Predicting—and Preventing—Turnover

Research by Gilad Chen

Changes in job satisfaction are a good predictor of an employee’s turnover intentions.

When an employee’s job satisfaction improves, he or she will expect the aspects of their work that affect their satisfaction to continue improving into the future. But when satisfaction falls, the employee is inclined to believe that his or her job situation will probably continue to get worse.

Retention of talent is one of the main ingredients for success in many organizations. So companies are investing substantial resources into ways to retain their employees. Job satisfaction is a key factor in an employee’s decision on whether to stay or leave their organization. But it’s not just how employees feel about the job now—it’s about how their satisfaction has changed since last week or last month. That change in job satisfaction is a critical predictor of whether or not an employee is headed for the door, says Gilad Chen, professor of management and organization.

Chen and his co-authors studied levels of job satisfaction among four sample groups—two military and two managerial--and found that whether an employee planned to stay or leave depended on whether the person's job satisfaction was rising or declining over time.

Their results show that even when an employee’s overall satisfaction seems high, a negative change in job satisfaction increased the likelihood of that employee developing intentions to quit. An employee whose overall satisfaction decreases from a very high level to a slightly lower level is more at risk of leaving the organization than an employee whose satisfaction increases from a very low level of satisfaction to a moderate level of satisfaction, even though the first employee was more satisfied overall during that time period.

The study’s results are very robust, Chen says. He and his co-authors measured the change in satisfaction of civilians and members of the military at time periods that varied from 8 weeks to 6 months, and with newcomers as well as more tenured employees. The results were the same across all groups no matter how long or short the time period.

Part of this has to do with the human tendency to look at prior trends when projecting trends into the future. Any systematic change in a person’s work experiences shape their expectation for future work experiences. When an employee’s job satisfaction improves, he or she will expect the aspects of their work that affect their satisfaction to continue improving into the future. That leads the employee to want to stay with the organization. But when satisfaction falls, the employee is inclined to believe that his or her job situation will probably continue to worsen, and hence is more likely to leave.

Chen recommends that managers take the temperature of employee satisfaction on a regular basis, but especially when there is a major event in a company—new leadership, loss of a large client, a change in policies or a new initiative. “Think of any change as having the potential to trigger a trend,” says Chen, who emphasizes that it is important to step in to change downward trends quickly. “Because when things start spiraling downward, you need to think about how to improve employee satisfaction before the downward trend inspires them to leave.”

Managers should also consider the implications of how incentives and rewards are structured over time, says Chen. “It’s about the timing of rewards, not just financial rewards but promotion and recognition. If you throw all your weapons at one time, you’ll have a satisfied employee then, but you have to find ways to keep the employee at that level of satisfaction.”

“The Power of Momentum: A new model of dynamic relationships between job satisfaction change and turnover intentions,” co-authored by Chen; Robert E. Ployhart, University of South Carolina; Helena Cooper Thomas, University of Aukland, New Zealand; Neil Anderson, Brunel University, UK; and Paul Bliese, Walter Reed Army Institute of Research; is forthcoming in the February 2011 Academy of Management Journal. This research was partially funded by grant from the Army Research Institute for the Behavioral and Social Sciences. For more information, contact giladchen@rhsmith.umd.edu.