The Advertising Research Foundation published Phase 3 of its Attention Measurement Validation Initiative on 9 June 2026. The study ran four national campaigns across television, online video, social and digital display, and tested whether the leading attention methods agree on what they are measuring. At the channel level, they largely do. At the platform and placement level, they do not. That gap matters because the industry has been treating an attention score as a single comparable number a planner can move between line items. Phase 3 says it is not, and the consequences run through how buys are written, how vendors are compared, and how creative is judged.
Four points follow from the ARF's findings. They are worth taking in order because each one closes off a shortcut planners have been using.
1. Channel-level agreement is the floor, not the ceiling.
The headline result is the comforting one. Across the four campaigns, attention methods broadly agreed that television and online video deliver higher attention than social and digital display. That ranking is stable enough to plan against at the channel level, and it confirms what experienced buyers already assumed about the relative weight of a full-screen video versus a feed unit. The trouble starts the moment a planner zooms in. Below the channel level, the methods diverge. Phase 3 is clear that the disagreement is not noise around a shared answer. The methods are measuring different things and producing different rankings. A planner who uses one vendor to justify a channel mix and another to justify a placement mix is stitching together two incompatible views.
2. No clean line runs from attention to brand lift.
Phase 3 also pushed against the assumption that more attention reliably produces more brand lift. It did not. The relationship was inconsistent across the four campaigns, and lower-attention channels often carried their weight through frequency and reach rather than per-impression engagement. That finding should land hard with anyone who has been quietly retiring viewability and reach in favour of an attention-weighted plan. Attention is one input. It does not replace the older arithmetic of how many people saw the work and how often. The ARF's recommendation is to understand what each method actually measures and to treat the resulting score as a description of that specific signal rather than a universal currency. Read that way, attention earns a seat at the planning table without demanding the head of it.
3. Method consistency belongs in the buy, not the post-campaign deck.
The operational implication is the part most easily ignored. If methods diverge below the channel level, then comparing a YouTube line measured by vendor A with a TikTok line measured by vendor B tells a planner almost nothing about which one delivered more attention. The two numbers are not on the same scale. The fix is straightforward and unglamorous. Pick one method per campaign, write it into the insertion order, request the vendor's platform-level norms in advance, and hold the line when a sales team offers a different score from a different methodology that happens to look better. That discipline has to start at the brief, not the wash-up. Planners should also weight attention as one input among reach, frequency and creative quality, not as a substitute for strategic judgement about what the campaign is trying to do.
4. Creative and frequency carry the channels attention undersells.
The Phase 3 finding that lower-attention channels can still drive lift through weight and frequency is a quiet rebuke to the idea that display and social are simply worse media. They are different media. A digital display unit that earns a fraction of a second of attention can still build memory structures if the creative is built for that fraction of a second and the frequency is set accordingly. The implication for production is concrete. Brands running attention-led plans should not pull budget out of low-attention channels and assume the lift moves with it. They should brief creative that works inside the attention each placement actually earns, and let the media plan reflect that the job of a display unit is not the job of a fifteen-second pre-roll.
Phase 3 does not retire attention as a metric. It retires the idea that an attention score is portable across vendors, platforms and placements. The practical response is to hold one method consistent across a campaign, treat the resulting numbers as descriptive rather than comparative, and keep reach, frequency and creative craft in the planning conversation where they belong. Attention is a useful input. It is not the plan.