Churn Interview Guide: 12 Questions & Notes

A cancellation survey tells you which box a departing subscriber ticked. A churn interview tells you what actually happened — in what order, over how long, and what nearly changed their mind. This churn interview guide gives you the full working kit: who to invite and when, twelve questions in the order to ask them, a dedicated probe set for payment problems that surveys almost always miss, a note-taking template you can copy into your call tool, and the method for coding what you hear back into a shared reason taxonomy. Run even five of these conversations a month and you will know things about your churn that no dashboard will ever surface.

Use this alongside your survey, not instead of it. The survey gives you volume; the interview gives you the story behind the volume. If your survey questions need tightening first, our exit survey questions page is the companion to start from, and our overview of why subscribers churn sets out the broader voluntary and involuntary picture these interviews will keep confirming.

Recruiting the interview: who, when, and what to offer

• Who. Recent cancellations first — people who cancelled in the last 14 days remember the sequence. Mix them deliberately: some long-tenure, some early exits, some high-value plans, some who accepted a save offer and left anyway later, and some whose subscriptions lapsed on failed payments. Do not only interview your happiest or angriest former subscribers; both groups lie to you in opposite directions.
• When. Invite within 48 hours of cancellation, hold the conversation within two weeks. Later than that, memory rewrites the story into a single tidy reason and the useful detail — the failed charge, the confusing email, the moment they stopped logging in — is gone.
• How to ask. One short email from a named person, not a marketing automation. Say plainly: twenty minutes, we are trying to understand what happened, there is no pitch and no save attempt. That last promise matters, and you must keep it — the moment an interview becomes a retention call, every future invite gets ignored.
• Incentive honesty. If you offer a gift card or account credit for their time, say the amount, say it is for the conversation whether they praise you or not, and pay it even if they are blunt. Never tie the incentive to returning, and never offer a bigger incentive mid-call if they say something you like. You are buying candour, not compliments.
• Volume. Five completed interviews a month, coded consistently, beats fifty unstructured chats. Stop a round when three consecutive interviews surface no new reason or moment — that is saturation for that segment, and it is a finding in itself.

The 12 questions, in arc order

Ask them in this order. The arc moves from story to specifics to the counterfactual, and jumping straight to "was it the price?" contaminates everything after it. Let silence work — count to five before you follow up.

1. "Take me back to when you first signed up. What was going on for you then, and what did you hope this would do?" Establishes the original job. Half of churn is a mismatch between this answer and what followed.
2. "How did the first few weeks actually go?" Surfaces onboarding failures while they are still framed as experience, not blame.
3. "When did you first think about cancelling — even briefly? What was happening at that moment?" The single most valuable question in the set. You want the moment, not the summary.
4. "What happened between that moment and actually cancelling?" Reveals the trigger chain: a second failure, an unanswered support ticket, a renewal they did not expect.
5. "Walk me through the day you cancelled. What did you do, step by step?" Exposes billing confusion, cancellation friction, and payment problems in their own sequence.
6. "Was there anything about billing, payments, or renewals that surprised you at any point?" Bridges into the payment probe set below without leading.
7. "If the price had been different, would anything else have had to change too?" Separates a pure price exit from a value exit wearing a price label.
8. "What did you use us for most — and what did you never get round to using?" The gap between those two answers is your activation problem, named by the customer.
9. "Did you talk to our team at any point? How did that go?" Service failures hide here, because subscribers rarely tick "support was bad" on a survey about a product.
10. "What are you using instead now — or are you just going without?" "Going without" is a different competitor than any named rival, and it changes what you build next.
11. "What would have had to be true for you to still be a subscriber today?" The counterfactual. Listen for whether the condition was ever in your control — a feature, a fix, a working card, a pause option.
12. "Is there anything I should have asked you about leaving that I didn't?" Close with this every time. It regularly produces the most quotable, most useful minute of the call.

