Is the problem real?

You have an idea. Before you build it, you want to know if it solves a problem people actually have. So you ask them: “Would you use an app that does this?”

They say yes. Nearly everyone does. It costs nothing to be encouraging, and most people would rather not disappoint you. Rob Fitzpatrick’s book on customer interviews is called The Mom Test for exactly this reason, and its subtitle states the problem: how to learn if your business is a good idea “when everyone is lying to you”. They aren’t lying to be cruel. They’re being kind.

The fix isn’t finding more honest people. It’s asking questions that a kind answer can’t distort.

Opinions versus evidence

A question about your idea, their opinion, or what they would do one day invites a guess about the future, and a guess costs the person nothing. A question about what already happened has a true answer, and the true answer is different for every person.

1. 1Every interview is a series of choices about what to ask.
2. 2"Would you use it?" and "How much would you pay?" ask for a prediction. People predict generously.
3. 3"When did it last happen? What do you do about it today? What have you spent?" ask for memories. Memories are specific.

The strongest evidence of all is money or time already spent. Someone paying for a clumsy workaround, or doing the same tedious check every morning, has the problem badly enough to act on it. That’s the kind of person who might pay you to fix it.

Try it

The idea is an app that tells small online shops when their site goes down. Start with the script you’d probably write first, then fix it.

Good problems are ones people already fight

Paul Graham’s essay How to Get Startup Ideas argues that the best ideas usually come from problems the founders have themselves, rather than ones they made up because they sounded like startups. That’s the same test from the other side. If a problem is real, someone is already living with it, and you can ask them how.

Some things to listen for:

• A workaround: a spreadsheet, a manual check, a tool used for something it wasn’t built for.
• Money already spent: even on something that works badly.
• Detail: “Black Friday, six hours, customers emailed me” is a real memory. “Yeah, that happens sometimes” usually isn’t.

And some things to discount: compliments, “I’d definitely use that”, and promises about next month.