Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus convallis sem tellus, vitae egestas felis vestibule ut.

Error message details.

Reuse Permissions

Request permission to republish or redistribute SHRM content and materials.

How to Predict Who's Likely to Leave

July CoverWilliam Wolf, managing director and global head of talent development at Credit Suisse, joined the global financial services firm 18 months ago to further enhance its people analytics—a role he performed for clients as a partner at McKinsey & Co. The HR department at Credit Suisse has collected data on employees during their entire life cycle for many years, but wanted to delve deeper into turnover.

For instance, HR professionals already knew who left, segmented by demographics, performance ratings, regions and so forth. They knew why people left from information captured in interviews one month post-exit. But they wanted to go beyond these data to identify those at risk of leaving before the decisions are made.

“We already had data collected after the fact on who left and the reasons they gave us,” explains Wolf. “We needed to look at why we don’t have the compelling employee value proposition to keep that person here and at who else is at risk. The best way to learn that is to study the specific circumstances prior to the points of departure.”

The needed analysis was logistic regression—a statistics method used to predict an event with only two possible outcomes based on one or more predictor variables. Political scientists, for example, use logistic regression analyses to predict whether voters are more likely to vote Democrat or Republican based on age, income, gender, race or place of residence.

Wolf wanted to apply regression analysis to attrition data. Soon after arriving at Credit Suisse, Wolf assembled a small “people analytics” team made up of HR data analysts and a skilled quantitative methods expert he hired from a newly failed hedge fund to conduct predictive analyses.

The team began pinpointing variables common to people who stayed at Credit Suisse and variables common to those who left. They looked at 40 variables including:

  • Size of the person’s team.
  • Performance rating of the manager.
  • How long the person has been in a role.
  • How long the person worked at Credit Suisse.
  • Demographic traits.

“We came up with a group of variables that together can help predict whether a person is more likely to leave Credit Suisse or more likely to stay,” Wolf explains.

Wolf won’t share the variables that predict attrition at Credit Suisse, but he adds that they reflect corporate context and may differ for every company, and even for locations and groups within the same company.

The results have been eye-opening for HR professionals and managers at Credit Suisse. “If I know 10 facts about you in our context at Credit Suisse, I can estimate a probability of your leaving this year,” Wolf says. “Individual variables and groups of variables have given us extraordinary insight into the causes of voluntary attrition.”

Wolf gives the example of female attrition, an issue plaguing many financial services organizations. “The analysis revealed that at Credit Suisse there is nothing inherent about being female that makes one more prone to leave,” he says. Instead, “We need to better manage the way all people make transitions, such as coming back from an overseas assignment, becoming a supervisor or joining the company.” Women are more at risk, Wolf adds, because they tend to have more transition points in their lives than men do.

“The findings are shifting the discussion at Credit Suisse from being only about why women leave to addressing the transition problems that affect all groups,” Wolf says.

The analysis revealed some surprising results. For instance, “Our top-rated managers had greater attrition among their subordinates than among the managers rated at the second-highest grade,” Wolf explains. “We had to ask ourselves, ‘what is it about rock stars that makes people want to leave?’ ” Specific managerial training for top performers will be revisited in light of this analysis.

Sharing with leaders the variables that can lead to attrition helps them address risk factors. “The mitigation has to happen at the managerial level,” Wolf says. “We don’t want to create blanket HR policies to address the issues of the unhappy few. So, if we know people tend to leave if they have been in a position for more than four years, we aren’t going to institute a policy to move people every four years. But a manager can have a conversation with an employee that is better informed.”


​An organization run by AI is not a futuristic concept. Such technology is already a part of many workplaces and will continue to shape the labor market and HR. Here's how employers and employees can successfully manage generative AI and other AI-powered systems.