How to run a customer interview

Customer interviews are the cheapest and most useful research a startup can do. They need no budget, only time and the right questions. But a badly run interview is useless or even harmful: it tells you what you want to hear and sends you in the wrong direction.

The point of an interview is not to get praise for your idea or to sell your product. The point is to understand how the person lives or works, where the problem sits and how painful it is. So for most of the conversation the customer should be talking, not you. A good rule: the less you talk, the more you learn.

People are bad at predicting their own future behaviour. Ask "Would you use an app like this?" and most will politely say yes, then never use it. The past, on the other hand, is made of facts.

Bad question: "Do you need software that tracks the stock in your warehouse?" Good question: "When was the last time you noticed a product had run out? What happened then?"

The last question matters most: every good interview should lead you to the next one.

Speak whichever language the customer is comfortable in, Uzbek or Russian. If you can, go to their place of work: their shop or office. People show and tell you more in their own surroundings. Keep it short; twenty to thirty minutes is enough. Do not ask a busy shop owner for an hour.

Hospitality and politeness run deep here. Someone may say "great idea, I will definitely use it" simply to avoid upsetting you. Do not treat that kind of praise as evidence. Pay attention to facts and commitments: what the person has actually done, what they have paid for, whether they are willing to give you more of their time.

Voice or video calls on Telegram work too, especially with customers in the regions. But questionnaires and surveys do not replace conversations: they answer the questions you thought of in advance and never reveal what you did not know to ask.

Write things down straight after each interview. Record the customer's exact words as closely as you can, especially how they describe the problem. Keep one note per interview: who, where, the main problems, the current workarounds, notable quotes. If two of you go, one asks and the other writes.

After five to ten interviews, lay your notes side by side and look for what repeats. One person's opinion may be chance; a problem that comes up with five people is a pattern.

If none of these happen, the problem may not bother them enough, however pleasant the conversation was.

For many founders the hardest part is the first message. Keep it short and honest: "Hello! I am researching how shops keep track of goods sold on credit. I am not selling anything; I would just like to hear about your experience. Could you spare twenty minutes this week?" People like talking about their work, especially when they sense genuine curiosity rather than a sales pitch. If someone says no, do not take it personally. That is normal; message the next person.

Schedule five interviews this week. Write down five core questions beforehand, none of them hypothetical. Write your notes within ten minutes of each interview. At the end of the week, pick out the three problems that came up most often.