Marc Guldimann, CEO and Co-founder, Adelaide, describes how his company measures attention to ads and why duration metrics are so important.
By now, it’s common knowledge that attention is key to driving advertising outcomes. But when referencing “attention metrics,” marketers often mistake a part for the whole. Duration of viewability and eye gaze, for example, are frequently cited as valid proxies for attention. While both are vital contributors to a holistic understanding of attention, reliance on individual inputs carries risks currently associated with the binary metrics these inputs are meant to replace.
First, let’s explore the distinction between viewable duration and eye gaze.
- Eye gaze duration, or focus, is the amount of time someone is looking at something. It’s typically captured using eye-tracking hardware and serves as a good measure of how long creative holds attention.
Both have been shown to correlate with increased campaign performance but pose serious limitations when measuring media quality.
Viewable duration doesn’t consider important aspects of the experience, namely the number of competing ads on a page, the size of the placement relative to the screen, and the placement’s position. In the end, optimizing to the cheapest duration alone results in small ads on big, cluttered screens. Not an ideal recipe for attention.
Eye gaze duration is challenging to measure at scale, in most cases leveraging a panel of people who agreed to have their behavior monitored. It also suffers from noise introduced by creative, as duration of focus is primarily a product of creative quality.
Beyond shortcomings in measurement, duration as a standalone metric creates rather perverse incentives for publishers to produce sticky placements and show ads to people who are most likely to pay attention rather than the ideal audience.
So, how do we leverage the obvious power of duration into a metric that adequately measures media quality while controlling for its shortcomings?
At Adelaide, we use machine learning to assess several metrics proven to predict attention, including eye gaze and viewable duration. The model synthesizes these signals and returns a media quality score for every placement. By continually feeding our model additional data, we’ve trained it to closely proxy outcomes. The result is AU, an omnichannel metric that measures each placement’s ability to capture attention, prevent distraction, and contribute to outcomes—creative and audience held equal.
If we are thoughtful about the metrics we use and the incentives they create, the future is bright for media buyers and sellers alike.
For more information on Adelaide go to Adelaide – Attention Metrics for Digital Media (adelaidelift.com)