Outgrace, home
People & CultureHow-to guide

How to run exit interviews that expose real retention risks

A repeatable process, a fixed set of questions and a coding habit can turn departures into a short list of fixes. Here is how to build one.

A resignation tells you that someone has left. It rarely tells you why, and when the reason is finally given it is often the version that is easiest to say in a final week. In our experience the exit interview is the cheapest research a company runs and the least used. Most organisations ask a few questions in a hurry, file the notes in a folder nobody opens, and move on. This guide sets out a process that produces answers you can compare across leavers, code consistently, and turn into a short list of fixes a leadership team will actually act on. It is written as practitioner judgement, not as a summary of published research, and it should be read that way.

Before you start

Three things must be settled before the first interview. The first is ownership. One named person, usually in people operations rather than the leaver's line manager, owns the process and its data. A manager interviewing their own departing report tends to get polite answers, because the manager is also the person the leaver may be criticising. The second is confidentiality, stated in writing: who sees individual notes, and how comments will be reported. The third is a decision about what the company will do with the results. If nobody senior has agreed to act on a finding, the programme is a survey with a lower response rate and a sense of futility attached to it.

Before the first leaver walks out of the building, a small group should agree four things in writing. Keep the list short enough to fit on one page.

  • A named owner outside the leaver's reporting line, with authority to report findings to the leadership team.
  • A written statement on confidentiality: who reads individual notes and the smallest group size that will ever be reported.
  • An agreed review date each quarter at which a senior leader will read the themes and decide what to do.
  • A fixed list of leaving reasons, agreed before the first interview, so that answers can be counted.
  1. Step 1

    Fix the timing

    Ask for the interview after notice has been given but before the final week. A leaver who has already started a new job often gives guarded answers, and one who has stopped coming in may have checked out. In our experience the period between notice and the last fortnight produces the most useful and least rehearsed answers. What usually goes wrong is that the interview is booked on the last afternoon, when the leaver has no time to think and the interviewer has no time to probe.

  2. Step 2

    Use the same core questions every time

    Write eight to ten questions that cover the reason for leaving, what would have been needed to stay, the quality of day-to-day management, workload, pay and progression, and whether the leaver would recommend the company to a friend in a similar role. Use the same wording for every leaver. What usually goes wrong is that interviewers improvise, the questions drift from one conversation to the next, and the answers cannot be compared when the year ends.

  3. Step 3

    Offer a written alternative

    Some people will not say a difficult thing aloud, however kindly they are asked. Offer a short written form with the same core questions, returnable by post or by a link to the owner rather than to the leaver's manager. What usually goes wrong is that the written answers come back as single words or blank boxes. A form works best when it asks for one example for each question, not for a general opinion.

  4. Step 4

    Record the reason in two parts

    Ask the leaver to choose a main reason from a fixed list, then to describe it in their own words. The list gives you counts and the words give you the cause. Include an other category and review its contents each month, because a rising pile of other usually means the list is missing a real reason. What usually goes wrong is that the list is written by the leadership team to flatter the company, and leavers quickly learn to pick the most neutral option.

  5. Step 5

    Code the answers against a codebook

    Write the codebook before coding starts. Each theme needs a name, a one-sentence definition, a note on what does not count, and one example from a real answer, anonymised. Then have two people code the same ten answers independently and compare their results. Where they disagree, change the definition, not the answer. Date the codebook and keep the versions. What usually goes wrong is that the codebook is written after the coding has begun, so that themes quietly shift to fit the data.

  6. Step 6

    Look for clusters, not single cases

    Group the coded answers by team, manager, tenure band, location and role level. A single resignation is a story and deserves a conversation. Three leavers from the same team within a year, citing the same theme, is a pattern worth a meeting with the manager and the people lead. Set the threshold yourself and write it down before you look at the data. What usually goes wrong is that leaders react to the most vivid exit in the last month and ignore the quieter, repeated complaints from the year be

  7. Step 7

    Turn themes into a short list of fixes

    Take no more than three themes into each quarterly review. For each one, write down the fix, a named owner, a rough estimate of the effort it needs and the measure that will show whether it worked. Ask a plain question before agreeing anything: in six months, would we know whether this had changed? If the answer is no, the fix is probably too vague. What usually goes wrong is a list of ten actions that nobody owns, which produces a sense of progress and no change at all.

  8. Step 8

    Report back to managers and to the business

    Give managers the aggregate themes for their own areas, with the same confidentiality rules applied. Once a year, or after a significant change, tell the wider business what the exit interviews found and what has been done about it. What usually goes wrong is silence. When nothing visibly changes, leavers stop answering honestly and current staff conclude that the process is for appearances.

Where this goes wrong

The most common failure is treating the exit interview as a complaints desk. Leavers who have decided to go often have a grievance, and the interview becomes a place to vent. That can be valuable, but only if the notes are coded like any other evidence. The second failure is letting the data sit with the people who are most exposed to it. When a line manager reads the themes about their own team, the temptation to explain them away is strong, and a sensible owner keeps that reading in a separate step. The third failure is measuring the programme by its response rate alone. A high response rate with bland answers tells you that people found the process safe and had nothing to say, or that they did not believe it. A lower response rate with specific, uncomfortable answers is often the better result. Look at the quality of the comments as well as the count.

A fourth problem is confusing a stated reason with a cause. A leaver who chooses pay as the main reason may have been paid at the market rate and frustrated by a manager who never explained the bands. The reason is the start of the question, not the answer. Ask a follow-up in the interview, or in the written form, for one example of what happened before the decision to leave.

Should the line manager ever run the exit interview?

Rarely, and only with a second person present who is not in the reporting line. The manager can explain what happens next and thank the leaver, but the detailed questions work better with someone who has no reason to defend a decision. Where a small company has no one else available, the owner can be a senior leader from another team, provided the confidentiality rules are the same.

How many leavers do we need before the data means anything?

There is no fixed number that works for every company, and we do not have a published source to offer one. In practice a small firm with a handful of leavers a year will see patterns slowly. Treat the first year as a starting point, look across two or three years before acting on a theme that only appears once, and be wary of any conclusion drawn from a single team with very few departures.

Should leavers be anonymous?

Keep individual records confidential to a named owner, not anonymous in the sense of having no record at all. Report only groups large enough that a leaver cannot be identified from them. Choose the minimum group size yourself, write it into the confidentiality statement and apply it every time, including when a result is uncomfortable for leaders.

Does the process need legal review?

Yes, before it starts. Where exit records contain personal data, the rules on how long they may be kept, who may see them and what the leaver must be told differ by country and by sector. An employment lawyer or a data protection adviser in each relevant jurisdiction should confirm the approach before the first interview is booked.

This week, do three things. Name the owner and write the one-page confidentiality statement. Draft the eight to ten core questions and the fixed list of leaving reasons, then test both on two recent leavers if they are willing to talk. Finally, put a fifteen-minute slot in the next leadership meeting to agree who will read the first quarterly summary and what they will do with it. None of this is large. The value comes from the discipline of doing it the same way each time, so that the answer you get next year can be set beside the answer you got this year.

Newsletter

The Briefing

One email a week: the leaders we profiled, the ideas worth stealing, and nothing else.

Weekly, free, one click to leave. Read our privacy policy.