Last week, I explored how federal agencies can use HR data to build predictive models to evaluate and reduce costly employee turnover. An article published in Business Insider this month described how HR software company Workday built an app to help employers do just that. Workday claims its software can not only predict who is likely to leave at the individual employee level, but can also advise companies on what actions they should take to keep an employee from leaving, all by studying years of HR data and applying predictive analytics.
Complementing Predictive Models with Exit Surveys
I think predictive modeling using past HR data can be a powerful method of addressing turnover. To increase the predictive power of these models, employers should directly ask employees why they are leaving, and compile data on their answers. While gathering departing employee feedback can be challenging, direct employee opinions can help complement the workforce statistics and reveal trends that may not otherwise be as readily apparent. For example, an HR transaction database may show that an employee is transferring to another federal agency, but an exit survey could also reveal that the employee is leaving for a higher-paying position. Consistently surveying employees who leave can uncover trends of people who are accepting other positions for similar reasons, and direct the agency to review its compensation and rewards programs.
Designing an Exit Survey
An automatic electronic survey tool can be an effective method of gathering measurable data on why people are leaving. In my view, automatic and anonymous exit surveys are preferable to exit interviews in providing departing employees an opportunity to provide feedback. While employees may be wary of “burning bridges” during an in-person exit interview, anonymous surveys can allow for more honest feedback and provide a standardized format to compile and analyze the exit data.
Questions to ask in an exit survey
- Why are you leaving? Possibly the most important and obvious question to ask is the employee’s reason for leaving the agency. The results of this question can greatly complement an analysis of the types of employees who are leaving, and will reveal if employees are accepting other federal jobs (and at which agencies), are moving to the private sector (and possibly, to which companies), retiring, or are resigning for personal reasons.
- What made you start looking for another job in the first place? According to the Wall Street Journal, this question is even more revealing than asking “Why are you leaving?”, in that it generates the initial thoughts and experiences that spurred the employee to consider leaving the organization, and will generate a less “canned” answer.
- If accepting another position, is the position higher-paying? Are employees disproportionately leaving to accept more lucrative opportunities (in which case, is there one particular employer/agency that is drawing your employees away), or more troublingly, are people willing to accept lower-paying positions to leave the agency, and where are those employees going?
- Did the agency’s training and career development opportunities influence your decision to leave? An exit survey can be an important opportunity to evaluate an agency’s training and career development programs, as the survey could reveal that employees feel they need to look elsewhere to further their careers.
- Would you recommend the agency as a great place to work? This question can be an important link to the other questions in the survey. Linking the types of departures (question 1) to the employees’ overall satisfaction with the agency can help pinpoint which types of employees are most able to be influenced in their decision to leave. While some people may rate the agency as a great place to work but are leaving anyway (as might be the case in a retirement), other employees’ feedback could indicate which employees make up the agency’s avoidable turnover.
Analyzing the data and drawing conclusions
Questions in an exit survey should be designed to capture data that is not typically included in an HR database. Ideally, an exit survey would be automatically triggered when an employee has notified management of his/her decision to voluntarily leave the agency, to capture the maximum number of separating employees. The results can then be used to augment the demographic data that is included in HR database. For example, if the demographic data reveals that employees within a certain job title are disproportionately more likely to leave the agency, and the exit survey results reveal dissatisfaction with the agency’s training programs, a review of training opportunities for that particular job would be prudent as a way to reduce turnover. By augmenting predictive turnover models with employee feedback from exit surveys, the root causes of costly turnover can more easily be identified.
Matthew Albucher is part of the GovLoop Featured Blogger program, where we feature blog posts by government voices from all across the country (and world!). To see more Featured Blogger posts, click here.