Mike Follett, managing director of Lumen Research, explains the science of eye tracking – and offers five media applications for the data it gathers…

For many years, the notion of OTS – opportunity to see – has been central to media metrics and widely used to indicate the success of advertising.

As any media person knows, it enumerates the chances a reader or viewer will get to see an advertisement.

But of course, while we might have had a guess at how many times a person could have looked at an ad in a given campaign, we never knew for sure whether they really did. How could we? We weren’t tracking everyone’s eyeballs.

But on that very subject, what was once an impossible dream is now perfectly realistic. Eye tracking is now an entirely real, scientifically credible process, and it presents the advertising and marketing business with its very best opportunity to measure the metric that everyone cares about: attention.

Attention matters very much to brands because, as numerous studies have shown, it is the key ingredient in the success of advertising. Dentsu’s Attention Economy study, which uses Lumen’s eye tracking data for its mobile and desktop insights, proved that attention was 700% more powerful than viewability alone when it came to predicting changes in brand choice. There is a little more to it than attention = persuasion, but the fact remains that when it comes to attention, more is more.

So whereas the old calculation involved the opportunity to see, we can now establish, far more usefully, what was actually seen. We still don’t track every eyeball, of course, but large, representative panels give us detailed data sets that allow us to power models that predict the attention that will be paid to a given ad on mobile and desktop devices.

Here’s how it works:

1) Proprietary software turns the webcam of a phone or computer into a high-quality eye-tracking camera. The system calibrates the camera to the user’s eyes, estimating attention within a circle of one inch in diameter. The software looks for a glint of light in the user’s eye and uses this information to estimate where they are looking on the screen. (Needless to say, these users are all volunteers – your eyeballs are not being tracked right now!)

2) The software records placement data (what was on the screen, where and for how long); and attention data (what ads people actually looked at, and for how long).

3) This is a privacy-first process, incidentally. All processing takes place on the user’s device and no video data or personally identifiable information is uploaded to the database.

4) The technology is entirely scalable, and can be deployed in any country in the world, except China and North Korea. It can be used for ad hoc projects, where respondents are recruited for specific research tasks, or on a more permanent basis. In the UK and the US there are two ‘Passive panels’ of 1000 people, who download the software to their desktops and mobiles enabling their attention to be tracked across all the sites they visit across the day. All panellists are fully consenting and compensated for their participation. All data collection, storage and processing are entirely GDPR compliant.

5) The placement and attention data from the eye-tracking panel is then used to power predictive models that estimate likely levels of attention.

What happens next? There are numerous uses for data of this kind. Eye tracking can help brands measure, buy and amplify attention to their marketing, and it can help with planning, reporting and optimising attention. Specifically:

1) Some agencies and auditors use eye-tracking data within their planning and valuation tools, factor them to their reach and frequency estimates to understand the real attention a campaign will generate – and how much it will cost to ‘buy’ that attention in one platform versus others.

2) Some partners make use of an LAMP tag – a piece of JavaScript that records the viewability characteristics of an ad with each impression and then estimates how much attention it is likely to have generated. This attention data can then be matched to CPM data to give live insight into the cost of attention, as well as click and conversion data to help understand the performance value of attention, and brand uplift data to give insight into attention’s brand value.

3) Viewability data may be ingested from the walled gardens of Facebook, Instagram, YouTube, TikTok and Snapchat, to give clients consistent attention data for ads across those platforms as well as the open web.

4) The predictive model can be integrated into DSPs to allow brands to channel investment towards the inventory that is most likely to get looked at, and away from ad placements which, though viewable, are unlikely to be viewed.

5) There are also tools that can test and optimise creative, and use this data to create brand-specific attention models.

The smartest publishers, ad tech providers, agencies and brands are using these tools to reduce wastage and drive value. Eye tracking, so hard to imagine even just a few years ago, is not a technology for tomorrow’s world, but one that is delivering huge advantages right now.

Also published in: Performance Marketing World

Podcasts powered by AudioHarvest