Most customer discovery interviews are theater. The founder describes their idea, the interviewee politely nods, the founder walks away thinking "great, another validation." Six months later, when the product launches to that same interviewee, the sale doesn't happen. What went wrong?
The interview wasn't discovery. It was a pitch disguised as discovery.
Real customer discovery interviews don't validate your idea. They surface real pain, real workarounds, and real willingness to pay — often revealing that your idea is subtly wrong, that the real pain is elsewhere, or that the customer you targeted isn't quite the right one.
The 25 questions below are the ones I actually use with founders. They're designed to extract signal, not to confirm hypotheses. Use them in the order presented — the sequence matters as much as the questions.
First: the 3 rules that determine whether any interview is useful
Before the questions, three ground rules. Break any of these and you're back to theater.
Rule 1: Don't mention your product.
Not at all. Not "we're building something like this." Not "have you heard of any tools for X?" Not "if there was a product that..." The moment you mention your product, the interviewee shifts from "what's your reality?" to "give the founder useful feedback." Those are different modes and one is worthless for discovery.
Rule 2: Ask about the past, not the future.
"Would you use a tool that..." produces optimism. "Tell me about the last time you..." produces actual behavior. Behavior is data. Opinion is entertainment. Almost every question below is past-focused for this reason.
Rule 3: Follow the emotion, not the checklist.
If the interviewee gets fired up about something you didn't ask about, drop your plan and follow it. That emotion is the signal. Your prepared questions are just a starting point — the best discoveries happen in the tangents.
The 25 questions, in order
Warm-up (Questions 1-3) — build rapport, get honest baseline
1. "Walk me through a typical week in your role." Why: gets them talking about their actual life, not their aspirational life. You'll hear where time actually goes.
2. "What are you working on this quarter that matters most?" Why: reveals what they're actively investing time and attention into. If the problem you're solving isn't near this, you may be targeting the wrong buyer.
3. "What part of your job frustrates you the most right now?" Why: unstructured, invites them to share raw pain without you steering. Whatever they say first is what's most present in their mind.
Discovering the problem (Questions 4-8) — is this pain real?
4. "The last time you had to [do the thing your product solves], walk me through exactly what happened." Why: past-focused, specific, forces them to actually recall reality rather than reason about it. Whatever they describe IS the current workaround.
5. "What was the hardest part of that?" Why: locates the specific pain point in the workflow. Different from "what's frustrating in general" — this pins down the moment where they felt friction.
6. "How often does this happen?" Why: frequency = size of opportunity. A problem that happens once a year is much less valuable than one that happens weekly.
7. "Who else is affected when this happens?" Why: reveals if the pain is theirs alone or spread across a team/company. Multi-person pain = larger deals.
8. "What did you do to try to solve it? What worked, what didn't?" Why: reveals the alternatives they've already tried. If they've tried nothing, the pain might not be real. If they've tried many things without success, there's genuine unsolved space.
Measuring the pain (Questions 9-13) — is it worth paying to solve?
9. "How much time did that whole situation take?" Why: quantifies the cost of the workaround. Hours × their hourly rate = your ceiling for pricing.
10. "Did that cost you or your company anything specific?" (money, deals lost, customers unhappy) Why: connects the problem to business outcomes. Problems tied to revenue or cost are much easier to sell into.
11. "On a scale of 1-10, how frustrating is this problem for you?" followed by "why not lower / why not higher?" Why: the why questions are more informative than the number. If they say "6," ask "why not 3?" to hear what's already tolerable, and "why not 9?" to hear what's not yet a crisis.
12. "If this problem magically disappeared tomorrow, what would you do with the time/money saved?" Why: reveals if they've already priced the opportunity in their head. If they have, they'll answer specifically. If they hand-wave, the problem isn't priced yet mentally.
13. "Is this in your top 3 pain points right now, or further down the list?" Why: even a real pain isn't worth solving if it's not a priority. You need to know where your target problem sits in their attention hierarchy.
Understanding the current solution (Questions 14-17) — what will you have to displace?
14. "What tools or approaches do you use for [related area] today?" Why: enumerates the competitive landscape from their perspective. You'll learn about competitors you didn't know existed.
15. "What do you like about those tools?" Why: identifies what you'll need to at least match. Also reveals switching costs (positive features they'd lose).
16. "What do you wish those tools did that they don't?" Why: this is the closest you get to feature requests without triggering "solutioning mode." They're describing gaps, not designing your product.
17. "If [current tool] disappeared tomorrow, what would you do?" Why: reveals dependency. If they'd panic, the tool is deeply embedded (switching cost is high). If they'd shrug, they're not really committed to it.
Testing willingness to pay (Questions 18-21) — will they buy?
18. "Have you ever paid for a tool to help with this? What?" Why: past payment behavior predicts future payment. If they've paid for tools in this space, they will pay again. If they've never paid for related tools, they might not be a buyer at all.
19. "How does buying decisions get made in your company for tools in this area?" Why: identifies whether they're the buyer, an influencer, or a user. If they're not the buyer, you need to also interview the actual decision-maker.
20. "What's the range of pricing that would feel reasonable for solving this problem?" Why: don't ask "how much would you pay" — that produces lies. "What feels reasonable" produces a realistic range. Note: use this AT MOST once per interview; too much pricing talk shifts the conversation.
21. "What would have to be true for you to buy something like this?" Why: reveals objections and requirements before they become blockers. "It would need to integrate with X" or "I'd need to see 3 references" or "the CFO would have to approve anything over $500."
Testing your specific hypothesis (Questions 22-24) — carefully
22. "If I described a solution that [core benefit, not features], would you want to know more?" Why: this is the ONLY place you check your specific direction. Notice: benefit, not features. Not "an AI-powered dashboard" — "a way to see your team's workload in real-time without asking them."
