Validation Before Building

Surveys ask people what they would do. What people say and what they do are almost never the same thing. The founders who validate well do not ask. They build the smallest thing and watch what happens. A landing page with a real price and a real button is a better validation tool than any survey you will ever send.

The reason is simple. A survey asks someone to imagine a future where they have your problem and your solution. A real price and a real button ask them to act in the present. The gap between saying yes and paying money is the gap between interest and need. Surveys collapse that gap. Real tests do not.

A 2023 analysis of over two thousand startup post-mortems by the startup analytics platform CB Insights found that forty-two percent of failed startups cited no market need as the primary reason for failure. That number has remained consistent across more than a decade of data. Founders consistently overestimate demand and underestimate how much evidence they actually need before committing.

Before you build, offer to sell the finished thing. One person. Real money. If they buy before it exists, you have validated both the problem and the trust you have built describing the solution. A pre-sale is not a commitment. It is a conversation about a commitment.

The founders who pre-sell well describe the outcome, not the product. They say, here is what you will be able to do, and here is what it will cost. Not here is what I am building. The first version asks the buyer to imagine. The second version asks them to decide.

The hardest part of the pre-sale conversation is not the selling. It is the vulnerability of describing something that does not exist yet with enough specificity that someone can decide whether to trust you. That vulnerability is the test. If you cannot have that conversation honestly, you are not ready to sell the thing. If you can, you are closer to ready than you think.

Talk to five people who would pay. Hear their exact words. Notice the problem they describe differently from the one you assumed. Five conversations will teach you more than five hundred survey responses because conversations give you the language people actually use, not the language they think you want to hear.

Write down the words they use to describe the problem. Those words are the language of your first marketing, your first pitch, and your first hire. They are also the language you will use to write the landing page that converts strangers into customers. The founders who skip this step end up writing copy that sounds like a pitch instead of a mirror.

Rob Fitzpatrick, in his book The Mom Test, which has become required reading in startup accelerators around the world, makes the case that most founders ask questions that lead to the answers they want. A good validation conversation asks about the past, not the future. What did you try last time this problem happened? How much did it cost you? What did you do instead? Past behaviour is the most reliable predictor of future behaviour you will ever find.

Interest is polite. Validation is painful. Interest says that is interesting. Validation says I have this problem and I have tried three things to solve it and none of them worked. The first response is a compliment. The second response is a market.

The founders who confuse the two are the ones who spend months building something based on polite responses and then wonder why no one buys it. The polite response does not mean the problem is not real. It means the person you are talking to does not have enough skin in the game to tell you the truth.

Look for the signals that indicate real validation. They include: the person describing the problem without being prompted, the person mentioning how much the problem costs them, the person asking you practical questions about your solution, the person offering to introduce you to someone else who has the problem. These are not guarantees. But they are signals that move past politeness into genuine need.

You do not need a sophisticated validation process. You need a process that produces evidence before you spend money building. Here is the minimum version: identify ten people who have the problem you think exists. Schedule fifteen-minute conversations with each of them. Ask about the problem, not your solution. Write down the exact words they use. After five conversations, look for patterns in the language. After ten, make a decision about whether to proceed.

That process takes roughly two weeks. It costs nothing except time. It produces more reliable evidence than any market research you will ever buy. And it gives you the language you need for the next step, whether that step is building, pitching, or deciding the problem is not worth solving.

The founders who skip this process do so because they are afraid of what they might find. That fear is the reason the process is necessary. The founders who run it honestly are the ones who make better decisions, faster, with less money spent on things nobody wants.

Related reading: \2 · \2

Y Combinator — The Real Product Market Fit. YC describes product-market fit as a state where customer demand and usage overwhelm the company's ability to keep up. Y Combinator — Startup School Week 1 Recap. YC emphasizes direct user learning at idea, prototype and post-launch stages.

Evidence boundary: These sources support the underlying principles and context; they do not prove that one universal tactic works for every founder, market or company. SOE separates source evidence from editorial synthesis.