How to Run a Problem Interview Without Sounding Like a Researcher
Type: media
Stage: Stage 1: Problem Proof
Difficulty: beginner
A step-by-step guide to having natural conversations that reveal real pain — not polished feedback.
Overview
Most founders ruin their problem interviews before they start. They ask the wrong questions, hear what they want to hear, and walk away convinced they have a great idea. This guide teaches you to have natural conversations that surface real pain — not polished, performative feedback.
The core rule: ask about the past, never the future
Never ask a potential user whether they 'would use' your idea. People are terrible at predicting their own future behavior — especially when they want to be helpful to you. Instead, ask exclusively about past behavior.
The question that changes everything: "How did you handle this problem last week?"
Past behavior is evidence. Future intent is noise. If someone spent three hours in a spreadsheet trying to solve something last Tuesday, that's signal. If someone says they'd 'probably' use your solution, that's not.
The Rule of 10-20-30
Before you have enough evidence to move forward, aim for:
• 10 deep conversations with people who match your target audience
• 20 distinct manual workarounds — solutions they've already built themselves
• A problem that occurs at least 30 times per year for each person
If the problem only happens once a year, urgency is low. If they haven't tried to solve it on their own, it probably doesn't hurt enough. And if ten conversations don't surface similar pain, you don't have a market — you have an anecdote.
Strong signal vs. weak signal
Strong signal: a user who spends hours in a spreadsheet, pays for a workaround tool, or has built something themselves to manage the problem. They can describe the cost in concrete terms — time lost, money wasted, mistakes made.
Weak signal: someone who says your idea 'sounds interesting' but has never tried to solve the problem. They ask 'how would this work?' rather than 'when can I use it?' They can't describe the last time the problem affected them.
Treat weak signals as useful data — they tell you where interest ends and real pain begins.
Questions that reveal truth
Use these prompts to get past politeness and into real behavior:
• "Walk me through how you handle [X] today."
• "What's the most frustrating part of that process?"
• "When did this last happen? What did you do?"
• "Have you tried to fix this? What happened?"
• "What would you lose if this problem never got solved?"
Do not ask: 'Would you pay for this?' Do not describe your solution. Do not offer suggestions. Your job in this conversation is to listen and record exact words — especially the phrases people use to describe their pain.
What to do with your notes
After each conversation, write down the exact phrases the person used. Look for repeated language across interviews. If three different people describe the same problem using the same words, those words belong in your positioning.
Count the strong signals. Count the weak signals. If you have fewer than five strong signals after ten conversations, you either need to talk to different people — or reconsider whether the problem is real.