customer-interview-guide

Use this when messaging, ICP or roadmap decisions rest on guesses. Ten good interviews with recent buyers usually reveal more than a hundred survey responses.

Inputs

• The decision the research should inform (positioning, ICP, pricing, a feature).
• Who to talk to: recent buyers (last 90 days) are best; also recent churns and lost deals.
• What you already believe, written down, so you can test it.

The play

1. Recruit the right people

Recent buyers remember the story. Aim for 8 to 12 interviews across segments. Recruit with a short, honest ask:

Hi Maya, you started with us last month. Could I ask you 20 minutes of questions about how you decided? No sales pitch, just learning. As thanks, <a small, appropriate gesture>.

2. Ask about the story, not opinions

Follow the timeline of the purchase:

1. Before: "What were you doing before? What was annoying about it?"
2. Trigger: "What happened that made you start looking then?"
3. Search: "What else did you look at? What did you almost choose?"
4. Decision: "What made you pick this? Who else had a say? What almost stopped it?"
5. After: "What's different now? What would you miss if it disappeared?"

Read references/interview-guide.md for the full guide with follow-ups and the synthesis template.

3. Avoid leading questions

Not "Would you say our reporting saved you time?" but "What happened the last time you needed a report?" Ask for specific past events, not hypothetical future behaviour.

4. Capture the exact words

Record (with permission) and pull verbatim quotes. The buyer's phrases are the raw material for headlines, emails and battlecards.

5. Synthesise into decisions

Group findings into triggers, alternatives, decision criteria, objections and outcomes. A theme needs 3+ interviews. Each theme ends with a recommended change.

What good looks like

• Interviewees do most of the talking and tell stories with specific moments.
• The synthesis includes surprises, not just confirmation.
• Buyer phrases end up in the next round of copy.
• The common mistake is turning the interview into a demo or a satisfaction survey, which produces polite praise and no insight.

Rules

• MUST get permission before recording and anonymise quotes unless the person agrees otherwise.
• MUST ask about past events, not hypothetical opinions.
• NEVER sell during a research interview.
• NEVER present a single interview as a pattern.

Example prompts

• "Write a customer interview guide to find out why our last 20 customers bought."
• "Here are 9 interview transcripts. Synthesise themes and buyer language for our homepage."
• "Write a recruit email for interviews with customers who churned last quarter."