No Attention. No Impact.

In some ways everything, and in some ways nothing. A line in the sand was drawn last June when I was asked to guest edit a WARC guide on planning for attention. I called in experts to discuss three key structural factors for an industry on the precipice of change. These included measurement, application, and ethics. The edition included papers from heavy hitters including MARS, GSK, IPSOS, Spotify, Dentsu, Attention Capital and more. The eagerness to which these thought leaders contributed was testament to the appetite for change. Currencies are starting to form, the nature of measurement is becoming more advanced, capital investment is starting to flow, and the study of attention is a very fast growing field. All of this is a good thing because the audience measurement industry needed significant disruption to shake off the damage that legacy metrics have done. The Attention Economy is no longer coming, it is here, and every agency holding group, most publishers and many large brands are actively investigating or integrating attention data into their 2022 plans.

What hasn’t changed is that humans are humans. They get distracted by other humans, pets, children, their environment – basically, everything around them. Humans don’t care which part of your sales funnel they are in, upper or lower. They just don’t care. So while an attention economy is changing the way the world is trading media, it is in no way changing how people consume advertising, nor can you (as an advertiser) change someone’s viewing behaviour.

What we know is that viewing behaviour is driven largely by the functionality of the platforms. Put another way, platform functionality (i.e. the user experience) is the biggest driver of in-attention. Things like spatial clutter, scroll speed, scroll depth, page geometry and ad context all contribute significantly to the number of attention seconds achievable. This is why we see some platforms and formats deliver vastly different levels of attention to others.

Many of you will also be thinking about the role of creative. Our results show that creative is no match for platform functionality. For example, an ad that gains high levels of attention on a quality platform will gain lower levels of attention on a poorer quality platform. This finding is generalisable. Every single time we collect data, the level of attention gained from the same creative diminishes in line with the performance of the platform. Even though good creative helps, the platform is the biggest driver of attention. If creative execution was the driver of attention, it would get the same attention on every platform. But it doesn’t.

Politics, priority lists and validation of value: barriers to change such as politics and priority lists will make or break a transition to better metrics. Audience measurement stakeholders (like brands and agencies) don’t like uncertainty, and media owners don’t want to be made responsible for the variable impact of different creative executions. So, in these critical years of exploration, careful change management strategies are vital to manage politics and priorities. The building blocks of change include validation of the value of attention via rigorous and generalisable research, appropriate product development and application, and thought leadership – all at the same time.

New measures need to be proven to work. And by ‘proven to work’ I mean that attention can be validated as valuable by: (a) showing significant and repeatable improvement in ad effectiveness over legacy metrics, and (b) that the measurement practices can be established as both best practice and scalable. In other words, the industry needs to trust that attention can fill large gaps in the relationship between dollars per impression and effectiveness per impression. It also needs reliable vendors and technology that is rigorous, privacy compliant and scalable. It’s now measurement companies, not marketing companies, that will lead the way.

Appropriate application and use cases: a reminder that the ultimate problem here is a currency problem. And our currency is failing because the relative value of one impression over another can’t be quantified. This means that any measurement system, model, methodology or concept that relies on the notion of equitable impressions as a data source will fail. Budgeting, market mix modelling, media buying, media planning to name a few. Take budgeting for example (where someone considers Share of Voice). If you are overspending on low attention media platforms and underspending on high attention media platforms while your competitors are doing the exact opposite, even if you think by volume of impressions that you are spending the same, the principles of SOV fail and your brand will decline. Put another way, you spend the same, but the platforms you choose get less attention and you don’t know that SOV fails. But best practice attention data can be used as an index of, prediction of, or measurement of equitable performance. And there are three application verticals that are shaping the ecosystem:

Planning: attention data provides deep understanding to help direct investment and media dollars for maximum effectiveness. This can be applied to media plans, Share of Voice budgeting and creative planning and is often provided as an index of equitable performance.

Buying: deep and rigorous attention data can be used to build probabilistic models for the purpose of predicting attention on digital ad inventory at the point of the transaction on programmatic platforms. This means that an advertiser can reduce wastage by rejecting placements with low attention or increase bids on placements predicted to return higher attention, in real-time.

Measurement: attention data can be used to build tag-based technology to predict the attention performance of an impression after it is delivered. This may be used as a verification of the accuracy of a pre-bid model or to test the performance of different platforms and campaigns.

