The use and impact of Business Intelligence (BI) systems in management accounting: an empirical study in the South African context

Abstract

Business Intelligence (BI)systems are used forgathering, managing, analysing, and sharing information about external and internal environments to gain insights for better decision making. Significant investments are being made in implementing Business Intelligence systems, but the extant evidence is limited as regards their impact on individual user performance and organisational performance. BI systems may have an especially important role to play in support of management accounting functions. This is because BI systems have the potential to assist management accountants to produce planning, costing, budgeting, forecasting and performance reports, and help them provide relevant and timely information to other managers, thus aiding their organisations to stay ahead of competitors. Yet, evidence suggests BI use may be low among management accountants who are often observed to perform their tasks outside of BI systems, such as maintaining their individual spreadsheets. To address this problem, this study investigated the impact and use of BI systems at the individual level and organisational level, with a focus on the management accounting context. More specifically, this thesis examined the impact of innovative use of BI systems on the individual-level performance of management accountants and examined the extent to which organisational use of BI systems impacts a balanced set of performance measures, both directly and indirectly through impacts on the sophistication of management control systems. To understand the factors affecting the use of BI systems and the impact of BI systems on individual performance, this thesis developed an extended DeLone and McLean (D&M) information system (IS) success model. The factors affecting use of BI systems included system quality, data quality, information quality and service quality. Use was conceptualised as both routine use (e.g., accessing standard reports, executing existing queries and running ad-hoc reports) and innovative use (e.g., using predictive analytics, self-service tools, constructing queries and visualizations). Task complexity was considered for possible moderating effects on the relationship between BI use and performance. The model was tested through a cross-sectional design using data collected from a sample (N=363) of management accountants, and through a longitudinal design using matched data from the management accountants collected over three time periods (waves) approximately 12 weeks apart. Structural equation modelling with AMOS was used for all model testing. The individual level study confirmed that factors influencing the innovative use of BI systems by management accountants are system quality, data quality, information quality and service quality, along with self-efficacy and task complexity. This study has also determined that routine use is less likely under conditions of task complexity and does not rely on self-efficacy to the same extent as innovative use. In addition, innovative use was found to be significant for individual performance by improving productivity and decision-making effectiveness. Furthermore, longitudinal results indicated that innovative use and user satisfaction are significant for individual performance at later time waves, improving productivity and decision-making effectiveness. Management accountants using only routine BI features were not found to enjoy improved performance. The thesis has thus contributed new insights into BI use among management accountants, confirmed the attributes of a BI system that promotes their use, confirmed the importance of distinguishing between routine and innovative use, confirmed that innovative use is especially important to user performance when task complexity is high, and that routine use is less likely to lead to productivity and decision effectiveness among management accountants. Moreover, the longitudinal design with tests of cross-lagged structural models over the three-time waves allowed the study to also confirm the causal ordering of variables within the IS success model, which prior research had not adequately established. The thesis has also found that the adoption of BI systems into the operational functional areas of an organisation has an impact on organisational performance. The organisational impact was examined using the Balanced Scorecard (BSC) approach. The BSC defines a balance set of performance metrics including both financial and non-financial performance. This study also argued that an important intermediary performance outcome of BI systems is improved organisational Management Control Systems (MCS). MCS may help translate investments in BI into improved financial and non-financial performance outcomes. Data was collected using a survey of senior financial and other executives from 396 organisations drawn from a sample of Financial Mail (2016) top 200 companies and 196 public entities listed by National Treasury (PFMA, 1999). Model testing showed that BI systems can improve the non-financial performance of the organisation (customer performance, business process performance, and feedback and learning), as BI systems may turn data into actionable information. Furthermore, BI systems were found to directly improve the financial performance of organisations. This study thus provides sufficient evidence that adoption of BI systems is seen to improve the balanced set of performance measures. This is important for management to justify investments into BI systems within their organisations. The study also found that greater sophistication of MCS brought about by using BI systems impacts on organisational non-financial performance, particularly customer performance and business process, which then improve future financial performance. Feedback and learning performance was not found to have an impact on improvement of future financial performance in the sample. This thesis has made a number of contributions. A theoretical contribution has been made by developing and testing an extended IS success model for BI systems in the management accounting context. The model includes factors such as data quality, distinguishes routine from innovative use, and incorporates consideration of task complexity. The work also brings new insights into the causal ordering of IS success constructs through analysis of cross-lagged models using data collected over three time-waves. This helped overcome limitations that arise when data is collected using only cross-sectional surveys, which lack temporal precedence. Theoretical contributions were also made by identifying measures of financial and non-financial performance that are relevant to understanding the impacts of BI at the organisational level, along with the role MCS plays in translating BI into performance outcomes. The thesis results will assist organisations on how to invest in and deploy BI systems, as well as the types of interventions that may be needed to extend individual users from routine use to achieve more innovative use of BI systems