People Analytics, Trust Erosion, and Intention to Leave: The Role of Information Asymmetry

People analytics (PA) systems can enable data-driven decision-making but also have been described as surveillance. Falling back on privacy-calculus theory, we are leveraging a scenario-based survey among German employees to explain employees’ perceptions of PA deployment in their workplace. We find that analytical capability levels of PA—descriptive, prescriptive, and predictive—do not affect participants’ perception of constructs in our theoretical model. Employees’ privacy concerns about PA systems are strong enough to erode organizational trust to a level where employees are likely to consider leaving an organization, and risk perceptions outweigh employees’ perceived usefulness of PA systems. With different analytical capabilities not impacting perceptions, we interpret these effects as related to the employee’s realization that, with PA, managers have access to more information on employee behavior than the employee can get on their own behavior. We find that PA thus has an effect that reverses the traditional asymmetry in which employees generally have more information on their own behavior than their managers. Notably, employees’ perceptions of PA and the organizations using it are less negative when they are not aware of the information asymmetry. Our study contributes by highlighting implications of PA information asymmetry, as the resulting (unfulfilled) intention to leave can have negative consequences for employees’ wellbeing and performance. Further, we contribute theoretically to discussions about transparency of algorithmic systems. In our study, transparency does allow employees to perform an informed privacy calculus, yet they are not given the option to act according to it.