Prove the demand.
Then build your brand.

Enthusiasm is not evidence.

A compliment tells you somebody can imagine liking the idea. A repeat purchase tells you somebody found enough value to choose it again. Those are different kinds of evidence. Your first job is to stop treating them as interchangeable.

Record the finding’s status separately from its zone. You can directly observe somebody expressing an opinion. That does not transform their opinion into purchase behaviour. Keep the source, date, context and limit attached to the claim.

Four sweeps.
One evidence base.

Start where buyers have already left a record: what they search, where they talk, what they review and what they ask in public. The aim is to understand the pattern, including the evidence that does not support your idea.

“Neither the positive nor negative tail is automatically ‘the truth’; record source, status, date and recurring pattern.”Stage 01 · The Evidence

Keep the uncomfortable findings.

Separate what recurs from what is merely memorable. One vivid complaint may suggest a question; a pattern across named sources gives you a stronger reason to investigate it. Preserve the category, geography and period around each finding. A pattern from a different market can guide your next question without becoming proof about your intended buyer.

When the sweeps disagree, keep the disagreement visible. High search interest alongside weak paid behaviour is a question to resolve, not an average to calculate. Ask whether the sources describe the same buyer, offer and occasion. Write the alternative explanation before deciding which finding carries more weight.

The complete research brief is available in the workbench. It specifies an output for each sweep and a gaps section. If a source is closed, an honest record of what could not be obtained is part of the work.

Look past the claim.
Read the offer.

Choose the five rivals closest to the decision your buyer will make. Record the headline claim exactly, the price ladder as actually sold, the route to market, the presentation and the recurring complaints and praise.

Then compare their leaks with your own stated angle. Does your difference answer something buyers care about? Is it a real difference aimed at a problem nobody has? Does a rival already occupy it? An attractive gap on a positioning map is not evidence that anybody wants it filled.

Good stories still
need working numbers.

Begin by challenging the inputs. Which are known? Which are assumptions? Which are genuinely unknown? Keep acquisition spending separate from the contribution generated by the transaction. A profitable-looking sale can still be too expensive to win.

Write low, base and high cases that change meaningful levers. State the source, owner and date each uncertain input should firm up. If the decision rests on values nobody has checked, call the model illustrative. Do not let a clean spreadsheet disguise a weak evidence base.

A decision needs
something to happen.

Design the smallest honest test that could change your view. Specify the offer, intended audience, price, period, cost limit and success threshold before results arrive. A real pilot booking, a transparent deposit or a paid pop-up asks more of a buyer than a favourable comment does.

The test must be honest about what exists and when it will be available. Do not imply a product is ready when it is not. If you have not run a valid behaviour test, record that gap. “Not yet” is a useful conclusion when it tells you what to do next.

Make the result interpretable.

Before the probe, record what the buyer will actually see and be asked to do. Keep a copy of the offer and price. Afterwards, report the number of people who reached the decision as well as the number who acted. A count of bookings means something different when five people saw the offer than when five thousand did.

Record what could have affected the result: an unusual audience, an introductory incentive, a broken payment step or a constraint on availability. Those limits do not make the test worthless. They tell you how far you can carry its conclusion. Decide whether the next step is to repeat the test under clearer conditions, change the offer or stop committing resources.

Give one assumption
a place to live.

Use this small working sheet to record a claim you want to investigate. It is a blank exercise, not a completed evidence register or a commercial assessment.

The verdict stays yours.

The gate asks whether evidenced demand and surviving arithmetic justify building this brand. The record should tell a sceptic what you know, what remains open and why you have chosen to continue, change direction or stop.

Give “not yet” a next action.

An open question becomes useful when it has an owner, an action and a date. Name the missing evidence most likely to change the verdict. Describe the smallest credible way to obtain it and the commitment you will hold back in the meantime. Review the decision when the evidence arrives; do not leave a cautious-sounding conclusion standing in for a plan.

The working preview above introduces the stage’s method. The full product includes the surrounding teaching, briefs, working files and review mechanics across the whole ten-stage sequence.