By the Wavefield Research team · Published Nov 9, 2026
An exit survey is a short standardized questionnaire for people who are leaving: employees departing a job, or customers cancelling a product. It asks why they left, what the experience was like, and what might have kept them, in a fixed format whose answers can be compared across leavers and trended over time.
The pages ranking for this term are HR glossaries that define the thing and stop. This one covers what they skip: the difference between an exit survey and an exit interview, complete question sets for both departing employees and cancelling customers, and the timing and privacy rules that decide whether the answers can be trusted.
The two get used interchangeably, and they should not be. An exit interview is a conversation: an HR partner, a departing employee, an hour, and a write-up filtered through what both sides were comfortable saying out loud. An exit survey is an instrument: fixed questions, fixed scales, completed privately. Ten exit interviews produce ten documents nobody can add up. Ten exit surveys produce a distribution: 40% cited compensation, manager-related exits doubled in the Northeast, the would-return share fell nine points. Most lists of exit interview questions circulating online convert directly into survey form; the conversion is the upgrade, because it swaps an anecdote generator for a measurement.
The same instrument logic serves two different audiences, and this guide covers both. The employee exit survey belongs to HR, sits beside the engagement and pulse survey program, and explains attrition. The customer exit survey, often called a cancellation or churn survey, belongs to the product team and explains lost revenue. Different owners, same discipline: few questions, fixed categories, trended over time.
The scale of the blind spot is public data: 3.4 million Americans quit a job in September 2026 by the Bureau of Labor Statistics' count, and most of their employers recorded the reason nowhere more structured than a goodbye email. The practice is institutional where attrition is taken seriously: Washington State runs a standing statewide exit survey for departing state employees, a 10 to 15 minute instrument run by its Office of Financial Management expressly to learn why people leave state employment.
A working employee exit survey needs four blocks, about ten questions in all. These exit survey examples are phrased to be fielded as written:
Start with a single-select: what was the primary reason you decided to leave? Offer the honest option list (compensation, career growth, manager, workload, flexibility, personal circumstances, a better offer) plus Other. Then ask the counterfactual as an open end: what would have had to change for you to stay? The first question makes the data countable; the second catches the reasons your option list never imagined.
Four to six statements rated on the same 5-point agreement scale: I was paid fairly for my work. My manager supported my development. I had real opportunities to grow here. My workload was sustainable. Fixed statements are what make this quarter's leavers comparable to last quarter's, which is the entire advantage a survey holds over a conversation.
Two short questions: would you recommend this organization as a place to work (0 to 10, the eNPS framing), and would you consider returning in the future? Boomerang hires and referrals from alumni are real recruiting channels, and these two numbers tell you whether a departure closed the door or left it open.
What did we do well? What should the next person in your role know on day one? Departing people are the only respondents with nothing to lose, which makes these two answers some of the most candid text an organization ever collects. Budget real time to read them.
The scale statements should reuse the wording of your employee survey questions wherever the constructs overlap, because the comparison between what leavers said and what stayers say on the same items is the sharpest analysis this data supports.
The customer version has one structural constraint the employee version does not: it interrupts someone who is actively trying to leave. Nobody completes a twenty-question cancellation survey except the person who approved it. Four questions, one screen, every one skippable:
One single-select at the moment of cancellation: too expensive, missing a feature I need, not using it enough, switching to another product, a temporary pause. Mutually exclusive options, one screen, skippable. The distribution of this one answer is the churn report.
If they chose switching, ask where. A write-in works better than a brand list, because the competitors you fear are rarely the complete set of places customers actually go, and the surprises are the point.
What is the one thing we could have done differently? Open end, optional. Cancelling customers write short, specific answers, and the recurring phrases across a quarter of these are a product roadmap written by the people who gave up on you.
What would bring you back? This is the winback list builder. Pair it with a checkbox consent to be contacted when the named gap is fixed, and the exit survey doubles as a re-acquisition channel.
Two integrity rules. The survey never blocks the cancellation; a form that holds the exit hostage collects resentment, not data, and regulators increasingly read it as a dark pattern. And the reason categories stay fixed across quarters, because the moment someone improves the wording, the churn trend line restarts at zero. The same fixed-category discipline runs through the broader customer feedback survey program the exit survey belongs to.
