How to validate a startup idea

What validation is, and what it is not

Most ideas do not fail because the product was badly built. They fail because nobody needed it enough to pay. Validation is how you find that out while the idea is still cheap to change.

These feel like validation, but are not:

• Friends and family saying they would use it.
• A large market size from an industry report.
• An AI assistant telling you the idea is great.
• Waitlist sign-ups that cost nobody anything.

Real validation answers three questions with evidence you can point to: who has this problem badly, what they pay to deal with it today, and whether enough of them can be reached by you.

Step 1: Write the idea as a bet

A vague idea cannot be tested. Rewrite yours as one sentence that could turn out to be wrong:

Every blank is something you will check in the next steps. If you cannot fill a blank, that is the first thing to find out.

Step 2: Find who has the problem worst

"Small businesses" or "busy parents" are not a customer. Split the market into groups and pick the one where the problem is frequent, costly, and owned by someone who can pay.

Signs a group feels the problem badly:

• They complain about it in forums, reviews, and communities.
• They post jobs or hire freelancers to deal with it.
• They have built workarounds: spreadsheets, scripts, extra staff.

A problem people feel every day beats a bigger problem they feel once a year.

Step 3: Count the buyers you can reach

Forget the total market size. Count real organizations or people in your chosen group, then narrow it twice: how many actually have the problem, and how many you can reach in the next six months.

A short, named list you can contact is worth more than a billion dollar market you cannot. If the reachable number is tiny, either the group is wrong or the price has to be high.

Step 4: Find what they pay today

Your real competitor is whatever people do today: another tool, an agency, a freelancer, or their own time. Find out what that costs them, in money or hours.

If you find that nobody pays anything today, treat it as a warning, not an opening. "No competitors" usually means no market.

Step 5: Talk to buyers about the past

Interviews only work if you ask about what people did, not what they might do. People are polite about the future and honest about the past.

Do not pitch while you listen. Note who feels the pain and who approves the budget; they are often different people.

Step 6: Ask for something real

A yes that costs nothing is worth nothing. Before you build, ask the most interested buyers for a commitment: a pre-order, a deposit, a paid pilot, or a signed letter of intent.

If people who said "great idea" will not commit to anything, you have learned something important, cheaply.

Common mistakes

• Asking people who like you. Their job is to be kind. Yours is to find the truth.
• Trusting a tool that agrees with you. Chat assistants tend to side with the user. Ask for sources, not encouragement.
• Building first. A landing page, a list of buyers, and five honest conversations come before code.
• Picking the biggest market instead of the most painful one.
• Counting interest as demand. Clicks and sign-ups are interest. Money and signatures are demand.

A worked example

Take the idea: an app that pays adults to film their household chores and sells the videos to robot companies. Here is what a validation run found.

The run also ranked cleaning-robot teams second: they feel the problem only now and then, while home-robotics teams feel it daily. The version to build first was a narrow data pilot for home robotics teams, starting with a $2,500 pilot instead of an hourly rate.

Notice what changed: not whether to build, but who to build for and what to charge. That is what good validation gives you.

Source: figures above are from a Clarno validation run on this example idea (home-robot training data), not from a third-party market report.

Questions founders ask

What is startup idea validation?

Startup idea validation is finding evidence that a specific group already has the problem, already pays to deal with it, and would pay you instead—before you spend months building. It is evidence, not opinions or compliments from friends or chatbots.

How is this different from asking an AI chatbot if my idea is good?

Chatbots tend to agree with you. Validation asks who has the problem, what they pay today, and whether enough of them can be reached. Clarno pressure-tests ideas with simulated buyers built from market evidence, then returns a verdict with sources—not flattery.

How long should validation take?

Days to a few weeks, not months. The goal is a decision on the next step, not certainty.

Can I validate an idea without building anything?

Yes, and you should. Everything in this guide works before a single line of code.

How many people should I talk to?

There is no magic number. Stop when the answers within one group start repeating; a common rule of thumb is ten to fifteen conversations.

What if my idea is already a live product?

The same steps apply. Check who pays you today, why, and which group feels the problem most, then focus there.