What Counts as Evidence of Startup Demand?

Startup demand is not a single fact. It is a chain of evidence connecting a real problem to behaviour, a buying decision and continued value.

Search volume, interviews, waitlists and payments can all be useful. They support different conclusions.

Start with the claim

Before collecting evidence, write the exact statement you are trying to support.

Examples:

• This problem occurs weekly for operations managers.
• Teams already spend money or staff time addressing it.
• A budget owner will fund a different approach.
• Customers can adopt the product without unacceptable risk.
• The value repeats often enough to support the revenue model.

A general claim such as “there is demand” is too vague to test.

Level 1: attention

Examples include search volume, social engagement, newsletter signups and traffic.

Attention can show that a topic is interesting or a message is effective. It does not reveal whether the visitor has the problem, belongs to the target segment or intends to buy.

Useful conclusion: people notice or seek information about the topic.

Level 2: stated problem

Interviews, surveys and support discussions can reveal recurring pain and language.

Improve the evidence by asking about the last occurrence:

• What happened?
• What did the person do?
• Who else became involved?
• What did delay or failure cost?
• Which alternative did they use?

Useful conclusion: members of the segment report a specific problem.

Level 3: problem behaviour

Look for observable effort:

• spreadsheets or internal tools;
• manual reconciliation;
• consultants or outsourced work;
• repeated searching;
• escalation and rework;
• staff time;
• use of several partial products.

Behaviour shows that the problem is consequential enough to produce action.

Useful conclusion: the problem causes repeated effort or cost.

Level 4: buying evidence

A buyer identifies a budget, approval route, trigger or proposal requirement. They may introduce stakeholders, provide data, accept a security review or request terms.

Useful conclusion: a credible purchase process exists.

Use the B2B Buying Committee Mapper to find gaps between the user and the decision.

Level 5: payment

A deposit, paid discovery project, paid pilot or purchase is a strong signal because the customer gives up money and accepts a consequence.

Payment does not prove retention or scalable acquisition. A founder-led service can also hide delivery costs.

Useful conclusion: at least one buyer accepted this offer under these conditions.

Use the Willingness-to-Pay Evidence Ladder to identify the next stronger commitment.

Level 6: repeat value

Renewal, repeated use, expansion, referrals and continued payment show that value persists after the initial decision.

Useful conclusion: the customer continued to receive enough value to repeat or expand the relationship.

This is different from initial demand and should be measured separately.

Evidence quality still matters

At every level, ask:

• Is the evidence recent?
• Is the source independent?
• Does it represent the target segment?
• Is it direct or inferred?
• Is the sample large enough for the decision?
• Is there contradictory evidence?
• What selection bias is present?

One paid pilot may outweigh one hundred compliments for commercial commitment, but it still represents one context.

Weak proxies to label carefully

• An AI score measures the scoring framework.
• A simulated customer response measures model output.
• A founder creating a landing page measures founder action.
• A waitlist signup measures a low-cost expression of interest.
• A page view measures exposure.
• A market-size estimate measures assumptions about a population.

These are not useless. Problems arise when the proxy is relabelled as the outcome.

Build an evidence portfolio

Do not search for one decisive number. Combine:

• external market patterns;
• customer accounts;
• observed workflows;
• current spending and alternatives;
• buying-committee evidence;
• payment tests;
• delivery economics;
• adoption and retention.

Then state what remains unknown.

The goal of validation is not to prove the idea. It is to make the next investment proportionate to the evidence.

Read the SignalCraft methodology, score an important source with the Evidence Quality Scorecard, or examine why founder action is not customer demand.