How to Supercharge People Analytics: Add Psychology

In today’s organizational culture, companies increasingly use people analytics — the analysis of people-related data — to uncover useful insights. People analytics can be very effective, from viewing descriptive data in “dashboard” form to highlight important information, to more sophisticated analyses to identify important patterns, such as high-turnover-risk worker subgroups. People analytics can even suggest some kinds of actions — for example, if analyses indicate that new hires who complete on-boarding activities are more likely to stay employed longer, that suggests that you should make sure new hires complete those activities.

In our experience, however, people analytics often does not deliver all the value it could. Many results describe, and even predict, outcomes, but do not explain them. Without knowing why something is happening, it can be hard to know how to change it.

This is where psychology can help. Psychological variables, based in the scientific study of human cognition, emotions, and behaviors, have the potential to provide explanatory insight, which can yield concrete ideas for taking action.

There are many psychological variables that we know are strongly related to important outcomes at work. And they can explain more about why something is happening, not just that it’s happening.

Psychology in People Analytics

Imagine your organization is trying to protect against workers quitting. Your people analytics team tells you that employees with six months’ to 2 years’ tenure, in the Midwestern region, with technology skills, in the salary range of $50-80,000, are at the highest risk of quitting. That is valuable information, especially if those employees are particularly important for your business.

Your next steps, however, are not clear. What is it exactly that could help keep these employees from quitting? One way to begin adding psychological variables is to ask the departing workers why they’re leaving (exit interviews). Imagine you learn that workers are being offered higher pay and more flexible job arrangements elsewhere, and that’s why they’re leaving.

Those new variables add some valuable insights, and they begin to reveal some information about the root causes of your issue. At this point, you might consider offering higher salaries to retain workers. However, a research psychologist would recognize that exit interview results can be biased by low response rates or other factors (Giacalone et al., 1997); more importantly, memory is often inaccurate, even for one’s own thought processes, and exiting employees might explain their behavior in self-enhancing ways (Gramzow & Willard, 2006).

Now, imagine the valuable information that more psychological variables might yield. For example, research shows that workers’ relationships with their leadership and the meaningfulness of their work are associated with their propensity to quit. If your organization measures these variables, they can be included in a people-analytics model. It is possible that, all things considered, salary is not a top reason for quitting; instead, manager relationships and meaningful work are what’s really driving turnover.

With the added information from the psychological variables, there are now clear options for how to take action. Research has identified many ways to work on improving relationships with managers and help people find more meaning in their work (see for example Martin et al., 2010 and Dik et al., 2013).

Benefits of Adding Psychology to People Analytics

In short, here are the main benefits of infusing the science of psychology into people analytics:

• Stronger methodology. Research scientists are experts at how to study people. They can assess people analytics methodology and add valuable new variables.
• Stronger models. Statistical models are only as good as the variables included in them. If an important variable is omitted from people analytics models, the usefulness of the models will be limited.
• More insight into “why.” Knowing why something is happening can provide far more useful information than simply knowing how much, or where, it is happening.
• More direction on how to take action. The most effective analytics yield results that can guide actions. Psychological variables can help you go from passively describing to effectively prescribing action.