From bit to business: Addressing managerial and practical challenges of analytics adoption

In this work, we address four research questions in order to offer an answer to our general premise: “Addressing challenges of data analytics adoption”. Namely, [RQ1] How do companies organize their analytics?; [RQ2] How mature are companies with regards to each of the DELTA dimensions?; [RQ3] Can automated analytics form a solution for the predicted shortages of data scientists?; and [RQ4] Can we leverage new ways of representing data to improve predictive analytics models? Hereby, the main theme of our work is enabling the creation of relevant business value out of data by means of analytics techniques. We categorize these challenges into two streams: a managerial and practical perspective.

We start with a management view on analytics. Employees can quickly set up small (independent) analytics projects. However, in order to truly enjoy the benefits of data analytics, senior management needs to consider how they can establish a data-driven decision-making culture. Only then it will become possible to make a strategic and competitive difference. Specifically, we describe the current analytics environment, challenges and opportunities. This is applied to the concept of analytics maturity in order to portray a growth path and defining indicators of higher maturity. Finally, we zoom in on the observation that the required data science skills are rare in the job market. Automated analytics platforms pose a possible solution here, and also bring along many other advantages such as improved analytics-business communication. In order to answer these questions, our focus is mainly on quantitative methods, such as survey research and experimental research, but we obtain also new insights through qualitative methods, i.e. interviews, which improve the validity of our work.

For the practical challenges, we start from a finding from the first part, namely that companies focus on comprehensible descriptive analytics techniques rather than advanced predictive ones. In particular, we concentrate on network analytics in order to improve both performance and comprehensibility of analytical models. By means of two practical use cases, in accountancy and retail, we research how we can leverage similarities and social interaction between entities. Both studies show the potential of using this new type of data representation and techniques. For this purpose, our research follows more closely a design science methodology since it aims to answer a particular problem, relevant to reality.

In conclusion, this work proposes several ways to deal with both managerial and practical challenges of adopting analytics. As such, we present novel findings, insights and models which can improve the analytics maturity of companies and leverage the massive amounts of data they are storing nowadays.