Signal vs Pull

Imagine this is your project.

12 out of 15 people you interviewed said they would use the product.

What do you know now?

Most misleading signals are real.

The interview really did go well. The user really did return. The customer really did say they would buy. The mistake happens one step later. Watch the movement:

Every line sounds plausible. Each arrow adds an assumption — open them. The first observation can be true while the build decision is still unsupported.

Evidence does not get stronger because the conclusion resting on it matters more.

When founders talk about validation, several different things collapse into one word. Someone understood the problem. Someone signed up. Someone agreed to a pilot. Someone paid.

Each of these tells you something different. The useful question is “what does this signal actually prove?” — and then “is that enough evidence for the decision I am about to make?”

That changes how you test a product: you collect evidence for a specific decision.

A signal has no fixed value. It depends on how this customer buys.

Two potential customers both agree to another 30-minute call. Which one carries more information?

The same is true of money. A €20 payment can be meaningful in one test and almost meaningless in another. A signed pilot can be serious movement in one enterprise sale and a non-binding experiment in another. Sharing operational data may say more than “yes, we have budget.” Bringing procurement or legal into the room may tell you more than another enthusiastic interview.

So instead of memorising a universal hierarchy of signals, ask:

What is scarce in this buying system?

The more a customer moves something that is genuinely scarce in their context, the more information their behaviour contains. A meaningful action is evidence for this customer. It is not yet proof of the market.

Three questions, every time you look at customer evidence.

The strength of evidence depends on the decision it is being asked to support.

“Yes” is one signal. Following through is another.

People commit in conversation all the time, and often mean it. Then calendars fill up, another stakeholder says no, the data turns out hard to share, budget needs approval. So there is a distinction between stated commitment — what the person says they will do — and fulfilled commitment — what they do when it costs them something.

Four things a customer says. Before you open each one, ask yourself what “fulfilled” would look like.

A stated commitment tells you about willingness in the room. A fulfilled one tells you:

The willingness held when it cost something.

A strong-looking signal can still come from a weak test.

Eight people say the problem is urgent. Then you listen to the interviews. The founder described the problem first, asked “would it be useful if…”, explained the benefits, showed a polished demo, and asked whether the person could imagine using it. The answers may be sincere. The customer was still never given a chance to show their own reality.

If several answers are unclear, fix the test before collecting more signals.

Strong evidence is a pattern.

A founder gets the first payment. Someone has finally put money behind the product. Now another mistake becomes possible: asking one payment to prove the whole business.

The payment counts. Claim only what it shows. The same applies to one pilot, one champion, one enterprise logo, one viral post, one returning user, one large waitlist, one successful interview.

One payment can prove that one customer paid under those conditions. Does the pattern corroborate the conclusion?

Eventually you have to decide. There is no number of interviews that turns uncertainty into truth — ten can be plenty for one question and meaningless for another.

Do we have enough evidence for the decision we are about to make?

Call it decision sufficiency. How much evidence you need depends on how big the decision is and how hard it is to undo.

Which decision is in front of you right now?

Ask “has this bet earned the next investment?” rather than “have we validated the startup?” You are deciding whether the evidence justifies the next step.

Which of these is closest to your situation?

Now use the frame on your project.

Answers stay in your browser until you send them to yourself. The first and last lines read together are the point.

Different products have different buying systems.

No startup moves neatly from likes to interviews to signups to pilots to payments. Use this frame instead.

Select the strongest signal you have right now.

The table stops one signal from carrying a conclusion it has not earned.

Interest is a reaction. Pull is movement: the customer gives up something scarce — data, a stakeholder, a procurement process, money — and when it comes to cost them, the movement holds.

Founders often ask “do I have validation?” It compresses too many different things into one yes-or-no answer. A better question:

What has the market actually given me permission to do next?

Keep exploring, change the customer, run a cleaner test, ask for a stronger commitment, build — or stop investing in this version.

You need evidence proportionate to the next decision. So before you build the next feature or spend the next budget, ask:

If the last answer is “no”, the next step is a test. If the honest answer is “I am not sure which decision I am making” — that is a different problem, and the one worth talking through.

Not sure which decision you are making?

Sometimes the gap is not in the signal but in the decision: which customer, which commitment, which question. That is what we work on together — 60 minutes with a Beyonds mentor over your audit, through three things:

If the audit already names the missing signal, you may not need the session — run the test.

{{ srsAuditNote }}