Scratch your own itch is dangerous advice.

"Build something you need" is the most repeated piece of startup advice. I've watched it sink more good founders than bad ideas ever have. It's half right, and that's what makes it dangerous: more dangerous than simply wrong advice, because the danger hides in the gap between personal need and market proof.

Scratching your own itch gives you two real things: proof the problem exists for at least one person, you, and the drive to keep going long after the novelty wears off. That's worth a lot. It's also everything it gives you. "I have this problem" is evidence of exactly one customer — and that customer is the least objective judge alive of whether anyone else shares it.

The two things your itch might be

Here's the trap. Your itch is one of two things, and from the inside, they feel identical.

It might be idiosyncratic — peculiar to you. Your situation, your tools, your way of working throw up a problem most people just don't have. Build for it, and you build something only you want, and you'll be genuinely puzzled why nobody else shows up. The problem was always real. It was just real for a market of one.

Or it might be early — you feel the problem before the rest of the market does. Being early is the promising case, and the cruel one, because being early and being wrong are impossible to tell apart at first. You can be dead right about the problem and still fail, because the market isn't ready, and timing is the one thing a founder can't control.

The one question that separates them

Idiosyncratic and early look the same from where you stand: you've got the problem, and almost nobody else seems to. "Scratch your own itch" can't tell them apart. Worse, it talks you out of even asking, because it tells you your own need is proof enough.

It isn't. There's one question that separates them, and you have to ask it outside your own head: is there independent evidence that other people have this problem and are already paying to solve it — in money, in workarounds, in duct-taped tools, in time?

Jessica Wu, who built Sola, lived her problem: the manual operations grind of a quant researcher. Classic scratch-your-own-itch. But that's not why the venture worked. It worked because the same problem surfaced independently elsewhere — in an industry unrelated to hedge funds — through her co-founder's experience with hospital systems. The itch told her where to look. Something outside the itch told her the problem was real. She also did the thing most founders won't: she turned down aligned-looking early revenue that didn't fit, because the signal she trusted was whether people would pay and keep paying, not whether she personally wanted the thing to exist.

Use it to find a problem, never to confirm one

So use your own itch to find the problem. It's great for that. Then go and validate the problem outside yourself before you commit, because your own need proves the problem exists. It says nothing about whether it exists at scale or whether the moment is right.

A great way to find a problem. A terrible way to confirm one. The confirmation has to come from somewhere your own conviction can't reach.