Customer Discovery Interview Questions That Test Demand
TL;DR
Ask 10 questions about a recent problem, the current workaround, its cost, who decides, and what triggers a switch. Follow every answer with a neutral probe for a real example. Treat compliments and hypothetical interest as weak evidence. Use repeated, specific behaviour to choose whether to test, refine, or pause.
Ask these 10 customer discovery interview questions before you invest in building a product. Start with a recent real problem, then trace the workaround, cost, decision-maker, and switching trigger.
The goal is not to collect praise for your idea. The goal is to find behaviour that shows a problem is real enough to deserve the next test.
Use This Copy-Paste Interview Guide
Run the questions in order. Ask the follow-up only after the participant answers. A semi-structured interview uses open questions and responsive probes, rather than a rigid questionnaire, so the participant can describe their experience in their own words. Interview methodology
-
“Tell me about the last time you had to deal with [problem].”
Signal: A recent, specific problem event.
Ask next: “What happened from the beginning?”
Weak answer: The person speaks generally or cannot remember an example. -
“What were you trying to get done?”
Signal: The customer’s underlying job.
Ask next: “Why did that matter at that moment?”
Weak answer: The person repeats your proposed solution instead of describing their goal. -
“What did you do first?”
Signal: Actual workflow and behaviour.
Ask next: “What happened after that?”
Weak answer: The answer starts with “I would probably.” -
“How do you handle that situation today?”
Signal: The current workaround or alternative.
Ask next: “Can you walk me through the steps?”
Weak answer: There is no workaround and no meaningful downside. -
“What is hardest about the current approach?”
Signal: The friction your product may need to remove.
Ask next: “Can you give me a recent example?”
Weak answer: “It is annoying” has no delay, error, cost, risk, or lost outcome attached. -
“How often does this situation come up?”
Signal: Frequency and recurrence.
Ask next: “When did it happen before the example you gave?”
Weak answer: The person guesses instead of recalling events. -
“What happens when the problem is left unsolved?”
Signal: The consequence of doing nothing.
Ask next: “Whose time, money, risk, or result is affected?”
Weak answer: The consequence remains vague. -
“What have you already tried, searched for, or paid for?”
Signal: Active effort to solve the problem.
Ask next: “What made you choose that option?”
Weak answer: The person has only heard about alternatives. -
“Who else is involved when this needs solving?”
Signal: The user, buyer, approver, and blocker.
Ask next: “Who can approve spending or a process change?”
Weak answer: The participant assumes authority they do not have. -
“What would have to change for you to switch?”
Signal: A real adoption trigger.
Ask next: “Tell me about the last time you changed a tool or process.”
Weak answer: The answer asks only for “better features” with no trigger or trade-off.
Use this guide as a menu, not a script to recite. Let the participant’s answer decide which probe comes next.
What Customer Discovery Interviews Should Prove
Customer discovery interviews should show whether a defined group has a recurring problem, how they handle it now, and what makes changing worth the effort.
Write one hypothesis before each set of calls:
We believe [specific customer group] struggles with [problem] when [trigger], causing [consequence], and currently uses [workaround].
For example:
We believe independent restaurant owners struggle with supplier-order errors during busy service periods, causing stockouts and manual reconciliation, and currently use WhatsApp messages plus spreadsheets.
The interview does not exist to prove that sentence correct. The interview exists to expose the parts that are wrong.
Keep one group per round. “Small business owners” is too broad if shop owners, restaurant owners, and agencies have different workflows and buyers. A focused group makes repeated patterns easier to see.
How to Avoid Leading the Customer
Leading questions suggest the answer you hope to hear. Neutral questions invite the participant to describe what already happened.
Use these replacements:
- Instead of “Would automation make this easier?” ask “How do you handle this today?”
- Instead of “Is this a frustrating process?” ask “What part takes the most effort?”
- Instead of “Would you pay for a tool that fixes it?” ask “What have you already spent trying to solve it?”
- Instead of “Would our product be useful?” ask “What would make you change the current approach?”
Open-ended, neutral questions reduce the chance that the interviewer supplies the answer through the wording. Semistructured interview guidance recommends clear, non-leading questions and adaptable probing.
