How to predict who will quit — and whom to poach

Predictive analytics is already well-known for helping organizations better analyze customer interactions, predict systems failures and forecast emerging market trends. But turn the lens inward, and analytics can also help managers improve employee retention, hiring, leadership development and engagement.

When it comes to choosing candidates to hire, making promotion decisions, and ascertaining which employees may  leave the company in the next few months, managers typically rely on gut feeling rather than hard evidence. But advances in workplace analytics are changing that, allowing for more data-driven decision-making around personnel issues.

Here’s how predictive analytics can help reduce turnover and give your organizations insights as to whom it should target for direct recruiting efforts.

Reducing turnover

Turnover is expensive — recruiting, hiring, onboarding and training a new employee can cost upwards of 1.25 times their salary. That means businesses must make smart hires, and retain prized employees to get a decent return on their investment. Here, the ability to predict why and when talent will leave can provide valuable insights and save firms a substantial amount of money.

“When people start leaving, it often happens in waves — turnover contagion. Big data and savvy analysts and engineers can write algorithms that find these patterns and pinpoint the cause — who’s the manager? How long is their commute? When’s the last time they got a raise? When was their last promotion? Then, you can put plans in place to mitigate those factors,” says Dave Weisbeck, CSO for workplace analytics firm Visier.

New research from Harvard Business Review and ENGAGE Talent show how, using big data, firms can track indicators of turnover propensity and identify employees who may be at risk of leaving the organization, says Matt Pietsch, CRO at ENGAGE Talent.

“The research shows us that there’s four major elements that can predict why someone decides to make a change,” Pietsch says. “This can be useful for organizations to predict when their own people are considering leaving and what factors exist that might make valuable talent at their competitors open to making a move.”

The four elements are:

A life event: A divorce, a marriage, their spouse has been transferred or has accepted a new job in a different geographic location. This could also include a company event such as a downsizing, layoffs, lost funding for projects and departments, or even a natural event like a hurricane, tornado, flood, Pietsch says.

A management issue: Anything that creates unease or distrust in the workforce can be a concern for turnover, says Pietsch, including public scandal. “Is the CEO facing a sexual harassment suit? Is the company under investigation for fraudulent activity? Whatever the issue, it makes people feel uncomfortable at the company,” he says.

The work environment and job embeddedness: The HBR research defines job embeddedness to determine how connected to other employees and the larger community employees feel, Pietsch says. Are colleagues leaving? Have there been layoffs or downsizing in other departments? Are workers not being encouraged to be active in the greater community? Is the company becoming less well-regarded in the outside world? Is the company losing market share? Is management failing to bring new talent and new ideas into the workplace? Concerns such as these could all be potential indicators that a person may consider quitting, Pietsch says.

Market demand: Market activity is always a consistent predictor of attrition, says Pietsch, whether the economy is good or in a recession; some roles may be more prone to attrition than others, and that’s important to keep in mind. “If I’m a software engineer, for example, I can leave my job at 8 a.m. and have a new one lined up by 8:15,” he says. “But if you’re in a role that’s less in-demand, it could take you six to twelve months to find a new position — that is going to majorly impact attrition and when and how people leave your organization.”

Predicting turnover at competitors

Organizations can also use predictive analytics to determine who’s most likely to quit at competing organizations with an eye toward recruiting that talent, says Pietsch. By understanding the factors that impact attrition, it’s possible to improve the odds that a potential hire would respond favorably to a call or an email from a recruiter or a hiring manager.

To test this theory, HBR and ENGAGE Talent created a Turnover Propensity Index (TPI) score for more than 500,000 individuals working in various industries across the U.S. The TPI was based both on publicly available data about an individual’s employer — including changes in Glassdoor or analyst ratings, stock price variation, news articles, and regulatory or legal actions — and that individuals work status and history, including their number of past jobs, employment anniversary and tenure, skills, education, gender, and geography. Machine learning algorithms then classified each individual as unlikely, less likely, more likely, or most likely to be receptive to new job opportunities.

“First, we wanted to see how well the TPI predicted openness to recruitment messages,” wrote the research authors, Brooks Holtom, professor of management and senior associate dean at Georgetown University, and David Allen, professor of management and associate dean at TCU and distinguished research environment professor at Warwick Business School.

The researchers then sent e-mail invitations to a sample set of 2,000 individuals to view available jobs tailored to their specific skills and interests. Of the nearly 75 percent who received the e-mail, 161 opened the invitation and 40 clicked through. “Those who were rated as ‘most likely’ to be receptive opened the e-mail invitation at more than twice the rate of those rated as least likely (5.0% versus 2.4%),” according to the research. “Additionally, among those who opened the email, those rated as ‘most likely’ to be receptive were significantly more likely to click through it.”

This suggests that the TPI score could identify employees at greater risk of leaving, and that companies could use publicly available data to strategically target top talent that might be more open to an outside offer, the authors conclude.

“This helps hiring managers and recruiters understand which people are much more open to receiving a call or an email from a recruiter,” Pietsch says. “And those people are 60 percent more likely to have a new job within 90 days, according to the research. Not only that, in advance of someone accepting a call, we can also look at what they need to do to retain them once they have made a move.”

Job satisfaction: A key metric worth quantifying

When it comes to retaining or recruiting talent, an individual’s sense of job satisfaction is key. Here, Pietsch notes that there are five core elements that indicate job satisfaction: strong leadership, business stability, company resilience, growth opportunities and a positive work environment.

But not every employee weighs these core elements equally. So determining which are most important to an individual can help an organization target or retain them based on the personal and emotional factors that are important to them, he says.

“For example, if someone’s business stability score is very high, and they care deeply about company resilience, then they most likely won’t want to work for a startup, because those are inherently unstable businesses,” Pietsch says. “So you can tailor your outreach attempts to zero in on underlying factors and better the chances of attracting and retaining the right talent.”

The Bureau of Labor Statistics Job Openings and Labor Turnover (JOLT) survey shows an unprecedented labor market that’s making it complicated for organizations to hire enough talent for their open positions, Weisbeck says. The number of open positions has remained steady for months, as has the number of available job seekers; what’s creeping upward is the number of “quits” and “separations,” meaning people leaving their current employment for other positions.

“Being able to pinpoint the reasons your talent is unhappy, disengaged, actively looking for other roles or about to quit is a competitive differentiator,” he says. “Five years ago, we saw only large enterprises using data analytics in this way; now it’s no longer a nice-to-have, it’s a must-have, even for smaller companies.”