., Korthaus, A., & Fong, M. ICT Adoption and the UTAUT Model (2014, November).Paper presented at the International Conference on Educational Technology Information Technology (ICETIS2014),Bangkok, Thailand.

(17)

15

Wang, H. (2014). Distinguishing the adoption of business intelligence systems from their implementation: the role of managers’ personality profiles. Behaviour & Information Technology, 33(10), 1082-1092. doi:10.1080/0144929x.2013.869260

Wieland, A., Durach, C.F., Kembro, J. & Treiblmaier, H. (2017), Statistical and judgmental criteria for scale purification, Supply Chain Management: An International Journal, Vol. 22, No. 4,

https://doi.org/10.1108/SCM-07-2016-0230

W.W. Chin, B.L.Marcolin, P.R. Newsted, A partial least squares latent variable modeling approach for measuring interaction effects: results from a Monte Carlo simulation study and an electronic-mail emotion/adoption study, Information Systems Research 14 (2) (2003)

Yeoh, William and Koronios, Andy 2010, Critical success factors for business intelligence systems, Journal of computer information systems, vol. 50, no. 3, Spring, pp. 23-32.

(18)

16

Appendix: The research model and Survey questions

Construct Items Source

Roles of Information Managers

(forced choice question)

Question: Which of below activities best describe your scope and core competency in your

organization?

Answer A (strategy information manager) - Determine organizational products/

services, ICT structure

- Position the organisation in the marketplace - Assess emerging technology, and decide on

partnership with key technology partners - Participant/ decide on strategic buy/ make

IT standard;

Answer B (structure information manager)

- Determine / (re)design the critical business process; develop business models;

- Determine organizational technology (data, system, configuration) architecture

- Develop organisational I/C and knowledge model

- Design critical IC, ICT process

Answer C (operation information manager) - Design, perform and monitor changes in

business process and ICT process; - Acquisition, training and development of

skills of business & ICT professionals

Henderson Venkatraman and Oldach (1996); Maes (1999); Abcouwer and Truijens (2006) Performance Expectancy

(measured using 7-point likert scale)

PE1: business intelligence systems would have a positive impact on my overall performance and productivity.

PE2: business intelligence systems would help me make faster and accurate business decisions. PE3: Overall, I would find Business intelligence

Provost & Fawcett (2013);

Venkatesh et al. (2003)

(19)

17 system to be advantageous. Effort Expectancy

(measured using 7-point likert scale)

EE1: Learning business intelligence systems would be easy for me.

EE2: Generating actionable insights from business intelligence systems would be easy for me.

EE3: Using Business intelligence system would fit to my daily work activities.

Venkatesh & Davis (2003);

Oliveira & Popovic (2014)

Social Influence

(measured using 7-point likert scale)

SI1: My colleagues would think it is important to use Business intelligence system for our team.

SI2: My manager would think my role more important if I apply business intelligence systems. SI3: In general, my employer would support using Business intelligence systems.

Venkatesh et al. (2003)

Behavioural intention (measured using 7-point likert scale)

BI1: I am interested in using Business intelligence system in the coming month.

BI2: I am interested in using Business intelligence system in the coming 3 months.

Venkatesh et al. (2003)

(20)

18

Appendix: Linkedin Groups and BI Communities used for survey

Name Number of followers/

Members

Link

Linkedin group:

Business Analytics, Big Data, and Artificial Intelligence

186k https://www.linkedin.c

om/groups/62438

Linkedin group:

Big Data, Analytics, Business Intelligence & Visualization Experts Community

om/groups/23006

Linkedin group:

Business Intelligence Professionals (BI, Big Data, Analytics, IoT)

om/groups/40057

Tableau community 130k https://community.tabl

eau.com/community/fo rums

Power BI community 37k http://community.pow

erbi.com/t5/Forums/ct-p/Forums

Qlik community 200k https://community.qlik.

(21)

19

Appendix: PLS Path analysis for 3 groups (Strategy, Structure and

Operations)

Strategy group validity result:

Strategy group Bootstrapping results:

Structure group validity result:

Structure group Bootstrapping results:

(22)

20 Operation group Bootstrapping results:

Understanding Business intelligence system adoption