
Customer interview questions for startups: what to ask
17 customer interview questions, the ones to stop asking, how many interviews to run, a note template and a made-up interview done twice.
Most customer interviews go wrong in the same way. The founder describes the idea, asks whether the person would use it, and goes home with a polite yes. People are kind to someone who is excited about a project, and questions about the future get guesses back.
The fix is in the questions. Below are customer interview questions for startups that ask about what people did, in the order you use them, plus how to find people, how many to talk to, a made-up interview done badly and then well, and a note template.
Three rules for every interview
These rules come from Rob Fitzpatrick's book The Mom Test, a framework for asking customers about their problems without leading them. The book is short and worth reading in full, and it's on the list of the best books for startup founders. In short:
- Talk about their world, not your idea. Once you describe what you're building, people react to your idea and stop telling you about their week.
- Ask about real events in the past, not about the future or in general. "When did this last happen?" gets a fact. "Would you ever...?" gets an opinion.
- Listen more than you talk. If you talk for more than a few minutes of a 20-minute call, the call was about you.
Who to talk to, and how to find them
Talk to people who have the problem now and who would decide to pay for a fix. In a business, those can be two people: the one who lives with the problem every day and the one who signs off the spending. You want both.
Friends and family only count if they really are your customer, and even then they will go easy on you.
Places to find them:
- One introduction away. Ask everyone you know a specific question: "Who do you know who runs a small bakery?" Specific asks get names. "Do you know anyone who might be interested?" gets nothing.
- Where they already are. Industry forums, online groups, LinkedIn, trade events, local meetups. For a local business, walking in at a quiet hour often beats any message.
- At the end of every interview. Ask, "Who else should I talk to about this?" A good conversation can lead to two more.
The ask itself should be short, plain and about them. Say what you want to understand, ask for 20 minutes, and say you're not selling anything, because you're not. For example (made up):
Hi Ana, I'm looking into how small bakeries handle orders from cafés, and Marta at the Saturday market said you supply a few. Could I ask you about it on a 20-minute call this week? I'm not selling anything. I'd like to hear how it works for you today.
How many customer interviews to do
My rule: keep having conversations until you stop hearing anything new. The number depends on the industry and on who your customer is. As a general rule it's hardly more than 20 or 25. Sometimes 15 is enough, sometimes 10.
For a consumer app, where the number of possible customers is massive, aim for 20, 25 or 30. For something like a biotech product, where there aren't many potential customers to talk to, between 5 and 10.
One way to spot the point where nothing new comes up: after each call, mark the answers you had not heard before. When several calls in a row add nothing to that list, you have the picture. Only count conversations with people who have the problem.
How to run a 20-minute interview
Before the call, write down the three things you most need to find out, and the belief that would sink the idea if it were wrong. Pick five or six questions from the bank below. You won't get through all 17.
During the call:
- Open (1 to 2 minutes). Who you are, what you're trying to understand, that there are no wrong answers. Ask if you may take notes. Record only with their permission.
- Their world (3 minutes). How their week runs and where the problem fits in it.
- The last time it happened (about 10 minutes). This is most of the call. Get one real story and follow it: "What happened next?", "Why?", "Can you show me?"
- What they do about it today (3 to 4 minutes). Tools, people, workarounds, money spent, things they tried and dropped.
- Close (2 minutes). Ask what you should have asked, who else to talk to, and whether you can come back to them later.
If they ask about your idea, tell them briefly at the end, then stop talking and watch what they do. A question like "When can I try it?" is worth more than any compliment.
If there are two of you, one asks and the other writes. Write up your notes within the hour, while you can still hear their words. An AI tool can help you sort them later; what an AI co-founder can and can't do covers why the conversation itself stays with you. If you're trying out a potential co-founder, a round of interviews run together makes a good trial project; questions to ask a potential co-founder covers what to watch for while you do it.
Customer interview questions for startups: the list
Swap "this" and "X" for the problem in your own words. Questions in each group go from wide to narrow.
Their world
- Walk me through a normal week in your job. Where does X fit in?
- What tools, people or services do you use for X?
The last time it happened
- Tell me about the last time X happened.
- What happened just before? And what did you do next?
- How long did it take you, and what did it cost?
- How many times did that happen in the last month?
What they do about it today
- What do you do about this today?
- What have you tried that didn't work? Why did you stop?
- What do you pay for this now, in money or in hours?
- If you use a spreadsheet or a workaround for it, can you show me?
How much it matters
- What happens when it doesn't get fixed?
- Where does it rank against the other problems on your plate this month?
- Have you looked for a fix? What did you search for?
- Who would have to agree before you spent money on it?
Closing
- Is there anything I should have asked and didn't?
- Who else should I talk to about this?
- Can I come back to you when I have something to show?
