Exit Interview: 9 Questions That Uncover the Real Reason Staff Leave

A 42-table restaurant in Maslak ran the same 9 exit-interview questions for three years. Turnover dropped from 47% to 22% — because the truth hiding under "pay" answers finally surfaced.

Why Most Exit Interviews Fail

Asked "Why are you leaving?", almost every server says "better opportunity" or "pay". Those are politeness filters, not data. In the Maslak dataset, the first-stated reason differed from the reason that emerged by question three in 68% of interviews.

The interview must be conducted by someone who is not the person's direct manager — HR, accounting, or an external consultant. Promise 30-day confidentiality. Staff worried about future references will give you nothing useful.

The 9-Question Template

Three years of refinement narrowed the template to these:

• 1. If you had to pick ONE reason, what is it? Forces specificity past the safe answer.
• 2. Was your pay fair for the workload? What does the new job pay? Calibrates market data.
• 3. What specifically is better at the new job? Reject "just better" — get hours, breaks, management, team.
• 4. Name two people here you had friction with (roles fine if no names).
• 5. Three concrete things we could fix. Not "communication" — "publish rota one week ahead."
• 6. What would have to change for you to come back? If pay, ask the number.
• 7. Would you recommend this job to a friend? Why? The eNPS question.
• 8. What do you wish someone had told you in your first 30 days?
• 9. What's one thing you didn't say while working here that you regret? The unlock question.

What the Data Revealed

Of 87 leavers, 41% said "chaotic schedule" at question five — only 6% mentioned it at question one. 28% named the same mid-level manager (question 4). That manager went into coaching; the next year turnover halved at that branch.

thMenu's Staff module includes an Exit Notes field per departing employee. Structured answers feed a six-month "Top Pain Points" report that surfaces recurring themes management would otherwise miss.

FAQ

When should the interview happen? Mid-final-week, not last day. The person has relaxed but observations are still fresh.

Written or verbal? Both. Written first (9 questions), then a 20-minute follow-up to dig deeper.

Where do we keep the data? Anonymized by role, reviewed quarterly. The pattern matters, not the individual.

Found this helpful? Share it.