The advertising industry is in the business of persuasion. And yet, for all our keen abilities and insights in the art of persuasion, I often find the argument that advertising is a data-led profession to be anything but. This is not to say that data and insights have no place within advertising, merely that it is often treated as a “box ticking” exercise in practice rather than an exercise which leads (in advance) or diagnoses (after the fact) any given marketing strategy.
This hits close to home. The entirety of my career has been within, or involved with, the discipline of “measurement science” - the fancy title we give to outcomes-based behavioral research in marketing and advertising. Measurement science comes in a number of fun flavors and shapes, but has largely been concerned with dimensioning the “three Rs:”
Reach: Accurately describing who received a marketing message.
Resonance: Understanding what cognitive impact the message had on those reached.
Reaction: Understanding the extent to which the message drives a consumer to a desired action.
And yet, for digital media (particularly “programmatic” ad buying) these goals are boiled down to a much more simple version of the “three Rs” for your average ad buyer:
Reach: How many people were served the ad?
Reach (again): Did the campaign check the boxes for viewability?
Reaction (sometimes): Did someone click on it?
Anything beyond reach (and frequency) is merely the aforementioned exercise in “checking the box” - so long as a campaign meets an acceptable threshold for advertising viewability (e.g. how much of an ad appears on the screen, and/or for video how much of the spot is played) it is considered to be at the right level of quality. Having checked that box, the only real metric that media planners are buying against are the same ones which we’ve been relying on since the hallowed early days of TV advertising - reach and frequency (the latter concerned with how many times an individual is messaged).
It is true that our capability to understand reach has been dramatically improved in this digital world, and massive web platforms have enabled buyers to purchase reach in an abundance and on the cheap. Naturally, with a catch - the accuracy of “delivery” metrics like reach and frequency have been increasingly questioned within the advertising community given a propensity for fraud (i.e. misreporting, etc.) to penetrate the skunk-work of black box systems constructing the programmatic landscape. Our concerns in advertising measurement in recent years have thus disproportionately focused on issues of trust and fraud, meaning that as we’ve become more “advanced” in our ad buying technology many classical goals in measurement science have fallen on the wayside - small things like, you know, whether the ads are even doing anything. The emphasis has disproportionately been directed at the “delivery” of media (the first “R”) rather than the “impact” (the other two “Rs”).
In fairness, measuring impact is a much tougher problem than measuring delivery by an order of magnitude. Advancements in direct response measurement such as multi-touch attribution (e.g. trying to figuring out all the points which influenced those clicks) are certainly newer, though not applicable to many goals in advertising (e.g. brand), are rife with methodological challenges, and it’s not entirely an open-and-shut case that clicks on a digital ad mean much for advertisers that can’t also directly provide a service through the internet. In those cases, once the potential influence of an ad leaves the ad-tech nest and the intensely (and controversially) tracked digital media world, things get really complicated, real quick. Observational sales lift methodologies, geo-tracking, and other tactics have made headway for specific industries such as consumer packaged goods or retail, but their lack of cross-channel comparability (i.e. whether sales lift can be assessed across tactics on various social platforms, search, open web, etc.) has relegated these techniques to specialized cases.
As a result, efforts to improve delivery metrics towards having a more meaningful gauge on impact have spurred the most interest in modern media measurement. In recent years, the concept of “attention” as a metric for measurement has dominated measurement-related conversations. The hitch is that no one knows what this means.
When asked what metrics like “ad recall” or “awareness” mean, advertisers may use slightly different words but will generally provide answers that are logically consistent. Comparatively, attention is all over the map. On the one hand, it describes measurement which dimensions ad delivery beyond the rather low bar of “opportunity to see” an advertisement which is implied by reach. On the other hand, the suite of methodologies proposed by the current marketplace (which is expanding rapidly) is disparate enough to demonstrate that the specifics of the definition are far from industry-wide agreement.
The concepts of reach and frequency have dominated the ad buying landscape because what they mean is both universally accepted and (roughly) applicable to a multitude of different advertising channels. Standardized definitions allow for standardized measurement which can be used as a “currency” - things like the Nielsen TV rating which are used to measure value. Lacking definition, the potential for Attention risks squandering in the same manner of the digital impact metrics outlined above - a specialized metric for specialized cases. Useful tools, but not impactful at the industry level, and certainly not a currency.
I use the word “potential” because not only is there ample opportunity for a new set of currency metrics to be introduced into ad-buying, it is badly needed. The emphasis on reach and frequency has led to an industry which prizes the scale (breadth) of an execution above all else; a reductive view which tends to ignore how consumers actually engaged with media in differential ways (depth). This implicitly forces us to think about actual media engagement no differently than the classic TV viewership model, which both belies advancements in media technology and the increasingly complex manner in which consumers interact with media (in smaller doses, on smaller devices, on their own time).
It is for this reason that Attention metrics have been particularly top of mind for marketing opportunities in gaming. While the gaming ecosystem has an abundance of scale, the superpower of gaming as a medium are the long sessions and relatively undivided (you guessed it!) attention that consumers give to gaming experiences. There is no “second screen” in gaming, meaning that the potential for marketing in these environments is particularly potent for driving impact. Yet again, this immense potential for impact is often not the primary gauge advertisers use when considering opportunities in gaming, which is inevitably a contributing factor to lagging investment in the sector.
Unfortunately, the way forward is not abundantly clear. Getting a bunch of marketers to agree on just about anything is fairly tricky. In the meantime, we’ll likely see any number of stop-gaps put into place to capture some definition of attention, if not the eventual definition of attention. I spoke with Alex Lee at Digiday last week on our efforts at Activision Blizzard, which entails a “scorecard” of various metrics that provide a picture of “Attention” that can be attached to campaigns in a turn-key manner (something that isn’t always possible with third party methodologies). It’s not perfect, but it is something to keep the conversation moving forward while the industry (hopefully) moves towards a unified definition.
The conversation around attention is important not because of the specifics of an attention metric per se, but as a leading indicator of the inadequacy of the metrics that media is bought against today. In some ways the specifics of what Attention should be are less important than the fact that it must be materialized as a metric - consumers do not utilize media the way they did ~70 years ago, and yet our measures and means of assessing media are designed as if they do. These faults will only become more glaring as consumers increasingly adopt interactive forms of media - the only question is whether or advertising systems will be prepared for this world.