How to Do Customer Interviews That Actually Help

Why most customer interviews tell you nothing

The first time most founders talk to customers, they walk away thrilled. Everyone loved the idea. Everyone said they'd use it. Then the product ships and nobody shows up. The interviews were not the problem; the way they were run was.

People are polite, especially in India where saying a flat no to a friend or acquaintance feels rude. When you describe your shiny idea and ask "would you use this?", they hear "please say yes" and oblige. That is not data. That is encouragement, and encouragement does not pay your bills.

The fix is simple to state and hard to do: stop asking about the future and your idea, and start asking about the past and their life. What people have already done is the only reliable signal you have. The goal of a customer interview is not to validate your solution. It is to understand a problem deeply enough that you know what to build.

Talk to the right people, not the easy ones

The easiest people to interview are your friends, family, and the founder WhatsApp group. They are also usually the wrong people. If your product is for small kirana store owners, talking to your engineer friends tells you nothing useful.

Define who actually has the problem before you book a single conversation. Be specific. Not "small business owners" but "salon owners in tier-2 cities who manage bookings on paper". The narrower you are at the start, the clearer your answers will be. You can widen later.

Aim for 10 to 15 conversations in a tight group before you draw conclusions. Fewer than that and you are reading noise. You do not need a representative sample or fancy methodology at this stage. You need enough real conversations that patterns start repeating themselves. When the third and fourth person describe the same frustration in the same words, you are onto something.

• Find people through communities they already gather in: local trade associations, WhatsApp and Telegram groups, LinkedIn, industry events.
• Offer a short, specific ask: 'Can I get 20 minutes to understand how you handle X today?' Not 'Can I show you my startup?'
• Avoid only interviewing people who already agree with you. Deliberately seek out a few who might not.

Ask about the past, not the future

This is the single most important habit in customer interviews. Questions about the future ("Would you buy this?", "Would you pay 500 a month?") produce fantasy answers. Questions about the past ("What did you do the last time this happened?") produce facts.

Someone telling you they would pay for a solution costs them nothing. Someone telling you they already spend three hours every Sunday reconciling payments in a notebook is handing you gold. The second answer proves the problem is real, frequent, and painful enough to act on. Dig into the specifics of that real story rather than chasing hypotheticals.

If you want a fuller framework for low-cost validation, the principles here pair well with how to validate a startup idea without spending money. The interview is one tool in that larger toolkit, and it works best alongside cheap experiments rather than on its own.

• Good: 'Walk me through the last time you tried to solve this.'
• Good: 'What are you using today? Why that?'
• Good: 'What's the most frustrating part of how you do this now?'
• Avoid: 'Would you use an app that did this?'
• Avoid: 'Do you think this is a good idea?'
• Avoid: 'How much would you pay for this?'

Don't pitch. Shut up and listen.

The strongest urge in any founder interview is to explain the idea. Resist it. The moment you start pitching, the conversation flips: now they are reacting to your solution instead of describing their world. You learn nothing, and you have primed them to be polite.

Aim to talk for less than a quarter of the time. Ask a question, then stay quiet. Silence feels awkward, and people fill it with the honest detail they were about to skip. When someone gives a vague answer, your best follow-up is almost always "why?" or "tell me more about that".

Keep your idea out of it for as long as you can. If they ask what you are building, you can say "I'm exploring this space and trying to understand how people deal with it today" and steer back to them. You will have plenty of time to test your solution later. The interview is for understanding the problem first. The difference between the two is worth getting straight, and problem validation vs solution validation breaks it down clearly.

What to actually listen for

Not all problems are worth building for. As you run interviews, you are filtering for signals that a problem is real and worth money. The clearest signal is that people already spend something to solve it: time, money, or a clumsy workaround they have rigged up themselves.

Watch for emotion. When someone gets visibly annoyed describing how they handle something today, that is a problem worth your attention. Flat, neutral descriptions usually mean the pain is not big enough to drive a purchase. You want to find the thing that makes them lean forward and complain.

• They already pay for something to address it, even partially.
• They have built a workaround (a messy spreadsheet, a manual routine, a person hired just for this).
• They bring up the problem before you do.
• The problem happens often, not once a year.
• They get emotional or animated when describing it.

Run the interview without breaking the flow

You do not need a rigid script. You need a short list of topics and the discipline to follow the conversation where it goes. The best interviews feel like a curious chat, not an interrogation. Record audio if they allow it, so you can stay present instead of scribbling.

Right after each conversation, before you do anything else, write down the three or four things that surprised you and any exact phrases they used. Their words matter more than your summary, because their language is what you will later use in your marketing and your product copy. Memory fades fast, so capture it while it is fresh.

1. Open warmly and set context: 'I'm trying to understand how people like you handle X. No right answers, I just want to learn.'
2. Ask them to walk you through the last time they dealt with the problem, step by step.
3. Probe each step: what was annoying, what took long, what they wished was different.
4. Ask what they tried before, what they use now, and why they switched or stuck.
5. Only at the very end, if relevant, briefly mention what you're exploring and watch their genuine reaction.
6. Thank them, ask if they know one or two others worth talking to, and write your notes immediately.

Turning interviews into decisions

Fifteen conversations later, you should be able to answer a few hard questions. Is there a specific, frequent problem that several people described independently? Are they already spending time or money on it? Is the group of people who have it big enough and reachable enough to build a business around?

If the answer is yes, you now know what to build first, in their words, and you can move toward a small, sharp first version. If the answer is no, you have saved yourself months of building something nobody wanted. That is a win, not a failure. Either way, interviews feed directly into the rest of validation work, including how to find your first 100 users, because the people you interviewed are often your earliest customers.

Watch for the trap of hearing what you want to hear. If only one person out of fifteen is excited and the rest are lukewarm, that one person is not your market, they are your exception. Be honest with the patterns. The discipline of running interviews properly is exactly the discipline that separates founders who build the right thing from those who build the thing they fell in love with.