# The 6-Question Churn Call Script A working script for a 12-minute call with a customer who just cancelled. From saasfeedback.ai — https://saasfeedback.ai/blog/questions-to-ask-churned-customers Licence: use it, edit it, ship it. No attribution required. --- ## Before you dial - [ ] Pull their account: signup date, plan, MRR, last login, last 3 support tickets. - [ ] Note the one thing you already suspect. Write it down. You are testing it, not confirming it. - [ ] Call within 7 days of the cancellation. After 14 days the memory is a summary, not a story. - [ ] Never let the account owner of the relationship run the call. People soften bad news for someone they like. ## The opening (30 seconds, say it almost verbatim) > "Hi [name], this is [you] from [product]. I'm not calling to win you back and there's no > offer at the end of this — I promise. You cancelled last week and I'd rather learn something > from it than guess. Ten minutes, and I'll take whatever you give me. Is now bad?" Two things are doing the work here: naming the absence of a save attempt, and giving them an easy out. If they say now is bad, book a time. Do not push through. --- ## Question 1 — Establish the timeline **Ask:** "Walk me back — when did you first think about cancelling?" **Why:** Churn is a decision with a date, and that date is almost never the cancellation date. The gap between the two is where your intervention window lives. **Follow-up:** "What was happening that week?" **Failure mode:** Asking "why did you cancel" first. You get the rehearsed answer — the one they already told their boss — and the rest of the call defends it. --- ## Question 2 — Get the concrete last straw **Ask:** "What was the last thing that happened before you decided?" **Why:** Reasons are abstractions; events are specific. "It was too slow" is a reason. "I waited 40 seconds for the export on Tuesday and my client was on the call" is an event you can go fix. **Follow-up:** "Was that the first time, or had it been building?" **Failure mode:** Accepting the abstraction. If you hear a category ("too expensive", "missing features", "too complicated"), you have not finished this question yet. --- ## Question 3 — Find the replacement **Ask:** "What are you using instead?" **Why:** The answer is rarely a competitor. It is often a spreadsheet, an intern, or nothing at all — and each of those means something completely different about your positioning. **Follow-ups:** - If a competitor: "What does it do that we didn't?" - If a spreadsheet or manual process: "What made that good enough?" - If nothing: "What changed so that you didn't need this any more?" **Failure mode:** Getting defensive about the competitor. You are collecting, not selling. --- ## Question 4 — Locate the value gap **Ask:** "When you signed up, what did you expect this to do for you?" **Why:** Most churn is the distance between the promise they heard and the thing they got. This question measures that distance. Answers here usually indict marketing or onboarding, not the product. **Follow-up:** "Did that ever happen? When?" **Failure mode:** Correcting them. If they expected something you never claimed, that is a positioning bug and it is yours. --- ## Question 5 — The counterfactual **Ask:** "What would have had to be true for you to still be a customer today?" **Why:** This is the roadmap question. Phrased as a counterfactual rather than "what should we build", it gets you a condition rather than a feature request — and conditions are what you can actually prioritise. **Follow-up:** "And if we'd shipped that six months ago, would you have stayed, or was something else already in motion?" **Failure mode:** Treating the answer as a commitment. They are not promising to come back. They are telling you what mattered. --- ## Question 6 — The open door **Ask:** "Anything else you'd want the person who builds this to hear?" **Why:** The most useful sentence of the call lands here more often than anywhere else. The framing gives them permission to say the thing that didn't fit the other five questions. **Follow-up:** Silence. Count to five. Do not fill it. **Failure mode:** Wrapping up before asking. Never skip question 6 because you are at time. Skip question 4 instead. --- ## The close > "That's genuinely useful, thank you. If we fix [the specific thing they said], can I email > you and tell you? No pitch — just so you know it wasn't wasted." Roughly one in five say yes. Those are your win-back list, and they convert far better than a discount campaign because you are returning with evidence. --- ## After the call (do this same day) | Field | What goes in it | |---|---| | Customer | Name, plan, MRR | | Decision date | From Q1, not the cancellation date | | Trigger event | The concrete thing from Q2 | | Replacement | Competitor / manual / nothing | | Expectation gap | Q4, in their words | | Counterfactual | Q5, in their words | | One verbatim | The single best sentence, quoted exactly | | Theme tag | Your existing taxonomy | | MRR at risk | This customer, plus similar accounts still active | Keep the verbatim column. Aggregated themes persuade nobody in a roadmap meeting; one exact sentence from a real customer ends the argument. --- ## The rules that make this work 1. **Ask about behaviour, not opinions.** "What did you do" beats "what do you think". 2. **Never pitch.** The moment they smell a save attempt, the honest answers stop. 3. **Silence is a tool.** The second half of an answer is where the truth usually is. 4. **Ten calls beats a hundred survey responses.** Stop when new calls stop surprising you — usually somewhere between 8 and 15. 5. **Record it, with permission.** Notes are lossy; you will lose the exact words, and the exact words are the point.