Startup User Research

Startup user research is not a formal survey exercise. In the YC lectures, it means getting uncomfortably close to users, the industry, and the actual workflow before and during product building.

Adora Cheung's early-stage method

• State the problem in one sentence.
• Ask whether the founder personally has the problem or can get very close to those who do.
• Immerse in the industry; become a participant if necessary.
• Identify a narrow initial customer segment.
• Storyboard the full user experience from discovery to post-use feedback.
• Build the smallest viable product that solves the immediate problem.
• Talk to users directly, preferably in conversation rather than lab-like interrogation.

Quality of feedback

Feedback quality varies by relationship and payment. Friends and family may be supportive but low-signal. Random users who paid are more likely to be brutally honest. Surveys mostly capture extremes; direct conversation reveals the middle.

Manual learning beats premature automation

For Homejoy, manually cleaning homes and manually interviewing cleaners taught the team what needed to become software. Cheung's rule is to do manual work first, then automate once the team understands which steps predict quality or conversion.

Emmett Shear user interview addendum

Lecture 16 adds a more tactical interview method: choose a narrow user segment, find people who actually match it, ask about current behaviour rather than imagined preferences, and avoid selling the idea during the interview. This deepens the user-research side of Doing Things That Do Not Scale and Building Products Users Love.