The payment-problem probe set

Payment-driven churn is chronically under-reported in interviews because subscribers are embarrassed by a declined card or assume the failure was their fault, so they offer a tidier reason instead. Recurly research estimates 20–40% of subscriber churn is involuntary, which means a meaningful share of your interviewees left through the billing door whether they say so or not. When the conversation touches money at all, work through these probes:

• "Do you remember any payment ever not going through?" Then stop talking. Let them reconstruct it.
• "What did you see from us when that happened — an email, a notice in the app, nothing?" "Nothing" is a critical finding: your dunning never reached them.
• "Did you try to update your card? How did that go?" Friction on the update page — a login wall, a confusing form — loses people who were actively trying to pay you.
• "At any point did you decide to just let it lapse rather than fix it?" This is the exact boundary between involuntary and voluntary churn, and only the subscriber can tell you where it sits.
• "If we had paused the subscription instead of cancelling it at that point, would you have come back?" Tests whether a pause offer at the failure moment would have saved the relationship.

Record the answers verbatim. Before you design fixes, size what payment churn costs you with the recovery calculator and set your findings against the failed payment recovery benchmarks — interviews tell you the mechanism, those pages tell you the scale.

Note-taking template

Fill this in during the call, not from memory afterwards. One template per interview, stored where product, support, and lifecycle can all read it.

Field	What to capture
Subscriber snapshot	Plan, tenure, price paid, segment — facts only, pulled before the call
Original job	What they signed up to achieve, in their words (Question 1)
First doubt moment	The moment they first considered leaving, with date or trigger if given (Question 3)
Trigger chain	Everything between the first doubt and the cancellation, in order
Payment history (stated)	Any failure, surprise charge, or card-update friction they describe — verbatim
Payment history (actual)	What your billing record shows for the same period, noted separately after the call
Stated primary reason	Their summary reason, quoted exactly
Counterfactual	What would have had to be true for them to stay (Question 11)
Best verbatim quote	The single most useful sentence from the call, unedited
Interviewer read	Your honest interpretation, clearly labelled as interpretation, not fact

Keeping "stated" and "actual" payment history in separate rows is deliberate. When the two disagree — they remember no failure, your system shows three — that disagreement is itself the insight: your recovery communication never landed.

Coding answers into the reason taxonomy

Within 24 hours of each interview, assign one primary reason code using the same taxonomy your exit survey uses, plus the verbatim quote that justifies it. If the interview surfaces a reason your survey has no code for, flag it rather than forcing a fit — two flags for the same uncoded reason means the taxonomy needs a new code, with a written definition and route. Roll interviews up monthly next to the survey numbers: where the two agree, act with confidence; where they diverge — surveys say price, interviews reveal payment failures and service moments — trust the interviews on mechanism and the surveys on scale, and investigate the gap before spending on either story.

The bottom line

Churn interviews are slow on purpose. Twelve questions, asked in order, to people who left within the last two weeks, with the save pitch explicitly off the table, will tell you the sequence of events that actually ends subscriptions — including the failed payments and billing surprises that subscribers rarely report on a survey. Code every conversation the same day, keep the verbatims, and let the patterns set your product, billing, and recovery priorities. When the payment probes keep surfacing the same failure story, that is your signal to fix recovery properly — start by seeing what it costs and what it includes on our pricing page.

Frequently asked questions

How many churn interviews do we need before the findings are useful?

Five well-recruited interviews in a single segment usually surface the repeating moments and trigger chains. Keep going until three consecutive conversations add no new reason or moment. Depth and consistent coding matter more than raw count, and a small set of comparable notes beats a large pile of unstructured chats.

Should we offer an incentive for a churn interview?

You can, and for busy or higher-value former subscribers it often decides whether they accept. State the amount plainly, pay it for the conversation regardless of what they say, and never connect it to returning or to saying positive things. You are paying for candour and time, not for a save.

Can support run these interviews, or should product?

Either can, as long as the interviewer was not the person who handled the subscriber's cancellation or support case. A neutral interviewer hears more honest answers about service failures. Whoever runs it, the rule is the same: no pitch, no save attempt, and the notes get coded the same day.

Why ask about payments if the subscriber says they left over price?

Because stated reasons and actual histories often differ. A subscriber may summarise a messy ending — failed charges, a surprise renewal, a card they could not update — as simply too expensive. The payment probes check the sequence without accusing them of anything, and the comparison with the billing record shows whether your recovery messages ever reached them.

What do we do with interview findings that contradict our survey?

Treat the divergence as a finding, not an error. Use surveys for scale — how many people pick each reason — and interviews for mechanism — how that reason actually unfolds. When they disagree, investigate before acting: check billing records for the interviewed group and look at whether the survey's answer options are steering people toward the tidy reason.