We read markets
through behaviour.

Markets tell their stories through behaviour.

Most of what an organisation knows about a market comes from asking. Asking is a good instrument and we use it: the official consumer surveys are part of what Cadence reads every quarter. But a survey is a statement about intention, collected from a sample, at a moment that has usually passed.

Alongside it, a market leaves a public trail of what it actually did. What it searched for. What it booked and how long it stayed. What volume moved through the tills, in which categories. What the news around it sounded like. None of that requires anyone's participation, all of it is dated, and all of it can be checked against its source.

Read together, the two instruments answer different questions, and the useful moment is often where they disagree: a market that says one thing and does another is telling you something a single instrument would have missed.

Different signals. One consistent read.

Cadence brings cultural, consumer and economic evidence into one frame: what people search for and watch, what households say, what they buy and book, the rate conditions they sit inside, the tone of the news around them, and the advertising brands put behind all of it.

Each signal is read on its own terms and never blended into another. There is no composite, no index and no score, because a single number that averages six instruments is a number nobody can take apart again. What makes it one read rather than six feeds is that every market is measured the same way, in the same window, against its own history, every quarter.

Where two signals point different ways, Cadence reports the disagreement and infers nothing from it. Signals observed in the same window are observations about the same period, not evidence that one produced the other.

Evidence before conclusions.

We decide what would count as an answer before we look. Every test is pre-registered: the question, the criterion and what would falsify it, written down in advance, and the result published whichever way it goes.

That discipline has cost us our most commercially attractive claim. We pre-registered five tests of whether the signal predicts demand, and published all five results. None cleared the bar we had set. So Cadence is a descriptive read and nothing in it is sold as a forecast, which is a narrower product than we would have liked and the only one the evidence supports.

Inside a quarter the same rule applies in miniature. Before a figure publishes, we test whether the instrument reproduces it; where it does not, the figure is withheld with the reason stated rather than estimated in its place. Everything that does publish traces to a named public source and a date you can check.

From research to intelligence.

The research is the foundation. What an organisation buys is the quarterly read it produces, in three forms.

The country report

One market in depth, chapter by chapter, read against its own history. Every market opens as a public preview; the depth behind it is what you buy.

The European Control Room

All eleven markets on one footing: where they repeat, where they split, and where two independent measures of the same market disagree.

The Cadence Skill

The same evidence inside the AI assistant your team already uses, answering in your own words for the markets you hold, with the receipt attached.