Skip to main content

Distilling the wisdom from 100 years of research on voluntary job turnover

Thomas Lee

Tom Lee

Should I stay or should I go?

Management researchers have been studying this fundamental question, in the context of work, for over a century, producing more than 2,000 papers on the topic to help us understand, with increasing clarity, what motivates people to stay in their job or look elsewhere.

One of the field’s foremost experts is Thomas Lee, the Hughes M. Blake Endowed Professor of Management at the University of Washington Foster School of Business. Below, Lee discusses his field and the findings of his recent papers distilling 100 years of evidence into a modern set of best practices for managing employee retention and turnover.

What was the early thinking on voluntary turnover?

The vast majority of early research looked for predictors of turnover in industrial settings. The concern, primarily, was work disruption. So researchers looked at biographical data—what past behaviors might make a worker more or less likely to leave? In the 1950s, researchers began looking more systematically, adding personal characteristics to the predictors. But findings were very ad hoc, and largely proving guesswork rather than theory.

The first breakthrough came in 1958, with the work of James March and Herbert Simon, who theorized that the reasons people stay or go had to do with their decision to participate, which was based on their perceived ease of movement (alternatives) and perceived desirability of movement (attitude).

In 1977, William Mobley’s “intermediate linkages model” first articulated the steps an employee takes from dissatisfaction to leaving an organization, followed by a paper identifying major predictors of turnover. This launched the modern era of turnover research, and sparked an explosion of studies testing his ideas. Most found them to hold up.

When did you begin studying voluntary turnover?

I came into this field in the early 1980s. The ideas of Mobley were so intuitive that the field had fallen into a creative rut. Like everyone else, my dissertation and early work in this area basically confirmed the same results.

But in 1990, I decided it was time to take some risks. Literally the day I was awarded tenure at Foster, I drove over to Terry Mitchell’s house to discuss a new idea I had about turnover. Terry (a professor emeritus of management) was extremely renowned in the study of motivation, leadership, decision making and attribution theory. After that meeting with me, he became a turnover guy, too.

What are your major contributions?

The first was the “unfolding model,” which came out of that initial meeting with Terry. In addition to the personality and attitudinal factors that had been studied at length, we found that, in many cases, an external event or shock prompted people to quit. So maybe you were dissatisfied with your job, but it took another job offer or an abusive boss or some other event to prompt you to actually quit.

After publishing a series of papers exploring and proving this model, Terry and I decided to look at the other side of the turnover equation. I had always thought that the reasons for quitting are different from the reasons for staying. Without missing a beat, Terry counted all the reasons he stayed at the UW: because he was comfortable, because he was committed to his doctoral students, because he had trained everyone to his work style, because he could see Mount Rainier through his window, because he had gradually improved his Seahawks season tickets. That conversation eventually led to the concepts of links, fit and sacrifice: the dimensions of “job embeddedness.” Most people thought that satisfaction was the most important factor in staying in a job. We identified all of these other accumulated factors that made leaving just a little harder.

Our third major contribution is the new model of “psychological withdrawal states.” Most research has focused on enthusiastic stayers and leavers. But no one had given any serious thought to reluctant stayers and leavers: I want to leave, but I can’t. I want to stay, but I can’t. Our data shows that by identifying all of these distinct groups, you can improve the prediction of turnover substantially.

Has the study and prediction of turnover become more important?

There has always been a significant cost to losing employees and having to replace them. That cost has risen in the Information Age, when companies make their money on knowledge and innovation.

Today, companies care about retention and turnover not only during economic booms (as was the case when I joined the field), but also during recessions—when they can ill afford to lose their intellectual capital.

Is the Millennial generation a special case when it comes to retention and turnover?

Yes and no. The stereotype of today’s young adults is that they don’t just want a job, they want purpose, fulfillment. They are idealistic. This sounds, to me, a little like the Baby Boomers when they were just out of college. And, like the Boomers, perspectives may change when the Millennials have families, houses, savings.

Whether or not this generation is different, we do know that the relationship between employees and employers has changed. Through the 1970s, there was a strong social norm to staying with a company for an entire career. Today, there is no stigma to changing jobs and employers frequently. In fact, in many industries, there is a stigma to staying put too long.

So, is retention and turnover still worth our attention?

Absolutely. The general rule is that it costs two-times a person’s salary to replace them. It’s likely that people are going to leave eventually. But if you can keep them on the job for a little longer, motivate them to give a little more, there is real value in that.

Toward the ideal of perfect turnover prediction, where should researchers go next?

In another paper on the future of turnover research, we’re actually advising researchers not to seek the “holy grail” of perfect prediction. We’re saying, instead, to take what we have learned over the past 100 years—which is substantial—and approach the question from different perspectives. For instance, we recently published a study looking at turnover indicators over time. By measuring the trajectories of job satisfaction, engagement and embeddedness, we find that the predictive quality of surveys skyrockets.

Also, we have a pretty good grasp on the factors behind individual employee turnover. Now it’s time to begin looking at the dynamics of aggregate turnover—at the organizational level and even the industry level. I studied collective turnover rates in 1982, and it was the most widely ignored paper I’ve ever written. But now the statistical models and data analytics have caught up to the level that we can pull this off. It’s a great and glorious thing! If I were 40 again, I’d certainly be following this thread.

Tom’s Tips…

To predict voluntary turnover

  • Measure comprehensively – design surveys that capture job satisfaction (or organizational commitment), job alternatives, active job search, avoidance behaviors (like absences or lateness) and job embeddedness.
  • Track change over time – survey employees more frequently (at least three times a year); look for trends, which can be far more telling than single moments in time.
  • Analyze deeply – let the data drive decisions and policies (though always temper with your best managerial judgment).

To prevent voluntary turnover

  • Foster job embeddedness – build a culture conducive to healthy employee fit, connections and commitment to the organization and community.
  • Know your people – identify enthusiastic leavers and stayers and reluctant leavers and stayers—and manage appropriately.
  • Prepare for shocks – prevent or address events that push an employee to quit, such as supervisory abuse, a job offer or spousal relocation.
  • Understand Millennials – the new generation in the workplace may have a heightened desire to find purpose and personal development.