AI summary by TwinMind
How To Talk To Users | Startup School
Y Combinator
17m 31s
Introduction and Speaker Background
- Gustav: Group Partner at Y Combinator since 2017
- Previously worked at Airbnb, YC founder in 2007
- Topic: How to talk to users and future customers
Talk Agenda and Common Misconceptions
- Four main topics: why best founders talk to users throughout company lifetime, how to find/talk to users, what questions to ask/avoid, turning conclusions into MVP
- Debunks startup mythology: ideas don't come from lazy Sunday brainstorming or late night coding sessions
- Social Network movie example: inaccurate portrayal vs real photo of Brian Chesky with Amal (first Airbnb guest)
Airbnb Case Studies
Brian Chesky's 50 Airbnb Experiment (2010)
- CEO gave up apartment, lived in 50 different Airbnbs over months
- Real purpose: talk to 50 hosts daily, gather honest feedback
- Without hosts, Airbnb wouldn't exist today
- Brian and Joe still receive calls on personal cell phones from hosts
- They put personal numbers on website, shared with early hosts
- No hiding behind anonymous emails or "do not reply" addresses
Early Customer Connection Philosophy
- Direct personal connections vs scalable growth channels (Google ads)
- Most founders still hide behind info@ emails
- Search "do not reply" in email to see how many companies avoid customer contact
Finding Users to Interview
Personal Network
- Easy to interview known contacts
- Most likely to respond but potentially less honest to avoid offense
- Don't fear reaching out or rejection
Co-workers/Former Colleagues
- Great source with topic knowledge
- Useful for B2B software targeting startups
- Easy navigation to potential users
External Sources
- Common YC founder methods: LinkedIn, Reddit, forums, Slack/Discord communities, in-person events
- Users often outside personal/professional circles
Mock Startup Example: Carbon Emissions Tracking
Hypothesis
- Companies want to reduce carbon emissions but don't act on it
- Plan: interview potential customers, understand problems/motivations, determine MVP requirements
Target Audience
- Founders, CEOs, CFOs at startups/larger companies
- LinkedIn profiles with "carbon," "climate," "sustainability" titles
Research Goals
- Do companies care about carbon emissions?
- Why do they care/not care?
- Who in company cares most and why?
Outreach Message Examples
Known Contact (Former Airbnb CFO)
- Introduce self, mention shared Airbnb history
- Brief project description without details
- Request 20-minute phone/video call
Unknown Contact (LinkedIn)
- Similar structure, different tone for strangers
- Professional but approachable messaging
Interview Best Practices
Format and Setup
- Video calls, phone calls, or in-person preferred
- 5-minute video interview > 500-5000 survey responses
- Build rapport - interviewees answering questions never asked before
- Establish trust and comfort
Critical Interview Rules
- Don't introduce your idea until end of call (or not at all)
- Early introduction biases answers
- Role: listen, don't talk
- Use open-ended follow-up questions: "tell me about that"
- Take detailed notes if not recording
- Translate recordings to notes regardless
Sample Interview Demonstration
Example Conversation with Amy (Montevue Inc CFO)
- Company cares about carbon emissions
- Uses consultant for annual PDF reports
- Problem: don't take action on reports
- Issues: information too dense, unfamiliar terminology, inaccurate data, need expert interpretation
- Motivation: regulations coming up
- No product/solution introduced during conversation
Effective Interview Questions
Good Questions to Ask
- "Tell me how you do X today" (specific problem/task/goal)
- "What is the hardest thing about doing X?"
- "Why is it hard?"
- "How often do you have to do X?"
- "Why is it important for your company to do X?" (dive deep into motivation)
- "What do you do to solve this problem today?"
Ideal State
- Watch them perform the task (screen share during video call)
- See actual tools, documents, processes
- Understand behavior, not just words
- Changing behavior is difficult
Useful Follow-up Questions
- "What do you mean by that?"
- "Can you tell me more about that?"
- "Why is that important to you?"
Questions to Avoid
Problematic Question Types
- "Will you use our product?" (they'll say yes meaninglessly)
- "Which features would make product X better?" (that's your job, not theirs)
- Yes/no questions (want detailed explanations with concrete examples)
- "How would a better product X look like?" (they're not product developers/designers)
- Asking two questions simultaneously (confuses answers)
Focus on Problems, Not Features
- Natural tendency to think about solutions during interviews
- Users have good problems but generally bad solutions
- Examples:
- Early Gmail: users wanted inbox + email on same screen (real problem: Gmail too slow)
- Early Airbnb: guests wanted host phone numbers (real problem: insufficient platform trust)
- Users don't have incentive to say no to additional features
- Your job: determine most important features through prioritization
Post-Interview Analysis
Organizing Learnings
- Compile notes from 5-10 user interviews
- Use sticky notes/software to organize learnings
- Bucket findings into different problem categories
- Identify most important issues
- Write down conclusions from all interviews
Creating Solutions
- Use information to create solution hypothesis
- Don't over-intellectualize process
- Start designing MVP quickly with accurate information
- Test MVP with same users
Determining Problem Value
Value Assessment Criteria
- Are people paying for other solutions in this space?
- Example: PDF report creators for carbon emissions accounting
- Payment indicates valuable problem
- Do people have existing solutions they're happy with?
- Excel/Google Sheets = competitors to hundreds of startups
- Need dramatically better experience to justify switching
- Won't pay hundreds monthly for marginal improvement over free tools
- How easy is it to sell to this audience?
- Plumbers/contractors: difficult, rarely change tools
- Startups: easy, open to trying new things constantly
MVP Testing and Prototyping
Early Testing Methods
- Show users design prototypes (Envision, etc.)
- Airbnb example: testing with people waiting for interviews/coffee
- Use clickable prototypes on phones before building
- Key rule: don't tell users what to do, just watch
- Give specific goals: "try to make booking" or "search with dates"
- Don't provide step-by-step guidance
- Have users speak their thoughts aloud during testing
- Listen for word comprehension and screen understanding
Ongoing User Engagement
Building User Community
- Create Slack instance or WhatsApp group with customers
- Make users feel special with exclusive access
- Position as "exclusive access to future world-changing product"
- Show product progression continuously
- Build trust by reacting quickly to feedback
- Ship new versions/screens based on user input
- Enable connections between users doing similar work
- Users value networking opportunities you provide
Summary and Conclusion
- Covered four main areas: why best founders talk to users throughout company lifetime, finding and talking to users, effective questioning techniques, turning conclusions into MVP
- Emphasized continuous user engagement throughout company development
- Highlighted importance of listening over talking in user research
