A customer interview is a structured one-on-one conversation that explains what your metrics can only report: why customers buy, stay or leave. Where a customer survey measures the distribution of known answers across many people, an interview discovers answers you did not know existed — in the words customers actually use, which later become your best marketing copy.
When an Interview Beats a Survey
Short answer: use an interview whenever you do not yet know what the possible answers are. A rating grid can tell you support scores a 3.2; it cannot tell you that customers churn because your invoicing confuses their accounting department — nobody put that option in the dropdown. The three occasions where interviews earn their cost: shortly after purchase (motivation and friction are fresh), at defined relationship moments (renewal, project completion — see the B2B questionnaire guide), and at cancellation, where a single honest conversation is worth more than a year of dashboards.
The 5 Rules That Keep Answers Honest
Ban "would". Every question containing "would you" invites polite fiction. Replace it with "when did you last" — past tense forces facts. This is the core insight of Rob Fitzpatrick's The Mom Test, the standard reference for interview technique.
Their life, not your product. Spend the first half of the conversation on how they work and what the problem costs them. The product only enters in the second half.
Dig on specifics, not opinions. "The onboarding was confusing" is an opinion. "At which step did you get stuck?" turns it into a fact you can fix.
Embrace the pause. After an answer, count to three before speaking. The sentence that comes after the silence is routinely the most honest one of the interview.
Never defend, never pitch. The moment you explain why something works the way it does, the interview is over — the customer starts sparing your feelings.
After Purchase: 10 Questions
Send or schedule this within the first two weeks after purchase, while the decision process is still vivid. Questions 3 and 4 are the ones teams skip and regret: the almost-objection is your conversion copy, and the alternatives question maps your real competition.
| # | Question | What it reveals |
|---|---|---|
| 1 | What was happening in your work when you started looking for a solution? | The trigger moment — your real marketing timing |
| 2 | How did you handle this before? | The workaround you are actually replacing |
| 3 | What almost stopped you from buying? | The objection your website fails to answer |
| 4 | What alternatives did you look at, and why not those? | Your real competitive set and your edge |
| 5 | Walk me through your first hour with the product. | Onboarding friction, step by step |
| 6 | What surprised you — positively or negatively? | Expectation gaps in both directions |
| 7 | When did you last get stuck, and what did you do? | Support gaps and self-help behavior |
| 8 | What has changed in your day-to-day since? | The actual delivered value, in their words |
| 9 | If we removed the product tomorrow, what would you do? | Dependence and true substitutes |
| 10 | Whom would you describe this to, and how? | Referral language for your positioning |
Question 9 is the honest version of the NPS question. "What would you do without it" reveals attachment through consequences rather than a number. If the answer is a shrug, no score will save the renewal. For the metric itself, see the NPS survey guide.
Existing Customers: 8 Questions
1. What do you use most often — and what have you never opened?
2. When did the product last save you real time? What happened?
3. When did it last cost you time or nerves? What happened?
4. What do you do right before and right after using it? (Reveals workflow gaps and integration needs.)
5. What have you built around the product — spreadsheets, scripts, rituals?
6. Which colleague complains about it, and what do they say?
7. What would need to happen for you to pay twice the price without hesitating?
8. If you were in charge here, what would you fix first?
Question 5 is the quiet star of this set: every spreadsheet a customer maintains next to your product is a feature request they never filed. Collect these workarounds across ten interviews and you have a roadmap grounded in observed behavior rather than upvotes.
At Churn: 7 Questions
1. When did you start thinking about cancelling? What happened around then? (The real reason lives months before the cancellation.)
2. Was there a specific moment that settled the decision?
3. What did you try before deciding to leave?
4. What are you switching to, and what convinced you there?
5. What did we never ask you that we should have?
6. What would have had to change six months ago for you to stay?
7. Under what circumstances would you come back?
Three churn interview rules: not run by the account manager who owned the relationship, one to two weeks after the cancellation, and with an explicit opening line that this is not a win-back call. Break any of the three and you collect politeness instead of reasons.
Running the Conversation
Set the frame in 30 seconds
Purpose, duration, and the sentence that changes everything: "Critical answers help us more than kind ones."
Open with their world
First questions about their role and workflow. People talk more freely about their work than about your product — and the context makes every later answer interpretable.
Follow the guide loosely, the answers closely
The question set is a checklist, not a script. When an answer surprises you, dig there — the surprise is the finding.
Capture verbatim quotes
Record with consent or note exact phrases. "It just eats my Monday morning" is usable; your paraphrase of it is not.
Debrief within 24 hours
Write the three strongest findings while they are fresh, tag recurring themes across interviews, and count them — themes, not anecdotes, decide the roadmap.
Scaling Beyond 10 Calls: AI-Led Interviews
The honest constraint: an hour of scheduling, conversation and debrief per interview means nobody interviews more than a fraction of their customers. The classic compromise is falling back to a static form — which removes exactly the follow-up questions that made the interview valuable. The newer option keeps the conversation and removes the calendar: an AI-led interview runs your question guide as a real dialogue — one question at a time, follow-ups where answers stay vague, in whatever language the customer writes. Every buyer gets interviewed instead of a sample of ten, and the answers come back summarized by theme with the strongest quotes attached. Use the human hour where it beats the machine: deep-dives on the themes the scaled interviews surface.
Turn this question set into an AI interview
Paste your questions, get a share link. The AI asks them one at a time, follows up on vague answers, and summarizes themes across all interviews. No account for customers, EU-hosted, free for 14 days.
The short version. Ban "would" — ask about the past, where facts live. Spend the first half on their life, not your product. Run 5-10 interviews per segment for themes, then scale with AI-led interviews so every customer gets asked, not just the sample. Count themes across interviews; never let one vivid anecdote set the roadmap. And at churn: neutral interviewer, 1-2 weeks after notice, explicitly no win-back.



![NPS Survey: Create, Calculate and Actually Learn Something [2026]](https://www.teamazing.com/wp-content/uploads/2026/07/nps-survey-feedback-tool.jpg)
![GDPR-Compliant Survey Tools: The Checklist That Actually Matters [2026]](https://www.teamazing.com/wp-content/uploads/2026/07/gdpr-surveys-and-tools.jpg)
![eNPS: How to Calculate the Employee Net Promoter Score (Formula + Benchmarks) [2026]](https://www.teamazing.com/wp-content/uploads/2026/04/enps-employee-loyalty.jpg)