Ridderprint

Chapter 7 160 Some of these skills have traditionally not resided in the HRM function (Andersen, 2017; Green, 2017; Ulrich & Dulebohn, 2015). HRM professionals are often praised for their interpersonal skills, but their knowledge of business operations, quantitative methodology, and advanced analytics is usually less developed. On top of this, the HRM talents who have the right capabilities often move to work in other fields (Andersen, 2017). While some scholars are hopeful and observe that a quantitative mindset is developing within the HRM function (Van der Togt & Rasmussen, 2017), others grimly conclude that the function is not yet ready for analytics (Angrave et al., 2016). Either way, effective implementation of people analytics requires different capabilities than traditionally present in the HRM function. Depending on the ambition of the organization, contemporary HRM departments may choose to develop these capabilities internally, to bring them in from other parts of the organization (Andersen, 2017), or to team up with external parties (Kryscynski et al., 2017; Van der Togt & Rasmussen, 2017). 7.4.2.2 No Perfect Data Second, people analytics is worth only as much as the data to which it is applied. The garbage in, garbage out principle should be clear by now. Data quality issues affect analysis conclusions and thus the quality thereof. Decision-making based on unreliable or invalid data can lead to decisions that result in the exact opposite of what was desired. For that matter, organizations completely new to HRM data and people analytics are in a luxury position. They can closely consider what HRM and employee information is of strategic importance and, thus, what their ideal HRIS looks like (Andersen, 2017). With what interval should employee turnover be registered and the headcount updated? What information needs to be stored on the onboarding process or the exit interview procedure? What do we need to know about our employees’ training and developmental activities? Additionally, these organizations could consult strategic HRM literature to build their employee survey from scratch. There is a plethora of research examining the psychological factors of work, how these can be operationalized, and how they influence business and employee outcomes (e.g., Van Veldhoven, Prins, Van der Laken, & Dijkstra, 2015). Most organizations will already collect and store HRM data structurally. On the one hand, these organizations should continuously assess and improve the quality of their HRM data. One important element is clarity regarding the validity and reliability of measurements: what do our HRM data points actually represent and how stable are they. Organizations could look at the underlying factor structure of their employee surveys, the rater interreliabilities in the selection or evaluation processes, the stability of performance or potential ratings, or the amount of errors on the demographic data in their HRIS. Where issues arise, it could be worth to change the collection process or the data definition. However, this touches on the earlier discussed exploration-exploitation tradeoff: to what extent do we improve the current employee surveys and HRIS (exploration) or continue to analyze the longitudinal data collected in the current, potentially suboptimal system (exploitation)? On the other hand, organizations should start piloting people analytics initiatives as soon as data quality reaches an acceptable level. Chapter 6 demonstrated the value of the