Real audience research, kept alive so you can put the next decision to it.

People who live in Amsterdam and are between 23 and 30, recruited with deliberate variation in gender, neighbourhood, education, work, income, housing, background and how they spend. The group is narrow enough to research properly and broad enough that many companies have a decision worth putting to it.

This is not “all young people” and the citizens do not speak for Amsterdam. They are specific people. That is the point: a specific reaction, objection or difference in perspective is what changes a decision.

1. 1

   A long interview and a questionnaire. Each person talks at length about their life, money, taste, what they buy and what reaches them, then fills in a structured survey on the same ground.

2. 2

   Observations, checked word for word. What a person said is extracted as short, citable observations. Any observation whose words are not verbatim in the interview is thrown out at ingest.

3. 3

   Expert readings, kept separate. Notes from a psychologist, a behavioural economist, a brand strategist and a media planner are inferences across observations. They are labelled as inferences and name what they rest on.

4. 4

   Answers with a receipt. A citizen answers in one voice, then each material claim in the answer is filed: the interview says it, it is reasoned from something said, it is plausible but unbacked, or it was never researched. The underline under a sentence shows which.

5. 5

   Independence first. On a question, every citizen answers alone before anything is summarised, so disagreement is shown rather than smoothed away.

A citizen may reason to a situation the interview never covered — that is what it is for — but it may never invent its own past, a figure, or a share of anyone. Where the research runs out, the answer says so.

Each opens the same demo with the question filled in. Pages for specific industries follow once there are sector examples worth showing.

1. 1. Context, optional. A clothing brand pastes its website. The page is read once, summarised in two sentences the visitor can correct, and three questions are suggested for that business. The citizens get the context; they do not become the brand’s customers.
2. 2. A first reaction. “What would make a €180 jacket from this brand feel worth it to you?” Each citizen answers alone, in two to four sentences, with the receipt under every claim.
3. 3. In short. A synthesis above the answers: the decision-relevant thing first, where the three differ if they do, and what these interviews cannot settle. Three people are never turned into “most”.
4. 4. An objection, then a revised idea. The second and third questions push: what would make you hesitate, and if we fixed that, what would you do? Follow-ups count as questions.
5. 5. The gate. After the third answer the conversation stays on screen and the next question needs access.

3 people interviewed so far, in September 2026, 2 of whom have also completed the questionnaire. The first interviews were all with men; recruiting now runs across genders, neighbourhoods and incomes. The plan for this audience is twenty to thirty interviews, then the same questionnaire fielded to several hundred people so that how common a pattern is can be measured rather than guessed.

Each citizen is tested blind: it fills in the questionnaire before we see what the real person answered, and the two are compared. Early runs are in. We do not publish an accuracy figure until the retest is done, and we do not claim the citizens predict what Amsterdam will do. This is decision support — an audience pushing back before you commit — not a forecast.

Right now we run pilots. You bring one live decision — a campaign, a price, a product, a page — and we put it to the audience with you, then you keep access for the follow-ups. Self-serve access is being prepared; pricing and terms are settled with the first pilot clients rather than printed here.

Clients get controlled access to ask and test, with excerpts and aggregates. Nobody gets raw transcripts, row exports or the citizens themselves.

Need your own customers, another city or another age band? We research it with the same method and you get it to yourselves. A shared audience is faster and cheaper; a private one is yours. Either way it starts with one decision you are weighing now.