Interview customers so they tell you the truth

People are polite about your idea and honest about their past. So don't ask what they'd do — ask what they already did, what it cost them, and what they're paying to work around it today. Ten conversations like that will save you from building something nobody wants, and they're free.

Do this

1. Leave your idea at the door. You are not describing it, and you are not pitching — you're learning what already happens without you.
2. Ask about the last specific time the problem came up: when it was, what they did, how long it took.
3. Follow the workaround. Whatever they use today — a spreadsheet, a person they call, a service they pay — that's your real competition, and its price is the bar yours has to beat.
4. Ask what they've already tried and dropped. Money someone has already spent trying to fix this is the strongest signal you'll get.
5. Stop talking. Count to three after they finish; the second half of the answer is the useful half.
6. Close with "who else deals with this?" and ask for the introduction before you leave.
7. Write it up within the hour, in their words, not your summary — and log it against that person so the pattern is there on day 7.

The interview script — and how to read the answers

Five questions, in order. Ask, then stop talking.

1. "Walk me through the last time this came up." Good — a date, a sequence, names, a number. Weak — "oh, it happens all the time." That's a summary, not a memory. Ask again: "When was the last one?"

2. "What did you do about it?" Good — a workaround you could point at: a spreadsheet, a person they called, something they pay for. Weak — "nothing, really." That IS your answer. A problem nobody works around isn't painful enough to pay to solve.

3. "What did that cost you?" — time, money, or both. Good — "about four hours" / "$300 a month" / "we lost the job." Weak — "it's frustrating." Push once: "frustrating enough that it cost you what?"

4. "What have you already tried that didn't work?" Good — a named tool, a hire, a vendor they fired. Weak — "haven't tried anything." Note it and move on. Someone who has already spent money failing to fix this is the closest thing to a guaranteed customer you will ever find.

5. "Who else deals with this?" Good — two names, and an offer to introduce you. Weak — "oh, everybody." Ask for one name.

NEVER ASK THESE. Each one is designed, without meaning to be, to produce a false yes:

"Would you buy this?" Everyone says yes. Nobody means it. "Is this a good idea?" You're asking them to be kind to you. They will be. "How much would you pay?" They'll invent a number they have never actually paid. Ask what they pay NOW instead.

READING IT AFTERWARDS. Sort every conversation into one of three piles, and be strict:

EVIDENCE They already spend time or money on this, and they told you the number without you dragging it out of them. INTEREST They agreed it's a problem, but they do nothing about it today. Pleasant. Not proof. NOTHING They liked your idea.

Only the first pile counts. Five conversations that land in EVIDENCE beat fifty that land in INTEREST — and if all ten land in the bottom two piles, that is not a failed week. That's the week working exactly as intended, for the price of ten coffees instead of a year.