How to Run Product Discovery Interviews (Step-by-Step)
Product discovery interviews are the most direct way to understand user needs, but most PMs conduct them poorly. The biggest mistake is asking users what they want instead of understanding what they do and why. Users are experts on their problems, not on solutions.
This guide covers the complete process: finding participants, preparing questions, conducting the interview, and synthesizing insights into actionable product decisions.
Step-by-step guide
Step 1: Define the research question
Before scheduling interviews, define what you are trying to learn. A good research question is specific and actionable: "Why do users abandon the onboarding flow after step 3?" not "What do users think of our product?" The research question determines who to interview and what to ask.
Step 2: Recruit the right participants
Interview people who match your target user profile and have recently experienced the problem you are investigating. Recruit from: recent signups who churned, active users who use the feature you are improving, or prospects who evaluated but did not buy. Aim for 5-8 interviews per research question.
Step 3: Prepare an interview guide
Write 5-7 open-ended questions that explore the user's context, behavior, and pain points. Start broad ("Tell me about the last time you...") and narrow toward specifics. Do not ask leading questions or suggest solutions. Include follow-up prompts: "Tell me more about that" and "What happened next?"
Step 4: Conduct the interview
Start with rapport-building (2 minutes), then move to the core questions (20-25 minutes), and end with open-ended closing (3-5 minutes: "Is there anything else you wish I had asked?"). Listen more than you talk. When the user says something surprising, follow the thread instead of moving to the next question.
Step 5: Take structured notes
Capture direct quotes, observed behaviors, and pain points in separate categories. Mark surprising insights with a flag. After the interview, write a 3-sentence summary of the key takeaway while it is fresh. Do not try to synthesize during the interview.
Step 6: Synthesize across interviews
After completing all interviews, look for patterns: which pain points appeared in 3+ interviews? Which behaviors were consistent? Where did users disagree? Group findings into themes. Each theme becomes an opportunity to address in your product.
Common mistakes
Asking "Would you use X?"
Users always say yes to hypothetical questions. They cannot predict their own behavior. Instead, ask about past behavior: "Tell me about the last time you faced this problem. What did you do?" Past behavior predicts future behavior better than stated intent.
Pitching your solution during the interview
The moment you start describing your product, the interview becomes a sales call. Users will tell you what you want to hear. Keep the focus on their problems and current behaviors, not your solutions.
Insufficient sample size
One or two interviews can be misleading. Aim for 5-8 interviews per research question. If you see the same pattern in 5 interviews, you have a reliable signal. If findings are scattered across 8 interviews, the pattern may not be strong enough to act on.
Not synthesizing into themes
Individual interview notes are not actionable. The value comes from synthesizing across interviews to identify patterns. Block time after your last interview to compare notes and extract themes.
Tips
- Record interviews (with permission) so you can re-listen for details you missed
- Interview in pairs: one person asks questions, one takes notes
- Schedule interviews in clusters (3-4 per week) so patterns emerge while context is fresh
- Share interview summaries with the team within 24 hours
How Vantage helps
Vantage lets you add interview transcripts and research notes as context sources when creating a project. The AI uses these insights to ground PRD generation in real user needs, so your specs reflect what users actually said rather than what you assumed they meant.