How to Run Customer Discovery Interviews That Actually Teach You Something
Unicorn Rivals Team
Why Interviews Fail
Founders ask: "Would you use this?" Customers say: "Sure, sounds cool." Nobody learns anything.
Customer discovery isn't sales. It's structured listening to understand whether a problem is urgent, frequent, and expensive enough to build a business around. If you're new to startups, read What Is a Startup? first — discovery is how you test the "unproven model" part.
Pair interviews with the 7-day validation checklist before you commit to an MVP scope.
Who to Talk To
Rule: Talk to people who have the problem today, not people who might someday.
| Good | Bad |
|---|---|
| Last 30 days they felt the pain | "My friend might need this" |
| They already pay for a workaround | "Students are our future market" |
| Specific job title / workflow | "Everyone with a phone" |
Aim for 10–15 conversations before you change direction. Fewer than five is anecdotes; more than twenty without a pattern means you're stalling.
The Interview Script (30 Minutes)
1. Context (5 min)
- "Tell me about your role and what a typical week looks like."
- No pitch. You're learning their world.
2. Problem history (10 min)
- "When did you last face [problem]?"
- "What did you do? What did it cost — time, money, reputation?"
- "How often does this happen?"
Listen for past behavior, not future promises.
3. Current solutions (10 min)
- "What tools or hacks do you use now?"
- "What's frustrating about them?"
- "What would make you switch?"
If they haven't tried to solve it, urgency may be low.
4. Close (5 min)
- "Who else should I talk to?"
- Optional: "We're exploring [one sentence]. Can I follow up in two weeks?"
Do not demo unless they ask. Discovery ends when you understand pain, not when they're impressed.
Questions to Avoid
- "Would you pay $X?" — hypothetical pricing lies. Ask what they pay today for alternatives.
- "Do you like my idea?" — compliments aren't data.
- Leading questions — "Don't you hate spreadsheets?" biases answers.
Better: "Walk me through the last time you used a spreadsheet for this."
Taking Notes That Matter
After each call, log:
- Verbatim quote — one sentence in their words
- Severity — 1–5 (how much it hurt)
- Frequency — daily / weekly / rare
- Current spend — money or hours on workarounds
- Surprise — what contradicted your hypothesis
Patterns across calls beat any single enthusiastic yes.
From Interviews to Decisions
| Signal | Action |
|---|---|
| 8+ people describe same pain unprompted | Explore problem-solution fit |
| People pay for ugly workarounds | Strong MVP candidate |
| Polite interest, no past behavior | Keep interviewing or narrow ICP |
| Wrong persona entirely | Pivot the segment, not the feature list |
When you're ready to build, decide B2B vs B2C based on who showed up in interviews, not who you wish would.
Remote and B2B Tips
- Record with permission (Zoom, Grain) — focus on listening
- Ask for process, not opinions: "Show me your last invoice workflow"
- End with two referrals — warm intros beat cold LinkedIn
In Unicorn Rivals, customer acquisition isn't abstract — Growth research and five resources make "did we find the right market?" visible on the leaderboard.
Related reading
Practice listening under pressure
Unicorn Rivals simulates forty founders competing for the same market — customer wins, rival steals, and runway clocks don't pause for perfect interview data.
Curious about Unicorn Rivals?
Join the iOS beta list — early access players get 500 💎 diamonds at launch.
Join Early Access →