Attention as the new currency

The industry has counted impressions for thirty years, knowing that an impression describes delivery rather than perception. Attention measurement is the attempt to count what actually reaches a person.

Why impressions mislead

An impression is recorded when an advertisement is served. Viewability adds a low bar: half the pixels for one second.

Neither describes whether anyone looked. Research using eye tracking consistently finds that a large share of viewable impressions receive no visual attention at all.

What the measurements show

Consistent findings across attention studies:

The ranking is stable and it does not correspond to cost per thousand. Some of the cheapest inventory delivers almost no attention, which makes it the most expensive per attentive second.

The recalculation

Buying on attention changes the arithmetic in specific ways.

A format costing three times more per thousand but delivering six times the attention seconds is half the price on the measure that matters.

Large formats, high-dwell environments and full-attention media move up. Cheap remnant display and small mobile formats move down, often to the point where their inclusion cannot be justified.

How it is measured

Three approaches, all estimates.

Panel eye tracking. A recruited panel with cameras, measuring actual gaze. Most accurate, smallest sample.

On-device signals. Scroll speed, dwell time, interaction, audio state. Available at scale, indirect.

Modelled prediction. Attention estimated per format, placement and context from panel data. The practical option for planning.

None is precise. All are considerably closer to reality than counting served impressions.

Practical application

1. Request attention metrics from measurement providers and sellers; several now offer them.
2. Compare on cost per attentive second rather than cost per thousand.
3. Drop the formats that deliver nothing, which usually frees a meaningful share of budget.
4. Reinvest in fewer, better placements.
5. Validate with a brand study or a holdout, because attention is a proxy for effect and not effect itself.

The caution

Attention is a better proxy, not a guarantee. An advertisement that is looked at and disliked has consumed attention and produced a negative outcome.

The creative still decides what happens with the attention that the media buying delivers.