Exit interviews vs. cancellation surveys: what each one tells you

Both methods try to answer the same question: why did this customer leave? They get very different kinds of answer, from different people, at different moments. Choosing between them is less about which is "better" and more about which question you most need answered right now.

What each one is

A cancellation survey is a short questionnaire shown during the cancellation itself, usually a list of reasons with an optional comment box. Stripe's hosted customer portal, for example, can show seven fixed reasons plus "Other" (opens in a new tab), with free text only when the customer picks Other. Dedicated cancel-flow tools let you customise the list and often attach a retention offer to particular answers.

An exit interview is a conversation after the decision. It might be a plain email that asks an open question, an AI voice call, a chat, or a scheduled call with a founder. What makes it an interview rather than a survey is that the next question depends on the last answer.

What a cancellation survey does well

• Coverage. Because it sits in the path the customer is already on, it hears from far more of the people who cancel than anything sent afterwards.
• Comparable data. The same list month after month gives you a clean trend line. If "too expensive" doubles after a price change, you'll see it.
• Low effort. There's nothing to read or tag. The output is already a chart.
• Immediate action. If your goal is to save the customer, the survey is the only method that runs before the cancellation completes, which is why cancel-flow tools pair it with offers.

Where cancellation surveys fall short

• You can only get answers you thought of. A list of reasons is your current set of guesses. The reason you haven't considered can't be picked from it, and the "Other" box is easy to skip.
• The answers arrive under time pressure. The customer is trying to finish leaving. They'll pick the nearest option rather than the accurate one.
• There's no second question. "Missing features" stays "missing features" because nothing can ask which one.
• It can become friction. A survey that's mandatory, long, or combined with offers that are hard to decline makes cancelling harder. That damages trust and draws scrutiny, so keep it short and skippable.

What an exit interview does well

• It finds reasons you didn't know to ask about. An open question has no list to constrain the answer.
• It can follow up. The second answer is often the useful one, and only a conversation can ask for it.
• There's no pressure. The decision has been made, nobody is trying to save the account, and that tends to produce more candid answers.
• It sounds like you. An email from a founder's address, or a well-run call, feels like being asked rather than being processed.

Where exit interviews fall short

• Fewer people answer. Anything that arrives after the cancellation competes with everything else in their day. Customers who quietly drifted away are the least likely to reply, so expect gaps.
• Answers take time to arrive. Replies come in over hours or days, not at the moment of cancellation, so they can't trigger a save.
• Free text takes work. Someone, or something, has to read and categorise the answers before you can see a pattern.
• Each channel has its own costs. Email has to reach the inbox and not look like marketing. A phone call needs a number on file and the customer's consent to talk. A scheduled interview needs a person's time on both sides.

One follow-up question, two examples

Here's what the difference looks like in practice. These are illustrative, but the shape is very common.

"Missing features"

A survey records "Missing features" and stops. An interview asks which one, and what they were trying to do when they noticed. The reply:

Now you know the feature (an accounting integration), the job (month-end reporting), the cost of the gap (an afternoon a month) and what replaced you (a tool that syncs). That's a roadmap item with a clear success test, not just a tally mark.

"Not using it enough"

A survey records low usage. An interview asks whether something changed or it never quite stuck. The reply:

That isn't a usage problem; it's an adoption problem. The product depended on one person. The fixes, such as onboarding a second user or handing over ownership when an admin leaves, are completely different from anything you'd do about "low usage".

Using both

The methods complement each other well. The survey gives you the distribution, meaning how often each known reason comes up. The interview gives you the meaning, meaning what those reasons actually are for real customers. The best setups join them by customer, so the button answer and the email reply sit on the same record rather than in two tools.

Two rules make the combination work. Keep the survey short and skippable, so the cancel path stays clean. And don't ask the same question twice: if the survey already got "too expensive", the interview should open by following up on that, not by asking why they left all over again. In our product the in-app Cancellation Capture and the email exit interview merge into one record for exactly this reason.

Which should you start with?

• If you have few cancellations and a close relationship with customers, start with an exit interview, even one you send by hand. Every answer is valuable, and depth beats coverage when the numbers are small.
• If you have hundreds of cancellations a month and mostly need to track known reasons, start with a survey. You'll want the trend line, and you can add interviews for the segments that matter most.
• If your main goal is saving customers at the moment they cancel, you need a cancel flow with a survey, because it's the only method that runs in time. Just keep it honest and easy to complete.
• If you mainly want to know what to build or fix, prioritise interviews. A tally of known reasons rarely changes a roadmap. A specific answer to a follow-up often does.

For the tools on each side, our comparison of SaaS churn tools groups them by exactly this distinction, and WhyLeft vs. Churnkey looks closely at one product from each approach.

Frequently asked questions

Do exit interviews get more responses than cancellation surveys?

Usually not in raw numbers. A survey in the cancel flow is seen by almost everyone who cancels through it, while an interview depends on people choosing to reply. The fair comparison is usable answers: a survey response that says only "Other" tells you less than an interview reply that names a specific gap. Measure both on your own customers before concluding anything.

What should I ask in an exit interview?

One open question to start, such as "What made you decide to cancel?", sent soon after they leave and from a real person's address. Then one follow-up based on their answer, asking about a specific moment or task. Don't ask them to rate anything, and don't pitch them in the same message.

Should a cancellation survey be mandatory?

We'd argue no. A required question adds friction at exactly the moment a customer is trying to leave, and forced answers are the least reliable kind. Making it optional costs you some coverage and keeps both the data and the relationship more honest.

Can AI run an exit interview?

Yes. Several tools now use AI to write the follow-up question, hold a text or voice conversation, or tag the answers. The useful test is whether the follow-up responds to what this particular customer said, rather than branching from a fixed script. Customers should also be able to tell they're talking to software when that's the case.