In this guide
Building is the fun part, and with an AI assistant it's faster than ever. That's exactly the risk: you can now build a polished product for a problem nobody has in a weekend. Ten short conversations take about the same time, and they tell you whether the weekend is worth it.
The goal of these calls isn't to pitch, validate or collect compliments. It's to hear, in people's own words, how they deal with a specific problem today, and how much it costs them. That's harder than it sounds, because people are kind about ideas.
Pick a narrow group
"Small businesses" is not a group. "Bookkeepers who work alone with 20 to 50 clients" is. "Barbers who take bookings by text message" is. The narrower the group, the more likely you'll hear the same story repeated, and repetition is the signal you're looking for.
A good group has three things:
- A shared task you can name in one sentence.
- A place they gather, online or off, where you can find more of them.
- Money they already spend on tools for their work. People who pay for nothing are hard to sell to.
Where to find them
- Your own network first. Ask friends: "Do you know anyone who works as a [role]?" Introductions get the most yeses.
- Online groups and forums where that profession talks shop. Read for a week before you post anything, and follow each group's rules.
- Professional directories and local listings, for trades and local businesses.
- Former colleagues and clients, if you're a freelancer who has done this work for others.
What to send
Short, specific, honest. Something like:
Hi Maria, I'm trying to understand how independent bookkeepers handle collecting receipts from clients. Could I ask you a few questions for 15 minutes this week? I'm not selling anything; I'm deciding whether to build something.
Expect about one in three people you message to agree. Thirty messages gets you ten calls. Offer two specific time slots rather than "whenever suits you".
The one question that matters
"Walk me through the last time you [did the task]."
Then stay quiet and let them talk. This approach is at the heart of Rob Fitzpatrick's book The Mom Test[1]: ask about specific past behaviour, because people describe what they actually did honestly, and describe what they might do politely.
Follow-ups that keep the story concrete:
- "When was that, roughly?"
- "What did you do next?"
- "How long did that take?"
- "What's the most annoying part?"
- "What have you tried to make it easier?"
- "What did that cost you, in time or money?"
- "Who else is involved when this happens?"
A fifteen-minute call, minute by minute
| Minutes | What you do |
|---|---|
| 0–2 | Thank them, repeat that you're not selling, ask if you can take notes. |
| 2–4 | Context: "Tell me a bit about your work. How many clients, what does a week look like?" |
| 4–11 | The story: "Walk me through the last time you…" and follow-ups. |
| 11–13 | Current solutions: "What do you use for this today? What do you pay for it?" |
| 13–15 | Close: "Is there anyone else I should talk to? Can I come back to you if I build something?" |
Notice what's missing: describing your idea. If they ask, tell them briefly at the end. If you describe it early, the rest of the call becomes them being polite about it.
Questions to avoid
| Don't ask | Why | Ask instead |
|---|---|---|
| "Would you use an app that…?" | People are kind about hypothetical ideas | "How do you handle this today?" |
| "Is this a big problem for you?" | Invites a yes to be helpful | "When did it last happen? What did it cost?" |
| "How much would you pay?" | Guesses about future spending are unreliable | "What do you pay for now to deal with it?" |
| "Don't you hate it when…?" | Leads them to your answer | "What's the most annoying part?" |
Real pain vs polite interest
Signs of a real problem
- It happened recently, and they remember details
- They've already tried to fix it: a spreadsheet, a template, a tool they dislike
- It costs time or money they can name: "three hours every Friday"
- They ask you: "Are you building something for this?"
Signs of polite interest
- "That sounds useful", with no story behind it
- They can't remember the last time it happened
- They've never looked for a solution
- They talk about other people having the problem, not themselves
How to take notes
Right after each call, write down four things while they're fresh:
- The story of the last time, in a few lines.
- Their exact words for the problem. These become your headline later.
- What they use today, and what it costs.
- A score: 2 for real pain, 1 for mild, 0 for none.
Hypothetical example of one call's notes
Story: Last Friday, spent the afternoon emailing six clients for missing receipts. Two replied, one sent a blurry photo. Finished on Sunday.
Her words: "chasing", "every single Friday", "the same three clients".
Today: email templates, a shared folder, phone calls. Pays for accounting software, nothing for this.
Score: 2. Asked twice whether I was building something.
Deciding what's next
After ten calls, add up the scores and read the quotes side by side.
| What you heard | What to do |
|---|---|
| 6+ people describe the same recent, costly problem | Move to a one-page test with their words as the headline |
| A few strong stories, but all different | Narrow the group and do five more calls |
| Strong pain, but they already pay for a good solution | Find out what's missing from it; that's your angle, or a reason to stop |
| Mostly polite interest | Drop it, and be glad it cost ten calls instead of three months |
Calls that tell you nothing
- Pitching instead of listening. If you talked more than a third of the call, it was a pitch.
- Calling friends who aren't in the group. Their opinions are kind and irrelevant.
- Stopping at three calls because the first three were enthusiastic.
- Hearing what you hoped. Read your notes the next day, and score them again.
Questions
Can I do this by survey instead?
Surveys tell you what people choose from your options. Calls tell you what they actually do, in words you'd never have thought to offer. Do calls first; surveys can help later, once you know what to ask.
What if I can't get ten calls?
That's information too: if a group is hard to reach for a 15-minute chat, it will be hard to reach with a product. Try a different channel or a different group.
Should I offer to pay for their time?
For busy professionals, a small gift card can help. But people who'd only talk for money are less likely to be your customers.
Write your one-sentence group, then send thirty messages this week. When you have six strong stories, put their words on a single page: a one-page test that costs $0.
Sources
Checked 14 November 2026. Numbers in brackets in the text point to these sources.
