TPA Digital
‘Attention’ in programmatic is picking up steam. But how useful is it really?
by Dani Stewart, Senior Client Partner at TPA Digital
The benchmark of viewability has helped humanise impression delivery and works to gauge the probability of an ad being seen. However, the notion of “attention” is gaining traction. The measurement of attention undermines viewability, with the idea that not all that is deemed viewable is actually viewed – and that, ultimately, it’s a consumer’s attention, rather than the opportunity to be seen, that drives sales.
This article will explore the idea of measuring attention and its feasibility when it comes to programmatic application.
What is ‘attention’?
Attention is exactly what was intended to be measured by the concept of viewability. Before all else, a fundamental of advertising is that your ad needs to have been seen to have any impact. Viewability is a proxy measurement for attention, on the assumption that if an ad is in sight, then attention must have been paid. However, the benchmark for a viewable display impression, by IAB standards, is just 50% of the ad in view for at least one second; this does not quantify whether the ad caught the consumer’s eye, just whether it had the opportunity to be seen.
Research conducted by Amplified Intelligence directly supports a link between attention and incremental brand choice, acknowledging that high attention is scarce and that the biggest jump in incremental impact occurs when humans pay attention at a low to medium level. If this is the baseline for attention, then more granular measurement is required to truly understand which placements are legitimately being viewed. Additional studies conducted by Lumen highlighted that only 18% of viewable digital ads are in fact looked at, and that attention is variable by media and device.

Who is leading the way?
The Attention Council (TAC) was formed in 2019, to focus on the quality of media through the “lens of attention”. TAC brings together industry leaders to provide thought leadership and promote the use of attention metrics across the media ecosystem.
The key players include Amplified Intelligence, which offers advanced technology that measures human attention across more than one screen; Lumen, an eye-tracking business specialising in attention measurement; Adelaide, which combines factors to create a comparative channel view of attention deemed “attention units”; TVision, which has technology that measures in-room behaviour and presence for TV; and Avocet, a DSP that has shifted its sights to quality view-duration as opposed to viewability.
TAC also has significant buy-in from advertiser members, such as Diageo, British Gas, Tesco and Microsoft.
Last year, Co-op and Dentsu made progress with the first campaign to use the Google custom algorithm to optimise towards attention as a key variable. Lumen’s eye-tracking technology was incorporated into the process to refine the algorithm, and initial results boasted an almost 20% increase in engagement over the six-week test period. Interests in the area have continued to be piqued this year as browser restrictions dominate, causing advertiser re-evaluation of how ad effectiveness can be proved.

This year, Mars has embarked on a rigorous testing journey to understand attention, its key finding thus far is that attention is a strong proxy for sales impact. The learnings from its ongoing testing will be used to inform future media planning strategies and metrics. Publishers are also beginning to take note, a repositioning towards high attention and engagement can be seen from the likes of The Ozone Project.
Is it ready to be put into programmatic practice?
Attention is often spoken about within a broader media lens. There is no doubt it is a beneficial metric for the programmatic ecosystem to observe, but there are practical implications to this shift.
Vendors are making progress in this space, both in measurement and reporting metrics, but there remains much to overcome for a seamless integration. First, classification; attention is yet to be formally classified. What are we deeming as the most reliable attention proxy, eye-line, dwell time? And for how long? Without consistency in classification, it will be difficult to feasibly apply this measure from planning through to activation.
Attention data is still external and, in most cases, still outside the walls of the DSP and Adserver. Advertisers heavily rely on their Adservers as a real-time source of attribution truth, whereas digital attention measurement tends to require a panel-based approach. This difference of approach is disruptive to existing workflows, but it will be interesting to see if adoption is accelerated by the ongoing reduction in cookie tracking capabilities.
Publishers also face a challenge through the adoption of an attention metric. It took top-tier publishers a lot of time and resource to optimise their site towards viewability metrics, so that buyers would want to purchase those ad opportunities. With no standards for attention, how can a publisher ensure they are not unfairly penalised – buyers ignoring their ad spots – because they haven’t set up their ad placements in a certain way? Publishers have a hard time ensuring their ads work against the multitude of buyer metrics, such as cost-per-click, cost-per-acquisition, return on investment, viewability, brand-safe and so on, adding another metric might hinder their ability to generate yield further, particularly in the short term.
Until attention metrics are standardised and integrated into the wider ecosystem on both the buy and sell side, it remains difficult to quantify the value of attention certainty into costs per mille. It is clear there is value in attention, and that its highly beneficial in theory, but application requires a shift in the way the ecosystem measures success. Significant advertiser buy-in is already under way through TAC, which is helping to push developments. However, more advertisers and publishers must be willing to accept the disruption to truly challenge an approach that has been built into so many aspects of programmatic over the last decade.
From here, ad tech will take note and an overhaul can begin.