Most failed businesses were not badly executed. They were built for a problem nobody was paying to solve, and the founder found out eight months and considerable money later.

Validation is finding that out in a week instead. Not by asking whether people like your idea — they will say yes — but by finding out whether they act.

Why asking people does not work

“Would you use an app that does X?” produces enthusiastic yeses and predicts nothing. People are answering a hypothetical, they want to be encouraging, and imagining a purchase costs nothing.

The fix is to ask about the past instead of the future. What someone did last month is a fact; what they might do next month is a guess dressed as an answer.

• Instead of “would you pay for this?” ask “what do you currently do about this problem?”
• Instead of “is this useful?” ask “when did this last cost you time or money?”
• Instead of “how much would you pay?” ask “what are you spending on this now?”

If the honest answer is that they do nothing about it and it costs them nothing, you have found something that is mildly annoying rather than genuinely painful. People do not buy solutions to mild annoyance.

A seven-day process

This costs almost nothing and it is deliberately uncomfortable, because comfort is what makes validation take six months.

Days one and two: define the problem precisely. Write one sentence naming who has the problem and what it costs them. “Small e-commerce owners lose sales because their product photos look amateur” is testable. “People need better marketing” is not. If you cannot name the person, you cannot find them to ask.

Days three and four: talk to ten people who have it. Find them where they already gather — communities, forums, local businesses. Ask about their current approach and what it costs them. Do not describe your solution. The moment you pitch, they stop telling you the truth and start being polite.

Day five: build one page. Who it is for, what changes for them, a clear price, and a button. The button can go to a form, a waitlist, or a payment link. It does not need a product behind it yet, but it must be honest about what happens when they click.

Days six and seven: put it in front of strangers. Share it in two communities where those people are, or spend a small amount on ads. Fifty to a hundred visitors is enough to learn something. Then count who acted.

Reading the result

Someone pays or leaves a deposit. Strong signal, even from one or two people. Money is the only response that cannot be misinterpreted.

People give email addresses but nothing more. Mild interest. It means the problem is recognized, not that it is worth paying to solve. Follow up and ask directly what would need to be true for them to buy.

Traffic arrives and nobody acts. Usually the offer, not the idea. Try describing the outcome differently or changing the price before abandoning it.

You cannot even find people to talk to. This is a result, and an important one. If ten people with this problem are hard to locate, finding a thousand customers will be considerably harder.

How to run the ten conversations

These conversations are where most of the value sits, and most people run them badly by talking too much.

Aim to speak for about a fifth of the time. Your job is to understand their situation, not to explain yours. The moment you describe your solution, the conversation becomes a politeness exercise and the useful information stops.

• “Walk me through the last time this happened.” Specific and recent beats general and hypothetical every time.
• “What did you do about it?” Reveals whether they act or tolerate. Tolerating is a bad sign.
• “What did that cost you?” In time, money, or a lost customer. This becomes your pricing anchor later.
• “Have you looked for something to fix this?” Active searching is a much stronger signal than passive frustration.
• “Who else has this problem?” Gets you the next conversation and tells you how big the group is.

Write down the exact words people use. Their phrasing is what should appear on your landing page later — describing the problem in your own vocabulary rather than theirs is a common and avoidable reason pages fail to convert.

The strongest test: sell before you build

Everything above is preparation for this. The definitive validation is taking money for something that does not exist yet, and it is far more acceptable than people assume as long as you are honest.

Tell people plainly: this is being built, here is what it will do, here is when you will get it, here is a discount for buying now, and here is a full refund if it does not arrive. Nobody objects to that framing. What they object to is being misled about what exists.

This works for services too. Sell the outcome, deliver it manually for the first few customers, and automate only once you know which parts actually need automating. Doing the work by hand first is not a compromise — it is how you discover what the product should be.

The discomfort of asking for money before delivery is precisely why this test works. It is the same discomfort your customers feel when deciding whether to buy, and putting yourself through it early is the cheapest information available.

Competition is a good sign

Finding existing competitors deflates most first-time founders. It should reassure them.

Competitors prove people pay for this. A market with no competitors usually means no market — someone almost certainly tried and it did not work. Being first is far more often a warning than an advantage.

What matters is whether you can be meaningfully different for a specific group. Not better overall, which is vague and hard to prove, but clearly better for one type of customer. Read your competitors’ negative reviews — that is a free list of the gaps their customers are already complaining about.

Validation mistakes

• Asking friends and family. They are answering a question about your feelings, not your business.
• Building first, validating after. Once you have spent three months building, you will interpret every ambiguous signal as encouragement.
• Treating a small yes as proof. Two enthusiastic people is a start, not a market.
• Validating the idea instead of the price. People wanting something and people paying your price for it are different findings.
• Stopping at the first no. One rejection is noise. Ten is a pattern.

When validation fails

A negative result feels like a wasted week. It is the opposite — you spent a week to avoid spending a year, and the information is usually more specific than “no”.

Often the problem is real but you targeted the wrong people, or solved the wrong part of it, or priced it for a buyer who does not have that budget. The conversations from days three and four usually contain the correction if you go back and read them properly.

The businesses that work rarely launch as originally conceived. They are the ones that adjusted early, cheaply, before commitment made adjusting expensive.

Frequently asked questions

How many people do I need to talk to?

Ten is enough to see a pattern. By the seventh or eighth conversation you will be hearing the same things repeatedly, which is the signal you have learned what this group has to tell you.

Can I validate without a website?

Yes. Offer the service manually to ten people and see if anyone pays. Doing it by hand for the first few customers teaches you more than any landing page, and it is faster.

Is a waitlist good validation?

Weak. Joining a waitlist costs nothing, and waitlist-to-customer conversion is usually far lower than people expect. Treat it as interest, not demand.

What if someone steals my idea?

Almost nobody will. Execution is the hard part and most people are busy with their own work. The far greater risk is building something in secret that nobody wanted.

Then check the arithmetic

Once people are willing to pay, work out whether the numbers hold. Our break-even calculator tells you how many customers you need to cover costs, and the margin calculator tells you whether the price leaves anything.

A validated idea that requires 400 customers a month when your market has 300 people in it is still not a business. Validation and viability are two separate tests, and both need passing.

If you are still deciding what kind of business to build, we compared the options in how to start an online business.