Marc Guldimann of Adelaide explains why we need to pay more attention to attention.

As attention metrics mature from a tactic to an integral part of the marketing stack, the specifics of how they should be used become increasingly important. One of the trickiest nuances is whether attention should be an input to value calculations or an optimization goal when buying media.

Keep in mind that it’s really only possible to optimize a system to one thing at a time. So, if you’re optimizing to the most attention, you’re not optimizing to awareness, sales, or other business outcomes.

First, let's look at what it means to optimize towards attention as an outcome. The attention paid, or the amount of time a person spends looking at something, is a product of media quality, creative relevance, and audience.

Here’s what optimizing each of these factors to drive the outcome of maximum attention actually does:

Media: Optimize media to placements that have a high probability of attention.

  • What it does: We’ve shown that media with a high probability of attention outperforms less attentive media across a range of business outcomes. Additionally, this tactic is proven to drive significantly more impact programmatically.

Creative: Optimize creative to hold attention.

  • What it does: Without strict guidelines or a more nuanced view of why people are paying attention, optimizing to attention could lead to a bunch of salacious ads that don’t drive impact. It’s relatively easy to capture attention with puppies, kittens, naked people (hopefully not in the same ad), or other elements that don’t help branding.

Audience: Target people who are most likely to pay attention.

  • What it does: Research has shown that people who are familiar with a brand are more likely to pay attention to its advertising. Because of this, optimizing audience targeting towards maximum attention is not a good idea when attempting to drive incremental awareness.

Taking these three impacts into account, it’s obvious that blanket optimization towards maximum attention is a flawed approach for most advertisers.

The most surprising effect to most marketers is the Attentive Audience Paradox: Optimizing to the most attentive audiences results in reaching people already familiar with a brand. Ultimately, this is the final nail in the coffin for optimizing every impression to the outcome of attention.

Fixation on the amount of attention paid to an impression is a common trap for newcomers to attention metrics and academics alike. After all, it is simple and makes the most sense at first blush. Unfortunately, while this way of thinking – defining attention as gaze measured in seconds – is technically correct, and the data that eye-tracking studies produce is exceptionally precise, thanks to the attentive audience paradox and nuances around creative, optimizing to the duration of eye gaze is not a good idea when trying to drive business outcomes. Duration of eye gaze is difficult to measure at scale, in most cases leveraging an opt-in panel of people who agree to have their behavior monitored, and it suffers from noise introduced by creative, as it is primarily a product of creative quality. Most importantly, if you’re optimizing towards duration of eye gaze, you’re not optimizing directly to brand or business outcomes.

The best optimization strategy is for algorithms to optimize towards the desired business outcome. So, how can advertisers best leverage attention metrics to execute this approach? The answer is by unpacking the outcome of attention and quantifying the value of its three aforementioned inputs – media, creative, and audience – in driving outcomes.

Example scenarios for optimization

Media: Attention metrics can be used to measure the opportunity for attention to placements, providing a signal to value media properly. Today, advertisers are incorporating this data into their media buying calculations to arbitrage an inefficient market. Tomorrow, we expect to see transactions denominated in media currencies derived from attention metrics.

Opportunities for arbitrage in the media market are substantial. Recently a large financial services company reduced customer acquisition costs by 25% using AU instead of viewability.

Creative: Attention metrics can be used to match creative with the quality of media required to drive the best results, as well as measuring whether people pay attention long enough to notice distinctive assets.

TVision and Realeyes recently found that lower-performing creative benefits from placement quality more than higher-performing creative. They also found that attention to lower-quality creative increases with frequency, while higher-quality creative wears out with fewer impressions.

Audience: Audience attention metrics can identify people who are engaged and will likely respond to lower-funnel messages.

At Parsec, a marketplace for media denominated using attention metrics that was the predecessor to Adelaide, we studied the correlation between attention and direct response actions. Our research showed that every second of attention on an upper funnel ad led to a 10% relative increase in clickthrough rate on subsequent lower-funnel campaigns.

These are powerful and relatively straightforward use cases for attention metrics, but they also demonstrate the importance of understanding the nuances, especially the Attentive Audience Paradox.

Our society has borne the consequences of optimizing to attention at all costs in the Facebook newsfeed. Hopefully, the advertising industry will take a more sophisticated and evidence-based approach to applying attention metrics rather than merely optimizing blindly to the maximum possible attention.