To validate a startup idea, find the one assumption your whole idea depends on — usually that a specific person has a painful problem and will pay to solve it — and test it with real potential customers before you build anything. Validation isn't proving you're right; it's finding the cheapest, fastest way to learn whether you're wrong.
Start by writing one sentence: who has this problem, and what does it cost them today? If you can't name a specific person — not "small businesses" but "a solo bookkeeper who loses an evening a week to chasing invoices" — you don't have an idea yet, you have a category.
The honest test: would the person you named recognise this as a real, present pain, or is it a nice-to-have they've lived with for years? Painkillers get bought; vitamins get put off. Most ideas die because the problem wasn't urgent enough to anyone specific.
Your riskiest assumption is the belief your idea depends on most that you're least sure about. List everything that has to be true for this to work — people have the problem, they'll pay, you can reach them, you can build it — then rank each by how badly it hurts if it's wrong and how confident you actually are. The highest impact times uncertainty is what you test first.
It's almost never the technology. It's usually demand: will a specific person actually pay for this, at a price that works, often enough? Founders avoid testing the riskiest assumption precisely because the whole idea rests on it — which is exactly why it has to go first, while changing your mind is still free.
Design the smallest experiment that could kill the assumption. For demand, that's usually 10–20 honest conversations with real potential customers — asking about their actual problem and what they do today, not pitching your idea and watching them be polite. For willingness to pay, it's a pre-order, a paid pilot, or a landing page with a real "buy" button and paid traffic.
Decide the pass/fail line before you run it. "I'll keep going if 4 of 10 say they'd pay and one actually does" is a test. "People seemed interested" is not. The point is to make the result un-foolable — to give the idea a real chance to fail.
Compliments are not validation; behaviour is. Someone saying "great idea" costs them nothing. Someone giving you their email, their time, or their money is data. Weight what people do over what they say, and be suspicious of any result that just happens to confirm what you wanted to hear.
If the test fails, that's a win — you've saved months. Either the idea is dead, or you've learned something specific that tells you how to reshape it. Then you name the next riskiest assumption and test that. Validation is a loop, not a one-time gate.
Try it now
Free idea validator — an honest score in seconds
Talk to 10–20 real potential customers about their actual problem, and put up a simple landing page with a clear offer to measure whether anyone wants it. Both cost nothing but time, and together they test the two things that matter most: is the problem real, and will anyone act on it.
Around 10–20 focused conversations is usually enough to see a clear pattern. You're looking for a repeated, specific pain and signs people would pay — not a statistically perfect sample. If the same problem and willingness-to-pay show up again and again, that's signal.
Asking if people like your idea measures politeness; validation measures behaviour. People will say "great idea" to be nice. Validation tests whether they'll actually give you their time, email, or money — the things that predict whether they'll buy.
Skip the guesswork. Kasspian is your honest AI business co-founder — it validates your idea, builds your business pack, and gets you paying customers.
Try it freeOne week, a dead startup and why it died. The next, a play to get your next customers.
No fluff, no spam. Unsubscribe anytime.