Everybody you ask will encourage you. Friends are kind, family is invested, and "would you use this?" is a question about imagination rather than behaviour. That is why so many validated ideas fail on launch day.
Real validation is behavioural: somebody spends money, or time, or gives up something. Here are the tests that produce that evidence, in the order of cost and reliability.
Quick Info
What counts as validation
Money changing hands, or a firm commitment
What does not
Positive conversations and survey responses
Best single test
Pre-sell before you build
Budget needed
Under £100 for most ideas
Time needed
Two to four weeks
Pass mark
Two strangers pay, without a discount
Why conversations mislead you
People answer hypothetical questions with their aspirations, not their behaviour. "Would you pay £30 a month for this?" gets a yes from somebody who would never actually buy, because saying no to an enthusiastic person is unpleasant.
—Friends and family are the least reliable sample you have access to.
—Survey respondents are not customers; they cost nothing to be.
—Social media engagement measures whether something is interesting, not whether it is worth money.
—"Let me know when it launches" is a polite no.
If your validation did not involve somebody handing over money or a signed commitment, you have collected encouragement rather than evidence.
Test one: ask about the past, not the future
Free, and the best starting point. Interview ten people in your target market about what they currently do, not about your idea.
01What did you do last time you had this problem?
02What did you use, buy, or pay for? How much?
03What was annoying about it?
04How often does this come up?
05What would have to be true for you to try something different?
Past behaviour is factual. Future intention is fiction. If nobody has ever spent money or effort on this problem, that is the finding, and it is worth knowing before you build.
Test two: search demand
Twenty minutes and free. Are people looking for a solution, and in what words?
—Search your problem and see whether anybody is selling a solution. Competitors are validation, not discouragement.
—Check the autocomplete suggestions and People Also Ask boxes for the real phrasings.
—Look at review sites for complaints about existing solutions. That is your differentiation, written by customers.
—Look in the forums and communities where your audience gathers, for the question asked repeatedly.
No search volume and no competitors usually means no market, not an untapped one. Untapped markets are much rarer than the phrase suggests.
Test three: pre-sell it
The real test, and the one people avoid. Offer the thing before it exists and take money for it.
01Write down precisely what you are offering, for whom, and at what price.
02Take it to twenty people who fit that description. Individually, directly, not as a broadcast post.
03Ask for the sale at the real price. Deposit, pre-order, or booking for a date.
04Be honest that it is not built yet. "I'm launching this in March, £200 to reserve a place" is a legitimate offer.
05If two strangers pay without a discount, you have a business. If nobody does, you have saved yourself a year.
Refund anybody if you decide not to proceed. That is what makes this ethical and it costs you nothing but the conversation.
Test four: the landing page test, done properly
A page describing the offer and a real action. Useful, and weaker than pre-selling because the commitment is smaller.
—State the offer and the price. A page with no price tests curiosity rather than demand.
—Ask for something costly: a deposit, a booking, an application, not just an email address.
—Send a small amount of traffic, £50 to £100 of ads or a few relevant community posts.
—Judge on the costly action, not on signups. Email signups measure interest; deposits measure demand.
—Twenty visitors and no action tells you little. Two hundred visitors and no action tells you plenty.
Test five: do it manually for one customer
For services and anything with a process. Deliver the outcome by hand, without any of the systems, for one paying customer.
—You learn what the work actually involves, which is always different from the plan.
—You find out whether you enjoy it, which is a genuine business risk people ignore.
—You get a case study, which is the most valuable marketing asset you will own.
—You discover the real cost of delivery, which decides whether the price works.
Almost every service business I know that works started this way, and almost every one that failed built the systems first.
What a pass looks like
01Two or more strangers have paid, at full price, without a personal favour involved.
02They have the problem you thought they had, described in their words.
03The price covers your cost of delivery with margin.
04There is evidence of repeat need or referral potential.
05You can find more of these people by a method you have actually tested, not one you assume will work.
That last item is the one people skip. Selling to three people you already know is not the same as being able to find customers.
What a fail looks like, and what to do
—Lots of interest, no payment: the problem is real and your solution or price is wrong. Change the offer, keep the audience.
—No interest at all: wrong audience or the problem is not painful enough. Change the audience or the problem.
—People pay but delivery costs more than the price: the model does not work. Reprice or narrow the scope.
—You could only sell to people who know you: you have not tested acquisition. That is the next test, not a pass.
—You hate doing it: a genuine fail. Do not build a business on work you resent.
A failed validation costing £100 and three weeks is the best value in business. A failed launch costs a year.
A two-week validation plan
01Days one to three: ten conversations about past behaviour.
02Day four: search and competitor check.
03Day five: write the offer down precisely, with a price.
04Days six to twelve: take it to twenty people and ask for the sale.
05Days thirteen to fourteen: deliver manually for whoever paid, and write down what you learned.
06Then decide, on the evidence rather than on how attached you have become.
The side business guide covers what to do once two people have paid.
Frequently Asked Questions
How do I validate a business idea?
Ask ten people about what they did last time they had the problem, then offer your solution at a real price to twenty people. Payment is the only reliable signal.
Why are surveys not enough?
They measure what people imagine they would do. Answering a survey costs nothing, so it predicts almost nothing about buying behaviour.
How much should validation cost?
Under £100 for most ideas. If validating requires significant spend, you are building rather than validating.
Is it ethical to pre-sell something that does not exist?
Yes, if you are honest about the timeline and refund anybody if you do not proceed. That is how most products have always been funded.
What if competitors already exist?
That is validation. A market with paying customers is easier to enter than one you have to create, provided you can be specifically better for some segment of it.
How many sales prove an idea works?
Two full-price sales to people who did not know you, plus a tested way of finding more of them.
Before You Go
Ask about the past, check that people are searching, write the offer down with a price, and ask twenty people for money. Two weeks, under £100, and a real answer.
If it passes, deliver it manually for the first customer before you build anything. If it fails, you have bought the cheapest lesson available. The build order is here.
Ask for the money. That is the whole test.
WRITTEN BY TARIQ SALLAM
Marketing Consultant. Entrepreneur. Content Creator.
I'm a marketing consultant, entrepreneur and content creator. I help businesses grow through practical marketing, websites, SEO, content and AI.
More About Tariq →
Keep reading