Customer Discovery Done Right: Ask About the Past, Never Pitch
Customer discovery is a set of short interviews where you ask people about their real past behavior, never about your idea. Stop when you hear the same problem repeatedly, then turn the notes into the problem and traction slides. Round Funded takes over for the raise, with the pre-seed investor list.
Stanford University describes the goal of customer discovery as finding out who the customers for your product are and whether the problem you think you are solving is important to them. That includes who has the power to make or influence the buying decision. It is not a pitch, and it is not market research by survey.
Steve Blank's process asks for one thing above all: keep going until the answers repeat. Until you hear the same thing over and over, you are not done. The Mom Test rules point the same direction: talk about their life, not your idea, and learn from actions, not opinions.
Why it matters before a raise
Investors at pre-seed are buying your understanding of a problem, and interviews are the cheapest way to prove you have it. A deck with real customer quotes beats a deck with a market-size slide and nothing underneath.
Y Combinator is blunt that you can learn more from a five-minute video interview than from 500 or 5,000 survey responses. Founders who skip the conversations tend to build for a problem that sounds big but does not hurt. The same guidance says to launch something mediocre quickly, then keep talking to customers and iterating, rather than waiting for a perfect product.
The Airbnb story shows the payoff. The founders offered to photograph early hosts' homes, then took the photos themselves. According to Y Combinator, listings improved, conversions improved, and they had amazing conversations with customers. Talking to users was the work, not a side task.
If you are still testing whether the idea deserves a company at all, start with how to validate a startup idea, then come back here for the interview craft.
How the Customer Discovery Process Works
The customer discovery process is simple: form a hypothesis about who has the problem, interview people in that group, record what they say, and revise. Repeat until the same pain keeps coming up.
Steve Blank's loop in plain terms
Your first job is to find out who the customer is and whether the problem matters to them. Your second is to find who actually makes the buying decision, which is often a different person than the user.
Do not wait for a finished product to begin. Interviews feed the product, and the product feeds the next round of interviews.
How many customer discovery interviews you need
There is no single magic number, and sources disagree. Use the table as a range, not a rule.
| Source | Guidance |
|---|---|
| Harvard Business School | Early discovery: five interviews per persona, with consistency in each |
| Syracuse University LaunchPad | 150 total: 50 interviews, then 100 by survey |
| Steve Blank | Keep going until you hear the same thing over and over |
A practical read: start with five per persona, then keep going until answers repeat. If every call sounds different, your group is too broad. Split it by persona and run the interviews again.
Run each interview as a team
Harvard Business School suggests a discovery team of two to three people: one asks, the others observe and take notes. Record with permission, as VentureWell advises, so you stay present instead of typing. In a well-run interview, the participant does most of the talking.
Customer Interview Questions: Good, Bad and Why
Good customer interview questions ask about specific past events. Bad ones ask for opinions or predictions about your idea.
Questions to avoid
These fail because the answer tells you nothing:
- "Do you think it's a good idea?" People tell you what they think you want to hear.
- "Would you use this?" or "Would you buy this?" Y Combinator notes people will probably say yes, and that it means nothing to you.
- "Wouldn't you agree that this is a problem?" This is a leading question, flagged by Harvard Business School.
- "Do you think our product is too expensive?" VentureWell says you are not ready to discuss pricing yet.
- Any pitch of your solution in the opening minutes. Y Combinator says introduce the idea at the end of the call, or not at all.
Questions that work
Open with the one Y Combinator recommends: "Tell me about the last time you ran into this problem." Then follow the story.
| Bad question | Better version |
|---|---|
| "Would you pay for a tool that does X?" | "What did you do the last time X went wrong?" |
| "Do you think this is a problem?" | "Walk me through the last time this came up." |
| "Would you use our product?" | "What are you using today to deal with this?" |
| "Is our pricing fair?" | "How is this cost handled in your budget today?" |
| "Do you like the idea?" | "What else should I have asked you?" |
Listen for emotion. Harvard Business School says the emotional response is key. But check for real stakes: one interviewee described a workflow with strong language, then admitted interns handled it and it worked pretty well. Frustration without consequences is not a problem worth building for.
Syracuse University LaunchPad says whatever the person does now to cope, whether a spreadsheet, a workaround or nothing at all, is your competition. Ask about it directly. Prepare about 20 questions, since you will usually get through only about 10 per interview.
Close every call the same way
End with "What else should I have asked you?" and "Who else should I interview?" VentureWell notes the best leads often come from people you have already talked to. Ask for two or three referrals each time.
How to Find Interviewees on LinkedIn and in Communities
Look where your customers already talk. Y Combinator asked its founders how they found early users, and common answers were LinkedIn, Reddit forums, Slack or Discord communities, and in-person events.
Keep two degrees of separation between you and the interviewee, per Syracuse University LaunchPad. Friends and family want you to feel good, which makes their answers useless.
For a LinkedIn search, filter by job title and company type, then send a short note asking for 15 minutes about how they handle the problem. Say you are researching, not selling. A template with placeholders:
Hi [Name], I'm researching how [role] teams handle [problem]. I'm not selling anything. Could I ask you three questions about the last time it came up? 15 minutes, whenever suits you.
Referrals beat cold notes. Warm introductions get a much higher response than cold outreach, so ask every interviewee for two or three names.
Posts that bring interviewees and later customers
Your own LinkedIn posts can pull in the right people before you ask. PostCat AI helps founders get customers from LinkedIn: it drafts posts in each person's own voice, answers comments, and traces clicks from a post to a sign-up or demo. PostCat AI is our sister product. Nothing is published from a person's profile without their consent: they approve the post, name someone who approves for them, or turn on a setting that publishes for them. Its guide on how to get customers fits the stage after interviews, when the same posts start attracting buyers.
