Ridderprint

Chapter 7 162 7.5.1 People Analytics Future research on people analytics should take three different perspectives. First, future research should explore the strategic value of people analytics and evidence-based HRM. On a macro-level, they could compare HRM departments and organizations with differing levels of people analytics maturity and evidence-based mindsets to explore to what extent this influences HRM efficiency and effectiveness, HRM’s credibility as rated by other disciplines, and business, HRM, and operational outcomes. I do foresee that a quantification of the benefits of people analytics on this level could be difficult due to inter-organizational differences. Hence, on a micro-level, scholars could examine how people analytics affects the decision-making process of individuals. For instance, to what extent does people analytics make decision-makers more confident, faster, or unanimous in their HRM decisions, or to what extent does people analytics increase their capabilities to predict the impact of their decisions. On a case-level and more focused on impact, scholars and practitioners could aim to quantify how changes in HRM policies and practices made in light of people analytics projects have affected operational and financial results. Second, scholars should explore the implications of new data sources. Modern technology is producing richer data which may hold value for a wide variety of HRM research topics. For instance, text mining applied to exit interviews could help to uncover motives for employee turnover. Characteristics of employees’ social networks could be predictive for their long-term career success. Sentiment analysis of messages on enterprise communication tools could help to gauge employee engagement. Sensors may help to explore work andmeeting patterns and optimize work floor designs. “Smart” work badges may help to map team cohesion and inclusion. E-mail software could be consulted to explore the health implications of overwork measured by after-hours emails. The possibilities seem endless and, as a result, many of the aforementioned analyses are already pioneered in practice. More scholarly involvement is needed to provide new and strong theoretical perspectives and to define the legal and ethical boundaries and responsibilities when it comes to these data. Third, HRM scholars should get involved in research on machine learning within the HRM domain in order to explore its strategic and practical implications. Machine learning will allow for more personalized HRM. Training courses, career steps, or mentors could be recommended to employees on an individual level, based on their current needs and interests and those of historic employees similar to them. Are such practices I-deals 2.0? How do they affect the psychological contract? Machine learning will also cause a wave of automation, which is already visible in the systems supporting HRM professionals in their HRM decisions. Here, scholars could examine in what HRM process (e.g., selection, remuneration, career planning) and in what stages of those processes (e.g., first screening, candidate ranking, employee selection), a decision-support system would be more or less valuable. Alternatively, scholars could explore how different parameter settings of decision-support systems or operationalizations of the outcome variables prevent or introduce bias; how quickly