How Measurement Panels Priced Whole Industries
The first serious question a sophisticated buyer asks us is the right one: do you have all of it?
No. Nobody does, and anyone claiming otherwise about the physical world is not being careful. What matters is a different question, and it is one the measurement industry answered decades ago.
The precedent
Television advertising was priced for seventy years on Nielsen panels covering a fraction of a percent of households. Radio, print circulation audits, retail scanner data, credit bureaus, economic statistics — all of them are structured samples, and all of them were sufficient to price entire industries.
They worked not because the sample was large but because it was consistent: measured the same way for everyone, so relative positions held even where absolute counts did not. A show with twice the rating of another really did have roughly twice the audience, whatever the true absolute numbers were.
That is the property that matters, and it is the one to interrogate.
Why relative measures survive partial coverage
Every number we publish is relative by construction. Share of Presence is your reach as a proportion of your category's tracked reach. Company scores are percentiles within a category. Event trajectory is overlap between editions, expressed as a percentage.
Two competitors in the same category face the same coverage process — the same organisers, the same list-publishing conventions, the same geographic mix. Whatever we are missing, we are missing similarly for both. So the ratio between them survives even when the absolute counts are a fraction of the truth.
This is not a rhetorical move. It is why we refuse to publish absolute counts of anything. "This company did forty-seven events" would be a claim about the world we cannot support. "This company's reach is in the 78th percentile of its category" is a claim about our record, computed identically for everyone in it.
Where it genuinely breaks
The honest failure mode is coverage bias that correlates with the company being measured. If a competitor's footprint skews toward small regional events that publish nothing, they will read lower than the truth — not because the sample is small, but because it is differently incomplete for them than for you.
That is a real risk, it is nameable, and it is testable. Splitting our record in half at random and computing the same rankings on each half shows whether ordering survives sampling. Comparing our record against editions where we hold a verifiably complete list shows where coverage is thin.
Those tests do not make the record complete. They tell us whether it preserves the answer — which is the only question that was ever relevant.
What we owe you
Three things, and we would hold any measurement provider to them.
Say what you measured. Ours is a share of tracked participation, stated on every number.
Publish the method. Not the proprietary weights, but what goes in, how it is computed, and where it stops.
Show the gaps as gaps. Where our record is too thin to say, the output is a dash with a reason, not an estimate wearing a confident face.
The alternative to a measured sample is not perfect knowledge. It is the brochure — and that has been the industry's only source for as long as the industry has existed.
Calculate your Share of Presence at whr.ai.