Skip to content

3 · Designing interview questions that get the truth

The fastest way to fool yourself is to ask questions that beg for a yes. People are polite — especially to a friendly founder — so a leading question harvests flattery, not facts. Good customer research is mostly the discipline of not asking for the answer you want.

The cardinal rule: ask about the past and present, not the hypothetical future. "Would you use an app that does X?" gets a cheap, meaningless "sure!" People are terrible at predicting their own future behavior. "Tell me about the last time you dealt with X — what did you do?" gets a true story you can learn from.

Bad question → better question:

Leading / hypothetical (avoid)Open / behavioral (use)
"Don't you hate how slow X is?""Walk me through how you handle X today."
"Would you pay for a tool that fixes this?""What have you tried to fix this? What did it cost you?"
"Do you like my idea?""What's the most frustrating part of X for you?"
"Would you use this?""When did you last run into this problem?"

Principles for honest interviews:

  • Ask open, not yes/no. Open questions surface what you didn't think to ask.
  • Dig into real stories, not opinions. "What did you actually do?" beats "what do you think?"
  • Don't pitch. The second you sell your idea, the conversation is about being nice to you, not about the truth. Listen first.
  • Stay neutral. Don't telegraph the answer you're hoping for in your tone or wording.

Where AI helps: paste your draft questions and ask it to flag which ones are leading, hypothetical, or double-barreled, and to rewrite them as open and neutral. AI is good at spotting bias in wording. Just remember the SBA frames interviews and surveys as direct research with real people — AI tightens the questions; the answers must come from outside the building (SBA, n.d.).

Check yourself. Why is "would you use this?" a weak question, and what kind of question gets you a more honest answer?

Sources