In an article on Forbes.com, HR analyst Josh Bersin extolled the benefits of talent analytics using a case study from a large financial services company:
“One of our clients… operates under the belief system that employees with good grades who come from highly ranked colleges will make good performers,” Bersin wrote. “So their recruitment, selection, and promotion process is based on these academic drivers.”
The firm conducted a statistical analysis of sales productivity and turnover, correlating performance and retention over the first two years against several demographic factors. They found that, of the six factors that corresponded with success, what did not matter was where candidates went to school, what grades they received, or the quality of their references. Within six months of implementing a new screening process, the firm increased revenues by $4 million.
However, for every one company that effectively harnesses their data, there are dozens that get it wrong by:
1. Overestimating performance as a predictor of potential. Research shows that only 30 percent of current high performers have management potential, and that most employees (more than 90 percent) would have trouble at the next organizational level.
2. Using subjective data. Too many companies dirty up their data sets with things like supervisor performance appraisals. Unfortunately, typical performance appraisals are a function of how much supervisors like their employees so, “high performers” are often those who successfully navigate office politics, not necessarily those who perform better.
3. Relying on incomplete data. Tomas Chamorro-Premuzic, recently wrote in The Guardian that “most organizations lack reliable systems for measuring employees’ performance … The result is … the equivalent of investing a great deal of money in weather forecasts without subsequently paying attention to the actual weather.”
4. Paying attention to irrelevant data. I recently read a Harvard Business Review article in which the author was describing the challenge of filling new positions for which data does not exist: “This poses different challenges, such as identifying patterns of your most successful hires, like the schools they come from, where they live…” Just because some of your best employees happen to be from the same school or town doesn’t have anything to do with whether they will be good programmers.
5. Believing that data eliminates uncertainty. On his ragan.com blog post, Jonathan Lewis wrote: “You can use data to reduce uncertainty, but don’t count on the data to eliminate it. The belief that uncertainty can ever be eliminated leads to unrealistic expectations, company paralysis, letdown, and frustration… We live in a complex and imperfect world, so no matter how big or little the data in our grasp, we will always have to make decisions with a certain level of uncertainty.”
Don’t get us wrong, we love data – Hogan’s research database has millions of data points, which we use to create, test and hone our assessments. The key to analyzing your company’s big data is to start with a valid, scientifically developed, objective tool like 360-degree feedback or personality assessment. These measures provide a picture of employees’ strengths, weaknesses, values, and work preferences. Using that information as a starting point, you can add in sales and performance data, demographics, and myriad other information to form a complete picture of how your organization, and your people, operates.