Collect behaviour, not encouragement

How to validate a business idea before building the expensive version

People are generous with ‘I would buy that’. A useful validation test asks them to spend money, time, reputation or effort—and makes it easy for you to hear no.

Quick answer

Write the idea as a falsifiable claim about one customer, problem, offer, price and channel. Check demand, market size, saturation and alternatives using reliable sources; interview people who recently faced the problem; then present a real offer and ask for a meaningful commitment such as a paid pilot, deposit, qualified introduction or scheduled onboarding. Model delivery costs and decide the pass, change and stop thresholds before seeing the result.

Reviewed for material changes on August 13, 2026

Turn the idea into something that can be wrong

Write one sentence: We believe this specific customer will choose this offer at this price through this channel because it solves this costly problem better than their current alternative. Replace any word that could describe half the population.

Choose the riskiest assumption. It may be that the problem is frequent, the buyer controls a budget, the product can be delivered profitably or the audience can be reached. Test that assumption before polishing the parts everybody already agrees with.

Use desk research to size the question

Look for evidence of demand, market size, geography, economic constraints, saturation and current pricing. The SBA recommends combining existing sources with direct customer research: public data answers broad quantifiable questions, while direct research gives detail about your particular audience.

Map direct and indirect alternatives, including spreadsheets, agencies, doing nothing and internal labour. Read complaints and switching stories, but do not turn a loud online minority into a market estimate.

Interview past behaviour, not your pitch

Find people who encountered the problem recently. Ask what triggered it, what they tried, who decided, how long it took, what it cost and what happened. Specific past events are more useful than a hypothetical opinion about your future product.

Do not spend the interview explaining the idea. Keep a consistent question set, write down exact phrases and look for repeated behaviour across people—not compliments to the interviewer.

  • When did this last happen?
  • What did you do first?
  • Which alternatives did you consider or pay for?
  • Who else was involved in the decision?
  • What made the problem urgent—or safe to ignore?

Ask for the smallest commitment that resembles the business

For a service, sell a narrow paid pilot. For a physical product, test a small batch or clearly described preorder only when you can meet the applicable sales, delivery and refund rules. For software, a concierge version can deliver the outcome manually before the full system exists. A landing page can qualify interest, but an email alone is weaker evidence than money or scheduled work.

State honestly what exists, what the customer receives and when. FTC guidance says US advertising claims and material disclosures must be truthful and clear online; other markets have their own rules. Do not present a concept render as a shipping product or hide conditions after the call to action.

Check whether demand can become a business

Estimate the selling price, direct fulfilment, payment fees, support, refunds and the time required to win a customer. Separate one-time setup from monthly costs and calculate how many sales cover them. A popular offer with no workable margin has validated a problem, not a company.

Note which costs are observed and which are guesses. Get supplier quotes or run one manual delivery where possible. The purpose is not a perfect five-year model; it is to expose an assumption that makes every sale lose money.

Make the decision you wrote before the test

Set thresholds in advance: for example, a number of qualified conversations, paid pilots at a stated price, an acquisition-cost ceiling or a delivery-time limit. When the test ends, choose go, change or stop and record the evidence behind it.

A failed test is useful if it isolates why: wrong audience, mild problem, weak channel, poor offer or broken economics. Change one important assumption and run a new bounded test. Do not keep changing the scoreboard until the result looks encouraging.

Sources and checks

Product limits and prices came from the companies themselves:

Plans move. We date every check so you know when to verify again.

Straight answers

Questions people actually ask

How do I validate a business idea for free?

Use public market data, competitor research and interviews first. You can also offer a manual pilot through direct outreach. Free validation can reveal a great deal, but some willingness-to-pay questions require asking for money.

How many customer interviews are enough?

There is no magic count. Speak to enough relevant people to see repeated behaviour and deliberately look for disconfirming cases. Ten tightly matched recent buyers can teach more than a large convenience survey.

Is a waitlist proof that an idea is validated?

It proves that some people accepted the cost of sharing an email under the conditions of that page. It does not prove they will pay, remain, refer others or cost less to acquire than the business earns.

Should I build an MVP before validating the idea?

Build only what is needed to test the riskiest assumption. A prototype may be necessary to test use; a manual service, mock-up or paid pilot may answer demand sooner and more cheaply.

What counts as strong validation?

Repeated behaviour from the intended customer under realistic conditions: paid pilots, completed preorders with clear terms, renewals, referrals or another costly commitment. The strongest evidence closely resembles how the finished business will earn and deliver.

Skip the blank page

Already know which website you like?

Paste the link. Copycat turns it into an editable Next.js first draft in under 15 minutes.

Copy a website for $100