Question 7 is often the most useful of the lot. Expect three kinds of answer: a tool stretched past its purpose (often a spreadsheet), a person patching it by hand, or a shrug. The first two are a sign the problem is real. The shrug means it may not be painful enough to pay for.
Questions to stop asking, and what to ask instead
| Instead of | Ask | Why |
|---|---|---|
| "Would you use an app that does X?" | "How do you handle X today?" | Nobody knows what they will do. Everyone knows what they did. |
| "Do you think this is a good idea?" | "Tell me about the last time X happened." | An opinion costs them nothing to give. A story has details you can check. |
| "Would you pay €50 a month for this?" | "What do you spend on this today, in money or hours?" | A price you suggest gets a polite yes. What they already spend is a fact. |
| "What features would you want?" | "Can you show me how you do it now? Which part annoys you most?" | People are bad at designing products and good at describing their frustrations. |
| "Wouldn't this save you a lot of time?" | "How long does this take you now?" | The first question tells them the answer you want. |
| "How often would you use it?" | "How many times did this happen last month?" | The future is a guess. Last month is a number. |
How to read the answers
Some answers sound good and mean very little:
- Compliments and "that's interesting". Usually manners, sometimes confusion. Thank them and go back to their story.
- "I'd definitely use that." Ask what they would give up to get it: their current tool, some money, an hour for a trial.
- Vague words, or no recent story. "Sometimes" and "now and then" are warning signs, and so is not remembering the last time. Real problems come with dates and numbers.
- Never having tried to fix it. If it hurt, they would have tried something.
And some answers are worth a lot:
- A specific, recent story with details they didn't need to give.
- Money or hours already going into a workaround.
- A next step that costs them something: a second call with their boss, a trial, a pre-order.
Interviews tell you whether the problem is real and how much it hurts. They don't prove that people will pay. That proof is a commitment: money, time, a pilot, a letter of intent. How to validate a startup idea covers what counts as enough evidence and what to do with your interview notes.
A made-up example: the same interview, twice
This example is invented, the bakery and every number included. Imagine a founder who wants to build ordering software for small bakeries that supply cafés.
The bad version:
Founder: We're building an app where cafés order from you online instead of texting. Would you use something like that?
Baker: Oh, that sounds great. Yes, definitely.
Founder: And would you pay €30 a month for it?
Baker: Probably, yeah, if it works.
The founder leaves with two yeses and nothing true.
The good version, with the same baker:
Founder: Tell me about the last time a café order went wrong.
Baker: Tuesday. A café texted at eleven at night to cut their order. I didn't see it until five in the morning, when the croissants were already proofing. I made 40 they didn't want.
Founder: What happened to those 40?
Baker: Sold some at the counter. Threw the rest away.
Founder: How often does that happen?
Baker: Once or twice a week. Usually early in the week.
Founder: What have you tried to stop it?
Baker: My partner checks the phone at ten every night. Last year we set up an online order form. The cafés never used it. They just kept texting.
The first interview said "build the app". The second changes the plan. The problem is late changes to orders, not ordering itself, and the cafés already ignored an order form, so a new app for them will probably fail the same way. Whatever the founder builds has to work with the texts they already send.
A note template for each interview
Fill in one note per call, with the same eight fields every time, so you can lay them side by side later. Here is the note for the baker's interview above. "Who" is a role and type of business, no names needed. Quote the problem and the words worth keeping exactly, and give urgency your own score from 1 to 10 after the call, with one line of why.
| Field | Note |
|---|---|
| Who | Owner of a small bakery that supplies cafés |
| The problem, in their words | "A café texts late to change the order and I don't see it until the croissants are already proofing." |
| How often | Once or twice a week |
| What it costs them | Up to 40 pastries each time, sold cheap or thrown away, plus a nightly phone check |
| What they do today | Partner checks texts at 10 at night. Tried an online order form; cafés ignored it. |
| Words worth keeping | "Threw the rest away." "They just kept texting." |
| Urgency, your score from 1 to 10 | 7. It costs money every week and she has already tried to fix it. |
| Next step | Happy to talk again; gave two names of other bakers |
After the interviews
Three things to take from your notes:
- What they do today is your real competition. The spreadsheet, the person doing it by hand and the shrug are what you have to beat. Market research for startups shows how to lay them out next to the other options.
- Their words are your words. The phrases in the "words worth keeping" row belong in your startup one-liner and on your website.
- These people come first when you have something to show. Someone who described the problem to you is not a stranger when you come back with a fix. How to cold email investors ends with how to write to first customers, and it applies here. Once they use what you build, interview them again about real use; how to build an MVP covers the changes that follow.
Ready to Move Faster?
Get actionable feedback and insights on your challenges and goals. 30 minutes focused on helping you move forward.
No sales pitch. No obligations. Just helpful advice.