How do you interview customers usefully?
By asking about what they actually did, not what they think or would do — because people are unreliable narrators of their own future behaviour and reliable reporters of their past.
The central rule. "Would you use a feature that did X?" produces a polite yes that predicts nothing. "Tell me about the last time you had this problem" produces evidence. Past behaviour is data; hypothetical preference is noise.
Questions that work:
"Walk me through the last time this happened." Specific, recent, concrete.
"What did you do about it?" Including nothing, which is informative.
"What were you using before, and why did you stop?"
"What made you start looking for a solution that week?" — the trigger, which is frequently the most commercially valuable thing you will learn, because it tells you when to reach people.
"How are you handling it now?" — the honest competitive set.
"What have you tried that did not work?"
"How much time or money does this cost you?"
What to avoid:
Pitching. The moment you describe your solution, the interview becomes a sales conversation and honest feedback ends.
Leading questions, which people answer agreeably.
Asking about price hypothetically, which produces answers unrelated to behaviour.
Talking. Aim to speak for a small minority of the time. Silence after an answer produces the most valuable material, because people fill it with the qualification they were holding back.
Accepting the first answer, rather than asking why two or three times.
Who to talk to: current customers, recent churners, people who evaluated and chose otherwise — the most useful and least contacted group — and people in the target market who have not engaged at all.
How many. Patterns emerge quickly for qualitative work; five to twelve conversations with a consistent segment usually surfaces the major themes, with diminishing returns after.
Record and transcribe with consent, tag the transcripts, and share quotations rather than summaries — summaries lose the specificity that makes findings persuasive.