Every founder who has done customer discovery interviews comes back with the same problem: everyone said they liked the idea. Nobody said they would not use it. And then nobody used it.
The issue is not the people — it is the questions. Questions like "would you use this?" invite positive answers because people are naturally encouraging. The fix is not to ask better questions about your product. It is to ask questions about their life and the problem they already have.
Quick answer: Ask about the last time they dealt with the problem, not whether they would use your solution. Past behaviour is real data. Hypothetical responses to a product that does not exist are not.
The rule: never mention your product in a discovery interview
The moment you describe your product, the interview shifts from discovery to validation. The person is no longer telling you about their actual experience — they are reacting to your idea. Their answers start to be shaped by what they think you want to hear.
A discovery interview should spend 80% of its time asking about the problem the person currently has. How do they deal with it? What have they already tried? What does their current workaround cost them in time, money, or frustration? You will learn more from these questions than from any reaction to your prototype.
The five questions that produce honest signal
These questions — adapted from Rob Fitzpatrick's The Mom Test — reliably produce answers that change what founders build:
- "Tell me about the last time you ran into [the problem]." — Forces them to recall a real episode, not a hypothetical.
- "What did you do about it?" — Reveals whether they cared enough to act. No action means low pain.
- "What tools do you currently use to handle this?" — Shows what you are actually competing with.
- "What do you wish your current approach did that it does not?" — Surfaces the real gap without leading them toward your solution.
- "If you could wave a wand and have this solved perfectly, what would that look like?" — Reveals their mental model of the ideal outcome, which is often different from what you are building.
Warning signs in interview answers
Learn to spot these signals that an interview is giving you false positives:
- "That sounds really interesting" but no specific personal example — They are being polite, not confirming a real problem they have.
- "I would definitely use that" — This is the most dangerous phrase. Follow up with: "Have you tried to solve this problem before? What happened?" If they have never tried to solve it, the pain is not high enough.
- No current workaround — If someone does not have a workaround, the problem is either not painful enough or not real. Painful problems get worked around with duct-tape solutions.
How many interviews do you need?
Five interviews with people who genuinely have the problem will teach you more than fifty interviews with people who might have it. Recruit carefully. Do not interview your friends, your family, or people who agreed to a call just to be helpful. Interview people who have done something about this problem in the last 90 days — posted about it online, built a spreadsheet for it, paid for a partial solution.
By interview five or six, you will start hearing the same phrases recurring. That convergence is your signal that you have enough data to move forward.
What to do after the interviews
Write down the exact phrases your interviewees used — not your paraphrase of their words. Founders paraphrase toward their own hypothesis. Raw quotes stay honest. Use these phrases in your landing page copy. Words that your target customers use to describe their own problem will convert cold traffic better than any marketing language you invent.
Related reading
- How to Validate a SaaS Idea in 2 Weeks
- How to Go From 0 to 100 Signups in Your First Month
- How to Use a Waitlist to Validate Your SaaS Idea
- When Is the Best Time to Launch a SaaS in 2026?
Frequently Asked Questions
How do I find people willing to do a discovery interview?
Post in communities where your target customer already hangs out — Indie Hackers, specific subreddits, Slack groups, LinkedIn. Lead with the problem, not with your product: "Building something for freelancers who struggle with client invoicing — would love 15 minutes of your time to understand how you handle it today." People are more willing to share their experience than to evaluate someone else's product.
Should I record the interviews?
Yes, with permission. You cannot listen actively and take good notes at the same time. Record and review the transcript later. The phrases you will want to reread are almost never the ones you thought to write down in the moment.
What if every interview confirms the problem is real?
Confirm willingness to pay before you build. After the discovery interview, follow up with: "We are building a tool that handles exactly this. We are planning to charge $X/month. Would you be willing to pay that for early access?" An email address on a waitlist costs nothing. A credit card number — or even a firm "yes, here is my card" — means the problem is real enough to pay to solve.
Ready to get your product discovered?
A free, permanent listing, indexed by search engines and AI assistants. Skip the queue from $19.