A customer interview script that gets past polite answers
A 30-minute customer interview script: questions about what people already did, follow-ups for vague answers, a worked example and how to score your notes.
In this note
A customer interview script that gets past polite answers asks about what the person has already done, not what they think of your idea. Open with "Walk me through the last time you dealt with [the problem]", follow every vague answer with "When was that? What happened?", ask what they have tried and what it cost them, and only bring up your product near the end, alongside a small ask such as a refundable pre-payment. Compliments, "I'd use that" and "sounds great" tell you nothing; a specific story, a workaround they built, or money already spent tells you almost everything.
Below: the full 30-minute script, follow-ups for vague answers, a worked example, and prompts for an AI agent to handle the scheduling and tallying.
Why do customers give you polite answers?
Because you asked them to judge your idea, and people are kind. Ask "Would you use an app that does X?" and most will say yes, since saying no to someone's project feels rude and costs them nothing. Rob Fitzpatrick's book The Mom Test (2013) is built on this: even your mom will lie to you if you ask whether your business idea is good, so stop asking.
The fix is three rules from the same book. Talk about their life, not your idea. Ask about specific things that already happened, not opinions about the future. Talk less and listen more. Every question in the script below follows them.
- "Would you use an app that…?"
- "Do you think this is a good idea?"
- "How much would you pay for this?"
- "Is [problem] a big problem for you?"
- "What features would you want?"
- "Would you recommend it to a friend?"
- "Walk me through the last time that happened."
- "What did you do about it?"
- "What do you use for that today? What does it cost?"
- "What have you already tried? Why did you stop?"
- "What happens if you just don't deal with it?"
- "Who else should I talk to about this?"
The 30-minute interview script
Keep the script on screen, but treat it as a checklist rather than a questionnaire. If they start telling a good story, let them, and come back to the list when it ends.
- Minutes 0–3Set it up"I'm not selling anything today. I'm trying to understand how you handle X."
- Minutes 3–15Get the last storyThe last time the problem happened, start to finish.
- Minutes 15–22Find the costWhat they've tried, what they pay, what it costs when it goes wrong.
- Minutes 22–27Show, then askYour one-line idea, then a small, specific ask.
- Minutes 27–30Close the loop"Who else should I talk to?" and permission to follow up.
1. Set it up (3 minutes). Thank them, say it's 30 minutes and you're not selling anything. Ask before you record; if they'd rather you didn't, just write. Warm up with: "Tell me a bit about your work and where [the problem area] fits in."
2. Get the last story (12 minutes). This is the interview. Ask: "Walk me through the last time you had to [deal with the problem]." Then keep them in that one story with follow-ups: "What did you do first?", "What happened next?", "How long did that take?", "Who else was involved?" A real problem comes with a date, a tool, a workaround and some annoyance. If they can't remember a recent time, that's your answer too.
3. Find the cost (7 minutes). Now ask about money and effort:
- "What do you use for this today? What does it cost you?"
- "What have you tried that didn't work? Why did you stop using it?"
- "What happens when it goes wrong? Has that ever cost you a client or a payment?"
- "If you could wave a wand and fix one part of this, which part?"
People who have already paid for a fix, or built their own spreadsheet to cope, are your market. People who have never tried to solve it probably won't pay you to.
4. Show, then ask (5 minutes). Only now describe what you're building, in one or two sentences, and stop talking. Watch for a question back ("Does it work with…?", "When can I use it?"), which means more than any compliment. Then make the ask. Pick one, and say it plainly:
- "I'm taking founding-member payments: $20 now, taken off your first year and refundable any time. Want in?"
- "Can I send you the beta in three weeks and get 20 minutes of your time to watch you use it?"
- "Would you introduce me to two other people who deal with this?"
A yes to any of these is a commitment. "Sounds great, keep me posted" is a polite no.
5. Close the loop (3 minutes). "Who else should I talk to about this?" turns one interview into two. Ask whether you can follow up by email, and write your notes within the hour, before you forget which story was whose.
