We live in an age of data. Companies target consumers based on what they watch on YouTube or what they click on Facebook. Businesses create incredibly accurate production forecasts based on mounds of data coming in from every angle. Hey, dating sites are even mining data to connect people for the perfect romantic relationships.
So, it’s no surprise that the HR/recruiting realm is also using an increasingly sizable amount of data to improve operations. Data use in HR is sometimes referred to as “people analytics.”
“People analytics is the science of predictive analysis applied to the understanding of people,” says David Solot, Ph.D., analytics product manager at talent development solutions firm Caliper. “It’s the idea of evidence-based practice expressed in a modern, big data world. What that boils down to is using data from multiple sources to make predictions about the future.”
Among the many applications HR departments have found for people analytics are hiring, onboarding employees, crafting career paths, building teams, reducing risk, and managing turnover.
Where HR Data Comes From
Data usage in HR is new enough that professionals in smaller operations might still be in the early stages of implementing data-driven practices. With so much data and so many sources, figuring out how to proceed can be intimidating.
“There are many types of data that HR professionals can add to their data lakes,” says Solot. “Some popular data sources are employee surveys, personality profiles, scorecards and dashboards, labor market data, and 360 feedback. The key is to build a data lake that has enough diverse data that it can help you answer all of the strategic questions that you are asked. You may already have a lot of this data on your existing systems. For others, you may need to reach out to a third-party vendor.”
Technology is so integrated into modern business operations that using data to drive decisions has become essential. Companies that fail to source and evaluate data to influence talent decisions will suffer from high turnover and reduced profits.
“The potential for ROI is staggeringly high,” Solot says. He points to a recent report from Bersin by Deloitte, which found organizations “using people analytics in a sophisticated and insightful way report 82 percent higher three-year average profit.”
Data may drive results, but ultimately, those results come from human beings who determine how the information gathered should be used.
“There’s a lot of emphasis these days on machine learning and AI,” says Solot. “However, machines will never bring the same type of judgement to a situation that a human will. Machines can do a hard analysis on a set of variables, but the outcome is only as good as the data and the algorithms that are put into it. A well-trained HR professional needs to look at the output from any people analytics study and ask: ‘Does this make sense? Is this applicable to our situation? What’s the upside and the downside of using this output?’”
For all its benefits, technology has limits. Computers don’t understand social cues, societal mores, or how certain actions may impact a brand.
“Consider the true but now-anecdotal example of the real estate agency that wanted to hire someone to do residential sales for a set neighborhood,” Solot says. “Their data indicated that the best salesperson would be someone who lived in that same neighborhood — makes sense so far, doesn’t it? — so they restricted their search in that way. However, the neighborhood they were searching in was predominately white, while the surrounding areas were more diverse. By restricting their search the way the data told them to, they were imposing a de facto racial bias on their hiring practices. They were sued for discrimination and lost. This is a good example of why we must question our people analytics results. Machine learning and AI are good, but they should be used to augment human judgement, not replace it.”