Talk to Your Users First
Summary of the video “How To Talk To Users | Startup School” by Y Combinator.
The best founders build products by talking directly to users before, during, and after development. Learn how to find users, ask the right questions, avoid common mistakes, and turn feedback into an MVP.
Why Talking to Users Matters
Users Keep You Honest
Users and customers are the only stakeholders actually paying you money, making them the most reliable source of truth about your product. They have no incentive to lie to you the way investors or friends might.
The Airbnb Hosts Example
Brian Chesky lived in 50 different Airbnbs over several months to talk directly to hosts every single day. This deep user engagement revealed critical feedback that shaped Airbnb's entire direction. Early founders like Brian and Joe even put their personal cell phone numbers on the website so hosts could call them directly.
Most Founders Hide Behind Email
Most founders today avoid direct customer contact, hiding behind anonymous or 'do not reply' email addresses. This was considered almost radical when Airbnb founders chose transparency instead.
Finding Your First Users
Three Tiers of User Access
Start with people you know (easiest to reach but potentially less honest), then move to coworkers or former colleagues who understand your domain, and finally reach outside your network to find truly unbiased users.
Common Channels for Finding Early Users
YC founders typically find early users through LinkedIn, Reddit forums, Slack or Discord communities, and in-person events. These channels let you identify people with relevant job titles or interests.
Outreach Message Strategy
Keep outreach messages short: introduce yourself, mention a connection or context, briefly describe your project without too much detail, and ask for a 20-minute call. Adjust tone based on whether you know the person or not.
How to Interview Users
Interview Format Matters
Conduct interviews over video calls, phone calls, or in person—never rely on surveys. A five-minute video interview yields far more insight than 500 or 5,000 survey responses because you can observe behavior and ask follow-up questions.
Build Rapport First
Make the interviewee comfortable and trusting before diving into questions. They will be sharing problems and feedback that maybe no one has asked them about before, so establishing psychological safety is critical.
Never Introduce Your Idea Early
Hold off on describing your product idea until the end of the call or not at all. Introducing your solution too early biases their answers and shifts them from honest problem-sharing to feature feedback.
Listen More Than You Talk
Your role is to listen and learn, not to pitch or convince. Use open-ended follow-up questions like 'tell me more about that' or 'what do you mean by that' to encourage deeper sharing.
Take Notes or Record
Always record the interview if possible, or take detailed notes during the call. You will need to translate recordings to notes anyway, so capturing information in real time is more efficient.
Questions to Ask (and Avoid)
Six Core Questions for User Interviews
Ask: (1) How do you do X today? (2) What is the hardest thing about X? (3) Why is it hard? (4) How often do you do X? (5) Why is it important for your company? (6) What do you do to solve this problem today? Dive deep especially on motivation and current behavior.
Follow-Up Questions Are Critical
People rarely say everything in one answer. Master follow-ups like 'What do you mean by that?', 'Can you tell me more?', and 'Why is that important to you?' to extract concrete examples and deeper understanding.
Questions to Avoid
Do not ask: (1) Will you use our product? (2) Which features would make product X better? (3) Yes/no questions (4) How would a better product look to you? (5) Two questions at once. These derail the interview and shift focus from problems to solutions.
Focus on Problems, Not Features
Users have good problems but generally bad solutions. Your job is to understand the problem deeply; feature ideation is your job, not theirs. Users will say yes to any feature you ask about because they have no incentive to say no.
Real Examples: When Users Suggest Bad Solutions
Gmail users asked to view inbox and email on the same screen—the real problem was Gmail was too slow. Airbnb guests asked for host phone numbers—the real problem was lack of trust in the platform. Always dig for the underlying problem, not the proposed feature.
From Interviews to MVP
Organize and Synthesize Interview Notes
After 5-10 interviews, organize your notes using sticky notes or software. Bucket learnings by problem type, identify which problems matter most, and write down your conclusions about what solution might work.
Create a Hypothesis and MVP Quickly
Don't over-intellectualize. Use your learnings to form a hypothesis about the solution, then build an MVP as fast as possible. Make sure your hypothesis is based on accurate information from users.
Test MVP with Original Users
Show your prototype (even just a design mockup in Figma or Invision) to the same users you interviewed. Watch them interact with it without telling them what to do. Give them a goal like 'try to make a booking' but don't guide them through each step.
Evaluate Problem Validity
Before investing heavily, confirm the problem is valuable: (1) Are people paying for existing solutions? (2) Do people already have solutions they're happy with (like Excel)? (3) Is this audience easy to sell to? Moving someone off Excel requires dramatically better experience.
Keep Users Involved Throughout
Create a Slack or WhatsApp group with early users to make them feel special and give them exclusive access to the evolving product. Show them progress, react quickly to feedback by shipping new versions, and let them connect with other users doing similar work.
Notable quotes
Users and customers will keep you honest. They are the only stakeholders actually paying you anything. — Gustav
A five-minute video interview teaches you more than 500 or 5,000 survey responses. — Gustav
Users have good problems but generally bad solutions. Your job is to understand the problem, not to ask them for features. — Gustav
Action items
- Identify 5-10 potential users in your target market using LinkedIn, Reddit, Slack communities, or in-person events.
- Craft outreach messages that introduce yourself, mention context, briefly describe your project, and ask for a 20-minute call.
- Schedule and conduct video or phone interviews, focusing on understanding their current process and pain points without mentioning your solution.
- Use the six core questions to dig deep: How do you do X today? What's hardest? Why? How often? Why important? What do you do now?
- Take detailed notes or record interviews, then organize learnings by problem type to identify the most critical issues.
- Build a quick MVP prototype (even just a design mockup) and test it with the same users, watching them interact without guidance.
- Create an exclusive Slack or WhatsApp group with early users and involve them in ongoing product development.
- Validate that the problem is valuable by checking if people pay for existing solutions and whether your audience is easy to sell to.