The industry needs to be wary of advanced viewability dressed up as attention. Attention and viewability are NOT the same yet they are often used interchangeably. Viewability is about how the ad is presented on the screen to a viewer, attention is how a viewer responds to the ad presented on the screen. One is device-based measurement, one is human-based measurement. If the measures were the same, they would not only be highly correlated, but they would move in the same direction (when one goes up the other goes up). They don’t. Viewability may serve to count the opportunity, but attention measurement is linked directly to the business outcome. And the difference in brand outcomes between viewability-optimised and attention-optimised are significant. With the attention economy in hyperdrive, many current vendors of viewability are attempting to capitalise and are making vast claims about measuring attention (often called advanced viewability). 

But viewability measurement technology falls short on measuring human viewing and there are a myriad of reasons why. Viewability measurement relies on code executed inside a browser: it cannot tell where the ad slot is on the page, it cannot tell if the browser is behind some other programme like PowerPoint; it certainly cannot tell if the user sitting in front of the computer is human. Viewability won’t tell you that even if there is a human behind the view, they are most likely highly distracted (usually by other things on the page). On average, between 70-80% of an ad has no active attention (eyes-on-ad) paid at all. Human attention metrics are the natural evolution of the viewable impression era. Best-practice attention metrics are collected passively from a human viewer. They monitor genuine, not simulated, platforms so that the respondent is viewing naturally. The gaze models should achieve accuracy to an eyes-on-ad level without the need for interruptive calibration. And of course, the process should apply GDPR standards for data privacy and storage.

A new measurement category needs ethical leaders to guide the developing conversation. It needs a clear and obvious way to sort the burgeoning new quick-fix attention measures from the rigorous attention measures. We need to be careful not to make the same mistakes as we did in the mid-noughties when disruption and blitz-scaling created (accidental) media companies who were the perfect combination of unconventional thinkers, risk takers, venture capital and commercial genius. Instant measurement appeared in an instant which sounded at the time like the exciting thing we all needed. Fast track to today and the industry is well aware of the negative flow-on effects of click bait style engagement metrics that made attribution promises they couldn’t keep. And worse, advertisers and brands have been devalued by the diminishing attention consumers pay to degraded content.

This is why I believe it is too soon for an attention CPM (aCPM). If we are planning to correct the course of a broken ecosystem that focuses on dubious metrics, we need metrics that value attention. CPMs are set on the notion of platform/format performance, without having any validation of platform/format performance.

With a CPM that is not commensurate with performance, you might end up paying more for a format that actually delivers lower levels of attention. This is a form of dirty data. Numbers without meaning, information without justification.

Applying dirty data to a measure that is accurate like human attention is like putting the wrong fuel type in your car – it will kill the engine.

For example Platform A turns out a lower aCPM than Platform B, but Platform B delivers twice the attention seconds (simply because the natural CPM of Platform A was just under half the price). Does this make Platform A better quality? No, it just means we are (still) chasing the lowest CPM, not paying for the value of attention. Considering platform effectiveness on an attention CPM alone does not automatically point you towards a quality-based trading currency. We are here to drive positive change in a broken ecosystem, so we need to be careful as an industry and take steps to get to an attention trading currency that incentivises quality. This will take time.

So, as we enter a new era of measurement, what have we learned from the last effort? We have learned that attention can just as easily be used to drive us to the cheapest, darkest corners of the internet. We need to truly understand attention metrics in order to use them to drive quality.

Measuring how humans respond rather than how advertising is delivered offers a more stable future. As a human-centred measure, attention is naturally resilient across changing devices and viewing environments. Over the next few years, attention metrics will be integrated into all major agency planning systems and the flow on effects from this will be that brands will be using attention metrics to minimise the effects of inequitable impressions. Collection technology will be significantly advanced with data collected from platforms other than digital video (such as cinema and podcasts). The applications in the ecosystem will have landed and this will include solutions for planning, buying and measurement. The industry will be flooded with vendors trying to capitalise on attention metrics. Some will be good, many will not. 

Karen Nelson-Field, CEO at Amplified Intelligence

Professor Nelson-Field is a globally acclaimed researcher in media science. She is a regular speaker on major circuits, including Cannes and SXSW, and has secured research funding from some of the world’s largest advertisers. Her first book, Viral Marketing: the science of sharing, set the record straight on hunting for ‘viral success’. Her most recent book, The Attention Economy and How Media Works, explains the stark reality of human attention to advertising. Her research has been noted in The NY Times, Bloomberg Business, CNBC, Forbes, Wall Street Journal, Huffington Post, and AdAge. She has a regular column, Attention Revolution, with Mediatel News. Karen’s commercial work combines tech and innovative methodological design to build attention measurement and insight products to guide the industry through a disrupting digital economy.

Professor Karen Nelson-Field : Karen Nelson-Field PhD | LinkedIn
CEO and Founder, Amplified Intellligence : Amplified Intelligence: Overview | LinkedIn