How to validate a startup idea before you start building?

People think validation is asking what they want. It's actually watching what they do.

Last time, I wrote about how products don't die because they're badly built, but because nobody asked whether the idea was worth building in the first place. I got a few messages asking the most obvious follow-up: okay, but how do you actually check? How do you validate an idea before you incur a six-month sunk cost?

The instinct is to share it with friends and family, post it to a founders' group, send out a survey, and run it past ChatGPT one last time just to be sure. The responses come back warm: "Great idea." "Go ahead with it." You feel validated, so you start building.

Here's the real answer: none of that was validation. People tell you your idea is good the same way they'd tell you a bad meal was lovely after you'd cooked it for them. That's politeness. It's not a sign that you've suddenly become the best cook in town.

The first step is to talk to people about their problem, not about your idea. The moment you describe your idea, the conversation is poisoned, because now they're being nice to you.

So don't describe it. If you're building an invoicing tool for freelancers, don't ask, "Would you use an invoicing tool?" Ask, "Walk me through the last time you had to chase a client for payment. What did you do? How long did it take?"

"Would you?" is a fantasy question. "What did you do last week?" is a fact. And if the honest answer is, "Eh, it's a minor inconvenience, I don't really think about it," congratulations — you just saved yourself six months building something nobody was going to use.

Another strong signal that a problem is real: people are already solving it badly. So this is a good time to research whether solutions like yours already exist. Look at what's out there, and read the reviews, especially the angry ones. Every one-star complaint is a user telling you exactly what they wish existed. Existing bad solutions aren't a reason to quit; they're proof the problem is worth solving. What should worry you is finding nothing at all, because that often means the pain isn't sharp enough for anyone to bother.

Sell it before it exists. Create a landing page with one clear promise and a single call to action, and price it low. If a stranger clicks "buy," they want it. One click from someone with no emotional attachment to you is worth a hundred "great ideas" from friends who have plenty.

Before you build the machine, be the machine. If your idea is an app that matches people with meal plans, don't build the app—just match a handful of real people with meal plans yourself, over email, manually, doing all the work behind the curtain the software is eventually supposed to do. It's slow, and it doesn't scale, and that is entirely the point. You're not trying to scale yet. You're trying to learn whether anyone actually wants the outcome badly enough to keep coming back. If they don't want it when you're personally handing it to them, no amount of polished software will fix that.

What validation actually gives you:

Notice that none of this requires a finished product. That's the whole idea. Every one of these tests is designed to get you the truth for the price of a conversation, a landing page, or a weekend of manual work, instead of the price of a full build and half your runway.

And validation isn't only there to kill bad ideas. More often it reshapes good ones. You go in thinking you're building one thing, you talk to enough people, and you realize the real pain is one step to the left of where you were looking. That's not failure. That's the whole process working. The idea you walk out with is almost always better than the one you walked in with, because this time it's built on what people actually do, not on what a very confident machine told you over breakfast.

You don't need permission to start validating. You need a landing page, a list of people to talk to, and the willingness to hear an answer you might not like. The founders who make it aren't the ones with the best ideas. They're the ones who found out the truth first.