A sample exit survey for employees, in fielding order: one screen of introduction naming who sees the data and which privacy promise applies; the primary-reason single select; the what-would-have-changed open end; six scale statements on pay, manager, growth, workload, recognition, and flexibility; the 0 to 10 recommend question; the would-you-return question; two closing open ends. Ten questions, seven minutes, and every item maps to a number someone will actually trend.
The customer exit survey template is shorter still: primary reason, destination if switching, the one-thing open end, what would bring you back. If you build either in Wavefield, describe the study in plain language and the AI agent drafts the full questionnaire with the response options, skip logic, and the confidentiality intro included, ready to edit before anything fields.
For employees, field between the resignation notice and the last day, when memory is fresh and the person still opens company email. A follow-up two to four weeks after departure catches more candor at the price of lower response. For customers, ask at the cancellation moment itself, one screen, and never as a gate the cancellation hides behind.
Confidential and anonymous are different promises. If responses carry a name or a token, say confidential and name who sees the raw data. And respect small-base leakage: in a twelve-person team, the person who left in March is identifiable no matter what the form promised, so aggregate before sharing anything with managers.
Leavers owe you nothing. A focused exit survey runs five to ten minutes; Washington State budgets 10 to 15 for its statewide instrument, and that is the ceiling for an audience with no stake left. Every question past the core costs completes from the one population you cannot re-survey.
One exit survey is an anecdote with checkboxes. The instrument earns its keep in aggregate: code the reasons, trend them by quarter and by team, and put the result next to your engagement and pulse data. A spike in manager-related exits from one department is a finding; a single angry questionnaire is a Tuesday.
Response rates deserve one extra sentence: expect lower numbers than an all-staff census, treat the gap as information about trust, and read survey response rates before deciding whether 40% of leavers is a sample or a symptom.
An exit survey is a short standardized questionnaire given to people who are leaving: employees departing a company, or customers cancelling a product or service. It asks why they left, what their experience was like, and what might have kept them, in a fixed format whose answers can be counted, compared across leavers, and trended over time.
The exit interview is a conversation, usually between the departing employee and HR, and it produces a narrative shaped by politeness and by whoever ran it. The exit survey is a written instrument with fixed questions and scales, so ten responses produce a distribution instead of ten anecdotes. Many organizations run both: the survey for comparable data, the interview for the stories behind it. Most published exit interview questions work fine in survey form once you give them fixed response options.
Only if they are actually built that way, and often they should not be. A truly anonymous exit survey carries no identifier at all, which also means nobody can follow up on a serious allegation it contains. The more common honest design is confidential: responses are linked to the person, raw data is restricted to one or two owners, and everything shared beyond them is aggregated. Whatever the design, the introduction must say which one it is, and small teams need aggregation regardless, because a team of twelve has no anonymous leavers.
For employees, the reliable window is between resignation notice and the last working day, while company email still gets opened. Some organizations send a second, shorter survey two to four weeks after departure, when candor rises and response rates fall. For customers, the exit survey belongs at the cancellation moment itself, kept to one skippable screen, optionally followed by a brief email survey a few weeks later for the fuller story.
A sample exit survey for employees: primary reason for leaving (single select), what would have changed your decision (open end), four to six fixed statements on pay, manager support, growth, and workload (5-point scale), would you recommend us as a place to work (0 to 10), would you consider returning (yes, maybe, no), and two open ends on what worked and what the next person should know. That is ten questions and well under ten minutes.
Only when someone owns the readout. The failure mode is collecting exit data into a folder nobody opens, which turns the whole exercise into a filing cabinet with better fonts. The organizations that get value assign an owner, code the reasons quarterly, present the trend beside engagement and pulse results, and track one number over time: the share of regrettable departures whose stated reason was something the organization controls.
Related: employee engagement surveys · pulse surveys · customer feedback surveys · survey response rates
Describe your exit survey in plain language and the AI agent builds the questionnaire, logic, and confidentiality intro. Per-project pricing from $99, no subscription.
Start a study