Validation

5 Checks Before you Build Anything

29 August 2026 · Updated 9 September 2026 · Validation

Most people test whether their idea is good. That’s the wrong test.

The five things worth checking are whether you can explain it in one sentence, how many real customer conversations you’ve had, whether the problem is urgent, frequent or costly, whether you’ve accidentally built for one market only, and whether you have a backup plan. The first carries more weight than the other four combined.

Why the usual test is the wrong test

Building stopped being the constraint somewhere around 2023. Finding someone who will pay is still the constraint, and it always was. Building used to be hard enough to hide that.

Two numbers make the case.

Stripe reported in February 2026 that median revenue in the first six months for new businesses on its platform fell 23% year on year in 2025, while revenue at the top decile rose 19%. In the same year, solo founding hit an all-time high. More people building than ever, and the typical one earning less than the year before. That gap isn’t build quality.

CB Insights’ analysis of startup post-mortems found the most common root cause of failure was building something the market didn’t need. That was 42% in the original study of more than 110 post-mortems, and 43%, described as poor product-market fit, in the 2024 update covering 431 VC-backed shutdowns.

Note what neither of those says. Not bad code. Not the wrong tech stack. Not a bad idea.

Check 1: Can you say it in one sentence?

Write down what you’re building. One sentence, aimed at someone who doesn’t work in your industry.

Most people can’t, and it isn’t a knowledge problem. It’s the opposite. Fifteen or twenty years in, everything connects to everything, and you’ve lost the ability to see what a stranger doesn’t already know.

The shape that works:

I help [a specific person] who [has a specific problem] to [get a specific outcome].

Three tests on what you wrote:

Findable. Is the person specific? “Small businesses” is not a person. “Bookkeepers at firms with four to twelve staff” is a person, because you could find one this afternoon.
Theirs. Is the problem theirs, or is it one that annoys you? A lot of first ideas are the founder’s irritation, not the customer’s cost.
Sayable. Could you say it out loud without wincing? If you’d rather email it than say it, it isn’t ready.

Score yourself out of ten, honestly. Most people are a three and think they’re an eight.

Check 2: How many real conversations have you had?

Not surveys. Not your mate saying it sounds great. Conversations, where you sit down with someone who has the problem and ask about their life rather than pitching your solution.

Rob Fitzpatrick’s The Mom Test is the method worth learning here. Ask about what they actually did, not what they think they’d do. People are unreliable about the future and quite reliable about the past.

Four questions that work:

What do you do about this today?
What happened the last time it came up?
What did it cost you, in time or money?
What have you already tried, and why did you stop?

The best signal that a conversation is real: you’re slightly afraid of the answer. If nothing you asked scared you, you were collecting encouragement.

How many conversations is enough?

There is no research-backed magic number. Anyone who tells you it’s exactly 28 has made that up.

I use 25 as a minimum, and the reason is practical rather than scientific. It’s enough to start seeing a pattern, and it’s too many to fool yourself with. Five conversations will tell you what you want to hear. Twenty-five will tell you what’s actually there.

It also isn’t a finish line. You keep having them.

Check 3: Is the problem urgent, frequent, or costly?

You want two of the three.

Costly means real money or real time being lost.
Frequent means it happens weekly or daily, not annually.
Urgent means there’s a deadline attached, often a compliance or legislative one.

A problem that’s cheap, occasional and mildly annoying is a very hard sell no matter how elegantly you fix it. People have been living with it for years and they’ll keep living with it.

The honest question underneath: what are they doing instead right now? If the answer is “nothing, they just put up with it,” that’s much harder than if they’re paying someone badly to handle it.

Check 4: Which market did you actually build for?

Most people build for one market by default rather than by decision. Currency, pricing model, compliance assumptions, and quiet assumptions about how things work all go in without anyone choosing them.

That’s fine until a second market becomes interesting, and then it isn’t one decision to change, it’s about forty.

This is not an argument for launching everywhere. Validate in one market. It’s an argument that a handful of decisions made now keep the core portable, and those decisions are cheap today and expensive as a retrofit in year three.

Check 5: Do you have a backup plan?

This one isn’t about the idea at all.

If you’re still employed, the best thing you can do is stay employed while you find out whether anyone will pay. The things that actually tell you are conversations and a small paid offer, and both fit around a job.

The all-in version is a story people tell afterwards. It’s rarely how it happened, and at 50 with a mortgage and dependents, “all-in or nothing” reliably resolves to nothing. That’s not a courage problem, it’s sensible maths.

The five, recapped

One sentence a stranger understands
Real conversations, aiming at 25 minimum
A problem that’s urgent, frequent or costly
A portable core rather than a home-only build
A way to keep going that doesn’t require quitting

Find out which one is holding you back

The Day Zero Readiness Scorecard does this properly rather than roughly. 10 questions, about four minutes, and it weights them, so you find out which of the five is actually the problem instead of getting a vague overall sense.

There’s no signup wall. Your result is on the page whether you give an email or not.

Take the scorecard →

Whatever your score, go and have one of those conversations this week. You’ll learn more in twenty minutes than in another month of thinking about it.

FAQ

How do I know if my business idea is good?

You mostly can’t, from the inside. The more useful test is whether the problem is real: is it urgent, frequent or costly for the person who has it, and are they already spending time or money working around it? That question is answered by talking to people who have the problem, not by refining the idea.

How many customer interviews should I do before building?

There’s no research-backed number. Practitioner convention lands somewhere between 20 and 30. I use 25 as a minimum, because it’s enough to see a pattern repeat and too many to fool yourself with. Five conversations will tell you what you want to hear.

What is the Mom Test?

A customer interview method from Rob Fitzpatrick’s book of the same name. The core idea is to ask about the person’s life rather than your idea: what they did last time the problem occurred, what it cost them, what they’ve already tried. Past specifics, not future hypotheticals, because people are unreliable about what they’d do and reliable about what they did.

I built an app with AI and nobody is using it. What now?

That’s the normal outcome, not a personal failing. Building got dramatically easier and finding customers didn’t change at all. Start by writing what it does in one sentence for someone outside your industry, then have 25 conversations with people who have the problem, asking about their situation rather than demoing what you made.

Am I too old to start a business at 45 or 50?

The research says the opposite. Azoulay, Jones, Kim and Miranda, in American Economic Review: Insights (2020), studied 2.7 million US founders and found the mean age at founding for the 1-in-1,000 fastest growing new ventures was 45.0, and a 50-year-old founder was 1.8 times more likely to achieve upper-tail growth than a 30-year-old. The same paper found founders with at least three years in the same narrow industry were substantially more likely to succeed.

Should I quit my job to work on my idea?

Not to validate it. Conversations and a small paid pilot both fit around full-time work, and they’re the things that actually tell you whether it’s real. Quitting converts a solvable question into a financial deadline.