Passing a screen and noticing it are different numbers

The single most common measurement error in signage is treating everyone who walked past as audience. The gap between traffic and attention is where a screen's real worth hides.

What a count actually tells you

People-counting is the solid end of measurement. A door counter, a sensor, or a camera-based counter can tell you, with reasonable confidence, how many people moved through a space and were physically in a position to see a screen. That number — call it the opportunity to see — is genuinely useful: it sizes the potential audience and lets you compare locations and dayparts on a like-for-like basis. But it has a hard ceiling on what it can claim. It tells you how many people could have seen the screen, not how many did. Treating opportunity-to-see as actual viewers is the foundational overstatement that every inflated audience figure is built on.

Attention is softer, and easily oversold

Measuring whether people actually looked — and for how long — is a different, softer discipline. Camera-based attention systems estimate gaze direction and dwell; in practice these are heuristics that degrade in crowds, variable light, and the messy reality of people facing where the model did not expect. The numbers they produce are useful for relative comparison (this placement holds attention better than that one) but dangerous when quoted as precise absolutes. The honest use of attention data is comparative and directional; the dishonest use is a confident "X% looked" presented as if it came from a meter. A measurement practice earns trust by being clear about which of its numbers are hard counts and which are soft estimates.

The passing-versus-noticing gap is the real signal

The interesting part is the ratio between the two. A location with huge footfall but poor attention — a busy thoroughfare where everyone is rushing past — may be worth less than a quieter spot where people wait and look, like a queue or a seating area. Counting alone would rank them backwards. The gap between passing and noticing is therefore not noise to be smoothed away but the actual signal: it tells you where a screen's content is landing and where it is wallpaper. Placements should be judged on attention-weighted reach, not raw traffic, and the only way to do that is to measure both numbers and respect the difference between them.

Use each number for what it is

The practical discipline is simple to state and rarely followed: use counts for sizing and comparison, treat attention as directional rather than absolute, and never collapse the two into a single confident "audience" figure that hides which assumptions it rests on. An operator who reports "this many passed, attention here runs higher than there, and here is the one place we could measure an actual outcome" is telling the truth and is far more useful than one who quotes a single big impressions number. Measurement is worth doing only if it makes decisions better, and it can only do that if everyone knows what each number actually measured.