The follow-ups that turn vague answers into real ones
Most of the useful information arrives in the second question, not the first. When you hear one of these, don't write it down as a finding. Follow it up.
| They say | What it usually means | Ask next |
|---|---|---|
| "I'd definitely use that" | A compliment about the future, not a commitment | "What do you use for this today?" |
| "It's a real pain" | A feeling with no evidence yet | "When did it last happen? What did you do?" |
| "I usually just…" | A habit described in general, which hides the details | "Can you show me, or walk me through last week's?" |
| "I'd pay for that" | A guess about a price they'll never be asked to pay | "What are you paying for it now, in money or time?" |
| "It would be great if it also did X" | A feature request, often off the main problem | "Why do you need X? What happens without it?" |
| "Everyone in my field has this problem" | Their opinion of other people | "Who specifically? Could you introduce me?" |
| A long story with a date, a tool and a workaround | A real problem | Let them finish, then ask what it cost them |
A worked example: an app for freelance bookkeepers
Say your idea is an app that chases a bookkeeper's clients for receipts and bank statements each month. You found the complaint in reviews of the big practice-management tools and in bookkeeping subreddits, the way mining Reddit for micro-SaaS ideas describes. Here is the same call run two ways.
The polite version. "Would you use an app that automatically chases clients for documents?" "Oh, definitely." "How much would you pay?" "Maybe $20 a month?" You hang up pleased, with no evidence the problem costs her anything and a price she made up on the spot.
The real version. "Walk me through the last time you closed a client's month." She describes last Thursday: three clients late, a spreadsheet with a column for "chased?", four reminder emails she wrote herself, and one client who sent photos of receipts by text at 11 PM. "What do you use to chase them today?" Email and a Google Sheet; she tried her practice software's client portal, but clients never logged in. "What does it cost you when they're late?" She pays a part-time assistant about five hours a month to chase, and twice this year a client's tax filing went in late because the paperwork came after the deadline.
Now you know the problem is real (a specific Thursday), what she tried and why it failed (clients won't log in), and what it costs her (an assistant's hours and late filings). You also learned what the polite version never would: clients reply to texts, not portals. That's your angle. You describe it in one sentence and ask for a $20 refundable founding payment. Whatever she says, it's a better answer than "definitely".
How do you read your notes after 10 interviews?
Look for repetition, not enthusiasm. After about ten calls, the same two or three complaints should keep coming up in people's stories. If every interview surfaces a different problem, your audience is too broad: narrow it to one kind of customer and run the next ten.
Score each interview on what happened, not how it felt:
- Strong: a recent, specific story, money or hours already spent on a fix, and a yes to your ask.
- Medium: a specific story and a workaround, but no to the ask. Note why they said no.
- Weak: only opinions and compliments, or no recent example. Don't count these toward your total.
What people already pay, in money or time, is also where your price comes from; pricing your first SaaS picks up from there.
Mistakes that waste a good interview
- Interviewing friends. They'll be generous and you'll believe them. Talk to people who have never met you; finding your first 100 users covers where.
- Asking two questions at once. They'll answer the easier one. Ask one, then wait through the silence.
- Taking a feature list home. Features are their guess at a solution. Write down the problem behind each one.
- Letting the agent hold the call. It can find people, book times and tally notes. The conversation is where you learn what to build, so it stays yours.
Where this fits
Interviews sit in the middle of validation: validating an app idea in a weekend shows how to fill a week with calls, and the asks at the end of each one lead into getting your first paying customer.
Two things make the next 20 calls easier: an agent that keeps your interview list and notes in files between runs and messages you on Telegram or WhatsApp when someone needs a reminder, and other founders to show your script to before you use it. Heyron members get both: their own agent on its own always-on computer, and the Solo Software Studio community, a place to post your script and notes, see what's working for other solo founders and work toward a first paying customer together. See what's included.
Frequently asked questions
What questions should I ask in a customer interview?
Ask about the last time the problem happened, what they did about it, what they use today and what it costs them, and what they've tried that didn't work. Avoid "Would you use…?" and "How much would you pay…?", because those ask for opinions about the future, which people get wrong.
How long should a customer interview be?
About 30 minutes. That's enough for one or two detailed stories and a small ask at the end. If a call goes well and they're happy to keep talking, let it run, but book 30 so more people say yes.
How many customer interviews do I need?
Plan for 20 to 30. The same problems usually start repeating after about ten; if they don't, narrow who you're talking to. Stop when new interviews stop telling you anything new.
Should I show my product in a customer interview?
Only at the end, in one or two sentences, after you've heard their story. Showing it first turns the call into a pitch, and people start reacting to your idea instead of describing their problem.
Can an AI agent do customer interviews for me?
It shouldn't hold the conversation; that's where you learn what to build. It can find people who are talking about the problem, keep your interview list, draft reminders for you to send, and turn your notes into a table of who said what.