23. "What questions would you want answered before you took a call about it?" Why: reveals what needs to be in your pitch/website/marketing. If they say "how it integrates with X," you need to prominently address integrations. If they say "how you keep our data secure," lead with security.
24. "Are there 2-3 other people in your company or network who deal with this same problem I should talk to?" Why: at the very end, get referrals. If you asked well, the interviewee should be happy to introduce you. This is how you scale customer discovery from 20 interviews to 100.
Closing (Question 25)
25. "Is there anything I should have asked you but didn't?" Why: reveals what THEY think is most important that you might have missed. Often the most valuable question in the whole interview.
The 7 questions to NEVER ask (they poison your data)
Bad question 1: "Would you use a product that does [X]?" Why bad: everyone says yes to be nice. Zero signal. Ask about past behavior instead.
Bad question 2: "How much would you pay for [product]?" Why bad: people underprice in their heads because they know they'll be quoted higher. Or they lie. Ask about reasonable ranges instead of specific numbers.
Bad question 3: "What features would you want?" Why bad: puts them in "product design mode" which is their weakest mode. You want their pain, not their design suggestions.
Bad question 4: "Do you think [my idea] is a good idea?" Why bad: they'll be nice. They don't know your idea in context. This question has zero validation power.
Bad question 5: "How often would you use it?" Why bad: they'll optimistically overestimate. Ask about how often the problem currently happens, not how often they'd use a hypothetical solution.
Bad question 6: "Do you know anyone else who might use it?" Why bad: refers your ASKED question forward. Ask instead about people with the underlying problem — question 24 above.
Bad question 7: "What do you think of the name/logo/pricing/design?" Why bad: this is opinion territory. Their opinion on design has zero correlation with their willingness to buy. Save this for later usability testing, not discovery.
How to actually run the interview
Format: 30-45 minutes. Video call is best (you catch facial expressions). Phone works. In-person is great but expensive.
Recording: get permission. Say "I'd love to record this so I can focus on listening and go back to your specific words later. Is that OK?" Almost everyone says yes.
Structure:
- First 3 minutes: rapport, thank them, remind them of the format
- Next 5 minutes: Warm-up questions (1-3)
- Next 15-20 minutes: Problem/pain/measurement questions (4-13)
- Next 8-10 minutes: Current solution + willingness to pay (14-21)
- Last 5 minutes: Hypothesis testing + referrals + closing (22-25)
During: listen 80%, talk 20%. If you catch yourself explaining or defending, stop.
After: within 24 hours, write down: the 3 most surprising things they said, the specific quotes worth remembering, and what this interview changed about your thinking.
The signals to look for across 15-20 interviews
One interview tells you nothing definitive. Fifteen to twenty tell you a lot. Look for:
Green light patterns:
- 70%+ describe the same specific pain in similar words
- 50%+ have already paid for something in the space
- 30%+ actively describe an unmet need matching your solution
- Referrals happen spontaneously ("you should talk to X")
Yellow light patterns:
- The pain is real but scattered — different interviewees describe different flavors
- People pay for tools in the space but none of them describe your specific solution as compelling
- Frequency of the problem varies wildly across your target segment
Red light patterns:
- Nobody spontaneously mentions the problem you're targeting (you have to remind them)
- People describe workarounds that are already good enough
- Nobody has paid for tools in this space at all
- Every interview feels different (means your targeting isn't clear)
A specific example
A founder I worked with had an idea for a B2B tool. Ran 20 interviews using this framework.
What she expected to hear: "yes we have this problem, we currently use X but X is bad, we'd pay $Y."
What she actually heard:
- 18/20 confirmed the general problem
- But when asked about the specific moment (question 4), 12 of 18 described a DIFFERENT specific pain than the one she was targeting
- Only 6/20 had ever paid for a tool in the space
- Those 6 all bought a specific competitor she hadn't heard of
What she did: shifted her product to target the specific pain 12 people described, went after the same buyers who'd bought the competitor she hadn't heard of, and repositioned her landing page around that competitor's known weakness.
Result: first 10 customers in 8 weeks. Would have been zero without the discovery — she'd have built for the pain she assumed, not the pain that was actually there.
What to do this week
Before running interviews:
- Make a list of 30-50 people in your target segment (see how to get your first 10 customers for outreach)
- Prepare the 25 questions. Print them. Have them next to you during calls.
- Practice on 2 friends first — you'll be bad at first, that's normal.
Running the interviews:
- Book 10-15 real ones over 2-3 weeks
- Record every single one (with permission)
- Take notes on paper during — don't type
- Write up the 3 surprises within 24 hours of each
After the interviews:
- After 5, look for patterns. After 10, refine your targeting. After 15-20, decide.
- If the signals are green: start pre-selling using method 1
- If the signals are yellow: interview 10 more with sharper targeting
- If the signals are red: consider whether to pivot or kill the idea — better now than after building
Where the interviews fit in your validation: Customer discovery interviews are Method 5 (cold reject) in the validation post. They're the fastest, cheapest way to falsify your idea before you spend money on it. They should ALWAYS come before pre-sales, and pre-sales should ALWAYS come before building.
The founders who nail this aren't smarter interviewers. They're better listeners. Every founder who fails at discovery does the same thing: they defend their idea instead of testing it. The 25 questions above are designed to prevent that failure mode. Use them.
---
If you'd like feedback on your specific customer discovery approach — the questions you're using, what you're hearing, and whether the signals are green/yellow/red — reach out via the contact page with a paragraph about your product and what you've heard so far. I'll spend 30 minutes on it and give you an honest read.