Customer Interviews That Don't Lead the Witness
By CodexierPublished 5 min read
Customer interviews are the cheapest way to avoid building the wrong thing, and one of the easiest to get wrong. People are kind. Ask them whether they like your idea and most will say yes, because saying no to someone's face is uncomfortable. The result is a notebook full of encouragement and no evidence. This guide covers how to ask about real past behaviour instead, which questions lead and how to rephrase them, how to recruit in Sweden, and how to turn a handful of conversations into decisions.
Ask about the past, not the future
The mechanism is well documented in Rob Fitzpatrick's book The Mom Test: questions about hypothetical futures invite optimism, while questions about specific past events force concrete answers. 'Would you pay for a tool that does X?' is a guess about a future self. 'How did you handle X last month, and what did it cost you?' is a memory, and memories contain the details you need.
Questions that lead and how to fix them
| Leading question | Why it fails | Better version |
|---|---|---|
| Would you use an app that does this? | Asks for a prediction and invites politeness | How do you handle this today? |
| Don't you find invoicing frustrating? | Suggests the answer | Tell me about the last time you sent invoices. |
| How much would you pay for this? | People guess low or high to be nice | What do you spend on this now, in money or time? |
| Is this feature important to you? | Everything sounds important when asked | When did you last need this? What did you do? |
| Do you like the design? | Collects opinions, not behaviour | Show me how you would do the task here. Think aloud. |
Follow-up questions do most of the work: 'Why was that?', 'What happened next?', 'Can you show me?'. Silence helps too. Wait a few seconds after an answer and people often add the most useful part.
Recruiting and scheduling
- Start with people who have the problem now, not friends who are willing to help.
- Existing customers, LinkedIn contacts in the right role and industry groups are good sources in Sweden.
- Offer a small thank-you, such as a gift card, and say how long the call takes. Thirty minutes is easier to accept than an hour.
- Mix roles: the person who uses a tool daily and the person who pays for it often describe different problems.
- Book video calls with screen sharing, so you can ask people to show rather than describe.
If you record, tell the person in advance, ask for consent at the start of the recording, say what it is used for and when it will be deleted. Recordings are personal data under GDPR, and a clear retention period is the simplest way to handle them.
Taking notes and spotting patterns
Have a second person take notes if you can, so the interviewer can listen. Write down quotes, not interpretations, and mark what was actually observed versus claimed. Directly after each interview, spend ten minutes writing the three most important things you learned while the memory is fresh.
Pain
Problems they described with real consequences: time lost, money spent, customers lost.
Workarounds
Spreadsheets, manual routines or paid tools they already use. A workaround proves the pain is real.
Triggers
Events that made them look for a solution. These tell you when to reach them.
Language
The words they use. Reuse them in your website and product.
Turning interviews into decisions
Interviews are worth nothing if they do not change a decision. Before you start, write down what you believe and what would change your mind. After five to eight interviews, sort your notes on a wall or board and check each belief: confirmed, weakened or unclear. Then decide what to build, what to drop and what to test next, for example with a clickable prototype. Our guide to testing a prototype before code covers that step.
Our product design sprint combines interviews, a prototype and user tests in a short, structured process, with prices on our pricing page. When you do not need outside help: if you have direct access to your customers and time to do the interviews yourself, the method above is enough to start. Book a call if you want a second pair of eyes on your questions.
Frequently asked questions
How many customer interviews do I need?
Five to eight with the right people usually reveal the main patterns. Stop when you start hearing the same stories and learn little new.
Should I show my prototype in the interview?
Not in the first part. Learn about their current situation first, then show it at the end if you want reactions, and treat those reactions as weak evidence.
Can I interview existing customers?
Yes, they are often the best source. Make it clear that it is not a sales call, and ask about their work rather than your product.
Do I need consent to record a customer interview?
Inform the person and get their agreement before recording, state the purpose and delete recordings when they are no longer needed. That keeps you in line with GDPR.
About to build something new?
Book 15 minutes. We look at your idea and your assumptions, and suggest which questions to ask customers before you spend money on development.
Book a free 15-minute call