If your first few interviews turn into committed early users, the design partner program is the natural next step.
Turn Customer Discovery Interviews Into Problem and Traction Slides
Record evidence as you go, then group it into patterns. Investors read two slides that your interviews feed directly: problem and traction.
Keep a simple evidence log
After each call, write down the date, the persona, the specific past event they described, what they use today, and one exact quote. Do not paraphrase the quote. Keep the recording, with permission, so you can re-listen.
Count patterns, not opinions. A note like "seven of nine ops managers described the same weekly spreadsheet reconciliation" is evidence. "People seem to like it" is not.
From notes to slides
| Interview output | Where it goes |
|---|---|
| Repeated pain with exact quotes | Problem slide |
| Current workaround as competition | Problem and competition slides |
| Who decides and who pays | Market and go-to-market slides |
| Interviewees who ask for early access | Traction slide |
| Referral chains you built | Proof of a reachable market |
The problem slide should read like the interviews: one person, one specific moment, one cost. Traction at pre-seed often means people who took a concrete step, such as joining a pilot, not people who said "sounds great." Check the result against your sense of product-market fit, and use a pitch deck review to compare your slides with 342 real decks from startups that closed the round.
Where Round Funded Fits After Customer Discovery
Once the interviews give you a sharp problem and real evidence, Round Funded handles the raise. The product surface for this stage is the pre-seed investor list.
Customer discovery tells you what to say. Round Funded helps you get it in front of the right investors.
| What you need | What Round Funded provides |
|---|---|
| Investors to contact | 60,000+ active VCs, angels and accelerators, filtered to a check in the last 6 months |
| Sending pitches | Outreach from your own Gmail, so replies land in your inbox |
| Evidence in one place | Shareable Data Room with deck, traction, team and ask, free on every plan |
| Tracking conversations | Fundraising CRM with pipeline stages from first email to close |
| Deck check | Free Pitch Deck Review against 342 funded decks |
Find pre-seed investors who fit your problem →
Step-by-Step: Run Customer Discovery Interviews Before You Raise
Follow these steps in order. Each one builds the evidence a pre-seed deck needs.
- Use the pre-seed investor list on Round Funded to see what investors at your stage back, then write down the one problem statement you need interviews to prove or kill.
- Define one persona and one hypothesis. Note who feels the pain and who holds the budget.
- Prepare about 20 questions, starting with "Tell me about the last time you ran into this problem." Remove every hypothetical and every pitch.
- Recruit five interviewees per persona from LinkedIn and communities, keeping two degrees of separation from them. Use a short research note, not a sales message.
- Run each call with a second person taking notes. Record with permission, let the interviewee talk most of the time, and hold your idea back until the end, if you share it at all.
- Close every call with "What else should I have asked you?" and "Who else should I interview?" Ask for two or three referrals.
- Log each interview the same day: date, persona, the past event, current workaround, one exact quote.
- Stop when answers repeat. Turn the repeated pain into the problem slide and the committed early users into the traction slide, then start your raise.
Frequently Asked Questions
How many customer discovery interviews should I do?
Harvard Business School suggests five per persona in early discovery, with consistency in each interview. Syracuse University LaunchPad recommends 150 touchpoints: 50 interviews and 100 survey responses. Whatever the count, Steve Blank's test applies: keep going until you hear the same thing repeatedly. See investors who back pre-seed teams once you have it.
What are good customer interview questions?
Start with "Tell me about the last time you encountered this problem," which Y Combinator recommends. Then ask what they do today, what it costs them, and what they tried before. Avoid "Would you use this?" because the answer is probably yes and means nothing. Review your slides with the pitch deck review before you pitch.
When should I tell interviewees about my idea?
Late, or never. Y Combinator says the most important thing is not to introduce your idea until maybe the end of the call, because doing it early can bias answers. VentureWell adds that you should avoid talking about your technology or pricing in the first conversation.
Can I interview friends and family?
Avoid them. Syracuse University LaunchPad says friends and family are trying to make you feel good, which makes their answers useless, and recommends at least two degrees of separation from every interviewee. Find strangers through LinkedIn, Reddit, Slack or Discord communities, and referrals from earlier interviews.
How do I use interview results in a pitch deck?
Turn repeated pains and exact quotes into the problem slide, and committed early users into the traction slide. Count patterns, not opinions. Round Funded's free Data Room lets you hold the deck, traction, team and ask on one shareable link for investors.
Are surveys a replacement for customer discovery interviews?
No. Y Combinator says you can learn more from a five-minute video interview than from 500 or 5,000 survey responses. Harvard Business School also warns against relying on interviews alone, so treat surveys as a second step that tests what interviews surfaced.
Should I record customer discovery interviews?
Yes, with permission. VentureWell notes recording lets you stay engaged instead of splitting attention between listening and note-taking. Harvard Business School also suggests two or three people per call, with one asking and the others taking notes. Then store quotes in your evidence log.
Make Customer Discovery Your Fundraising Foundation
Good customer discovery means asking about the last time a problem happened, never pitching, recording evidence, and stopping only when answers repeat. Five interviews per persona is a start, not a finish. Those notes become your problem and traction slides, and then the raise is a matter of reaching the right investors.
Start your raise with the pre-seed investor list →
Talk to customers first, then to investors. Round Funded helps with the second part.

