Skip to content

How to do it

  1. Check the problem, not the reception

    Someone is already spending money or hours on this, and you can name them. An enthusiastic response to your description is not this. The test is whether the problem existed before you turned up.

  2. Check for independent repetition

    Several people, who do not know each other, describing the same workaround in similar terms. One vivid story is an anecdote. The same spreadsheet three times is a product requirement.

  3. Check that budget exists and has moved before

    Find out who signs off, what they last bought for this area, and how long it took. If nobody has ever bought anything for this, you are creating a budget line — a much longer sale than most first-time founders can survive.

  4. Check that someone gave something up

    A pre-payment, a deposit, a signed pilot with a date, a scheduled hour with their boss. Something that cost them. Without this, the first three are a well-researched hypothesis.

Four conditions, rising in what they cost the other person. The first three can be satisfied by careful research; only the fourth costs them anything, which is why only the fourth predicts whether a business happens.Step 1: Someone already pays, research. Step 2: Workaround repeats, interviews. Step 3: Budget holder known, interviews. Step 4: Someone gave up money, commitmentSomeone already paysresearchWorkaround repeatsinterviewsBudget holder knowninterviewsSomeone gave up moneycommitmentcosts them nothingcosts them something
Four conditions, rising in what they cost the other person. The first three can be satisfied by careful research; only the fourth costs them anything, which is why only the fourth predicts whether a business happens.

Worked example

Scoring the clinic-billing idea against the four conditions.

People already paying to solve it4 clinics pay a part-time billeryes
Same workaround heard independentlythe spreadsheet, 3 timesyes
Budget holder identified, has bought beforeclinic owner, bought practice softwareyes
Someone gave something upno pilot, no deposit, no scheduled demono
Statuspromising, not validated3 of 4

Three of four is where most well-run validation exercises stop, and it is exactly the point at which founders start building. The fourth condition is the only one that costs the other person anything, which is why it is the only one that predicts anything — and it is usually one honest ask away.

Why the fourth condition is the only one that predicts anything

The first three conditions can all be satisfied by careful research. You can establish that people pay for alternatives, hear the same workaround repeatedly, and identify the budget holder without anyone ever agreeing to anything. That is a well-researched hypothesis, and it is where most good validation exercises stop.

The fourth is different in kind because it costs the other person something. A deposit, a signed pilot, an hour of their calendar with their boss in the room — these are the only signals that survive the person being polite, and the only ones that have ever predicted whether a business happens.

Validation does not expire, but it does decay

Evidence collected eighteen months ago described a market that has since had eighteen months of competitors, price changes and new tooling. If your validation predates your current plan by more than about six months, re-run the cheapest test — usually two conversations — before you treat it as current.

The exception is structural evidence. The fact that clinics pay a part-time biller does not change quarterly. The fact that a specific competitor left a specific gap absolutely does.

What goes wrong

  • Treating a large or enthusiastic waitlist as the fourth condition
  • Counting the same person twice because they said yes in two different ways
  • Declaring validation after the interviews and before any commitment
  • Moving the goalposts once three conditions are met and the fourth is inconvenient
  • Validating with people who would never be in your reachable market

Is your number plausible?

Each check catches a different way the arithmetic can be right and the answer still wrong.

  • Write down what would have made you stop, then check whether that thing happened and you continued
  • For each of the four, name the specific person or number behind it. A condition you cannot attribute is not met
  • If the evidence all arrived after you started building, it is justification rather than validation

Related answers

Is a signed letter of intent enough?

It satisfies the fourth condition if it names a start date and a price. An LOI with neither is a polite expression of interest, and founders regularly mistake one for the other.

How many people need to give something up?

One is the threshold for the condition and three is where it starts being a pattern. The useful question is whether the three are independent — three people from one company is one signal, not three.

What if I am validated on three conditions and stuck on the fourth?

Usually it is one honest ask away. Founders reach three of four and start building precisely because the fourth requires asking someone for a commitment, which is uncomfortable. That discomfort is the test.

Validate your idea freeScore the idea across market, competition, differentiation, monetisation and execution.