How to talk to customers so they don't lie to you

People don’t lie to you on purpose. They’re polite, they remember their own habits badly, and they’ll happily speculate about the future if you ask them to. A bad interview collects exactly that: compliments, promises and opinions. A good one collects facts about what the person already did, what it cost, and what they tried instead of you.

Who to talk to

Not friends, not colleagues. They’re on your side, and you can hear it in every answer. You need people who have the problem now or had it recently. Look where they complain: industry chats, forums, reviews of competitors, your own support inbox. Keep the ask short: you’re researching how people handle a specific task, it takes thirty minutes, and you’re not selling anything. Then actually don’t.

Ask about the past

The core rule: ask about what happened, not what might happen. “Would you use this?” is a question about the future, and the answer is worth nothing. “When did this last happen?” is a question about the past, and a specific story follows.

Questions that work:

• When was the last time you ran into this? Walk me through it.
• What did you do? What did you use?
• How much time or money did it take?
• What did you try before, and why did you stop?
• What was the most annoying part?

When the answer drifts into the general (“usually I…”), steer back to a specific case: “and last Tuesday, how did that go?” “Usually” is how people see themselves. Last Tuesday is what they actually do.

Also notice whether they ever looked for a solution. If the problem has bothered them for years and they never searched for or tried anything, they can live with it. The ones who’ve already built a workaround out of spreadsheets and calendar reminders are your audience.

Don’t pitch

The moment you show the product or describe the idea, the conversation changes. The person switches into polite-guest mode: they like it, it’s interesting, they’ll definitely try it. None of those words mean they’ll pay.

So don’t describe the idea at all until you’ve heard everything about the problem. If you must show it, do it in the last five minutes and treat the reaction as a bonus, not data. A compliment isn’t a fact. “I’ll buy it when it’s out” isn’t a fact either. A fact is something the person already did, or something they’re willing to do right now: give you access to their data, introduce you to a colleague, pay a deposit.

The other trap is leading. “That must be frustrating, right?” guarantees a yes. Ask an open question and sit through the pause. People tend to add the most useful detail when given room.

What to write down

Write down words, not conclusions. “He finds it inconvenient” is a conclusion. “Every Friday he spends two hours merging three spreadsheets by hand” is a note. A week later, conclusions blur into your hopes; words don’t.

One page per conversation works well:

• who they are and what situation they’re in;
• the last specific instance of the problem;
• how they handle it now and what that costs;
• what they tried and why they dropped it;
• exact phrases that stood out;
• the next step, if there is one.

Record the call if they agree, but write the summary right away, while you remember the tone. Two people beat one: one asks, the other writes.

When to stop

After each conversation, reread your notes and mark what repeats. When new interviews stop adding anything new, the same complaints and the same workarounds, you have enough for that segment. If every person tells a different story, you’ve probably mixed several segments, or there’s more than one problem. Narrow the group and start again.

Interviews aren’t a way to confirm an idea. They’re a way to find out where you’re wrong while the mistake costs half an hour of conversation instead of six months of development.

SUBXCAT can help you put together an interview script and work through the notes so they turn into decisions rather than a folder of transcripts.