Exit interviews: how to run them and actually use them

An exit interview is a structured conversation with a departing employee about why they are leaving and what the company could do better. Done casually, it produces polite fictions; done properly, it is the cheapest organizational diagnostics available — departing employees have unique visibility and unusually little reason to soften it. The difference between the two outcomes is entirely in how the interview is run and whether anyone ever reads the results in aggregate.

Getting truth instead of politeness

• Timing: in the final week, after handovers are settled — early enough that the person is still engaged, late enough that nothing they say affects their reference or final pay.
• Interviewer: someone other than the direct manager — HR or a neutral leader. The most common reason for leaving is the manager; nobody tells that to the manager.
• Framing: state plainly that the conversation cannot affect references or settlements, that themes are reported in aggregate, and that the purpose is improvement, not appeasement — and then honor all three.
• Format: same core questions for every leaver, recorded consistently. Comparability across interviews is what turns anecdotes into signal.

Questions that work

• "What prompted you to start looking?" — the trigger, which is usually more informative than the final reason.
• "What did the new role offer that this one could not?" — separates pull factors (career step, pay) from push factors (management, environment).
• "What would have made you stay?" — the direct retention question; answers cluster fast across leavers.
• "What should we fix first for the people still here?" — reframes the person as an advisor, which is when candor peaks.
• "Would you recommend working here? Why or why not?" — the summary judgment, and a leading indicator of your hiring pipeline.

Using the answers: the part most companies skip

Individual exit interviews are anecdotes; the value is in the aggregate. Record every interview in a consistent form, tag the themes (management, pay, growth, workload, environment), and review quarterly against the turnover numbers: which teams lose people, and what do their leavers say? Three departures mentioning the same manager or the same broken promise is not a coincidence — it is a finding, and it deserves the same follow-up an audit finding would get: an owner, an action, a date. Companies that run this loop learn more from ten departures than most learn from a hundred.

Record-keeping matters here too: exit interviews filed per person in the HR system (in EmployDB, as private records visible to HR only) survive HR staff changes — so the pattern three years of leavers describe is still readable when someone finally looks.