Do not show a prototype during the problem-discovery part of the call. Once the participant sees a proposed solution, the conversation shifts from their existing behaviour to their reaction to your idea. That is a different test.
Ask Follow-Ups That Get Evidence
A short answer is an invitation to probe, not a reason to move on. Use a neutral prompt that asks for detail without adding your own interpretation.
Keep these six probes beside your notes:
- “Can you give me an example?”
- “What happened next?”
- “What do you mean by that?”
- “Who else was involved?”
- “How did you decide?”
- “What did that cost?”
The best probe is often silence. Pause after a short answer. The participant may add the detail that makes the response useful.
Record exact phrases where possible. Customers often describe the problem in language that can later improve your positioning, outreach, and product copy.
Spot False-Positive Demand Signals
Positive reactions are useful for rapport. They are weak evidence for a build decision.
Treat these statements as prompts for another question:
- “That sounds useful.”
- “I would probably try it.”
- “My team would love that.”
- “Keep me posted.”
- “I know lots of people who need this.”
Look instead for a dated event, a named workaround, a clear consequence, recent research into alternatives, past spending, or a known buyer. A person who describes last week’s failed process and the hours it cost offers more useful evidence than a person who says your concept is “great.”
Feature requests need the same discipline. When someone asks for a dashboard, ask what they currently cannot see, what decision is delayed, and who needs the information. The requested feature is a clue. The underlying problem is the finding.
Turn Interview Notes into a Decision
Complete this note after every call, before starting the next one:
- Participant and segment: Who did you speak with, and why do they fit?
- Recent event: What happened, and when?
- Current workaround: What do they do now?
- Consequence: What does the problem cost or risk?
- Buying path: Who uses, approves, pays, or blocks?
- Active evidence: What have they tried, searched for, or bought?
- Counterevidence: What contradicts your hypothesis?
- Next test: What must you learn next?
After several calls, group notes by the same segment and problem. Do not count compliments. Count repeated behaviours.
Use three outcomes:
- Test: Several people describe the same recent problem, workaround, consequence, and switching condition. Build the smallest test that checks the riskiest remaining assumption.
- Refine: The problem exists, but the segment, trigger, buyer, or consequence differs from your original hypothesis. Rewrite the hypothesis and interview that narrower group.
- Pause: People do not recall the problem, do not use a workaround, and face little consequence when it persists. Choose a different problem or segment before building.
If recruitment is the weak point, use a defined screener and source participants beyond friends, family, and your existing network. Our guide to recruiting study participants in India helps you set that process up.
When to Bring in an Interview Research Partner
Bring in a research partner when you need participants with a narrow profile, a neutral moderator, sensitive-topic interviews, or more than one research format. Founder-led calls work well for early learning. Independent moderation helps when your presence may shape the answer.
At Qualfacto, we support qualitative research formats including in-depth interviews, focus groups, digital diary studies, ethnography, and usability and app testing. Our participant panel helps connect eligible people in India with qualitative research studies, while our in-depth interview guide explains the one-to-one format in more detail.
FAQs
How Many Customer Discovery Interviews Should I Conduct?
Start with one clearly defined segment and keep interviewing until your notes reveal repeated patterns and meaningful differences. Stop treating the calls as one batch. Review the evidence after each few interviews, then sharpen the next question or recruit a more precise profile.
Can I Ask Customers What They Would Pay?
Yes, but ask about past spending first. “What have you paid for before?” and “What budget paid for it?” reveal more than a hypothetical price answer. Use direct price testing after you understand the problem, current alternatives, and buyer.
Should I Interview Friends and Family?
Use friends and family to practise your guide, not as your main evidence. Their relationship with you can make praise more likely. Label those calls separately and recruit people who actually experience the problem and influence the relevant purchase.
What Is the Difference Between Customer Discovery and Usability Testing?
Customer discovery investigates whether a problem matters and how people currently handle it. Usability testing evaluates whether people can use a specific design or prototype. Use discovery before you commit to a solution, then use moderated usability testing to improve the solution you decide to test.
