Closing the loop of big data analytics: The case of learning analytics

Much of the literature on business analytics assumes a straightforward relationship from human behaviour to data and from data to analytical insights that can be used to steer operations. At the same time, more critical scholars have suggested that the implications of big data analytics can go beyond improved decision making, sometimes twisting or even undermining managerial efforts. We adapt a theory of reactivity, originally developed to study university rankings, to identify various unintended effects of the application of big data analytics in an organizational setting. More specifically, we study the perceptions of a sophisticated learning analytics system among staff members of an internationally recognized business school. We find evidence for four reactive effects: re-allocation of resources, change in values, redefinition of work and practices, and gaming, and map these to four underlying reactive mechanisms: commensuration, self-fulfilling prophecies, reverse engineering and narratives. The study contributes toward theoretically broader, but also more practical understanding of big data analytics: reactivity may dilute the methodological validity of analytics to describe organisational and business environment for managerial purposes, yet the understanding of reactive effects makes a more potent use of analytics possible in organisational settings.