The skill: Extracting truthful, specific, actionable information from customers instead of collecting compliments that feel like progress but teach you nothing.
Setting up the conversation
The first 60 seconds determine whether you get honesty or politeness. Use this sequence:
- Start with your vision — what you're trying to improve
- Frame where you are in the process
- Show vulnerability — what you need help with
- Put them on a pedestal — why you're asking them specifically
- Make the ask — 15 minutes
This earns honesty because it signals you're genuinely trying to learn, not just validating your idea. People give better answers when they feel like an expert being consulted, not a user being tested.
Questions that surface truth
The core principle: talk about their life, not your idea. Ask about specifics in the past, not opinions about the future. Talk less, listen more.
To understand their real behavior, use these five questions. They work in almost any context and force specifics — you can't answer them with platitudes:
- "Why do you bother?"
- "What are the implications?"
- "What else have you tried?"
- "How are you dealing with it now?"
- "Talk me through the last time that happened."
To understand why they use (or switch) products, trace the timeline:
- "When did you first start looking for a solution?"
- "What were you trying to accomplish?"
- "What did you try before this?"
- "What made you switch?"
This surfaces the real job being done — not feature preferences, but the actual outcome they were hiring a product to deliver.
To understand what blocks or drives switching, probe all four forces:
- Push: "What was frustrating about how you were doing it before?"
- Pull: "What caught your eye about the new solution? What did you hope would be different?"
- Anxiety: "What almost stopped you from trying it? What were you worried about?"
- Habit: "What was good enough about the old way? What made it hard to let go?"
If you only hear push and pull, you're getting half the picture. The deals that stall, the trials that don't convert, the signups that never activate — those are usually anxiety and habit winning.
Recognizing bad data
Three types of responses that feel like progress but teach you nothing:
- Compliments — "love it!" tells you nothing about whether they'd actually use or pay for it
- Fluff — "I would definitely..." is a prediction about future behavior, which people are terrible at
- Ideas — "it should do X" is them designing your product instead of describing their problem
When you hear these, redirect back to specifics: "Tell me about the last time you actually tried to do that." The goal is always to get back to concrete past behavior.
Knowing when you're done
After each conversation, apply the commitment test: did you leave with a commitment — time, reputation, or money — or just a compliment? Compliments cost nothing to give. If someone says "that's interesting, let me know when it's ready," you've learned nothing about demand.
Watch for zombie leads — people who keep saying encouraging things but never commit to a next step. Give them a concrete chance to say no. A clear no beats an indefinite maybe.
Stop interviewing when conversation 8 sounds like conversation 5. Three to five conversations per segment is usually enough for initial pattern recognition. After that, you're confirming, not discovering.
Interview Script
Do's and Don'ts
Written with ❤️ by a human (still)