An industry narrative is emerging that has the potential to exacerbate the structural flaws in digital advertising rather than usher us into a new paradigm. Some of our industry’s sharpest minds are claiming that we are entering the Outcomes Era.
Instead, I posit that we have been in the Outcomes Era for over a decade. It is high time we exit it and embrace our entry into the Quality Era.
There are numerous ways to frame this debate, all of which should lead to the conclusion that quality, not outcomes, will be the dominant currency of the next era of advertising, as it was in our bygone analog/broadcast era.
What era have we been in?
Does anybody seriously think this industry is in a good place with its obsessive focus on proving efficacy with outcomes? This obsession often results in media taking attribution credit for business outcomes it did not contribute to and the propagation of easily gamed metrics masquerading as outcomes.
All of this happened while we had wide-scale user/device identification with nearly no restrictions on its use for targeting or measurement. But our access to these signals is diminishing.
A sizable and growing portion of digital media exposures are no longer connectable to a single known user or the purchases they make. So how do we even stand a chance of building a healthy outcomes-based system when we couldn’t even do it in the third-party cookie era?
The only way this could possibly work is if the system adopts, and buyers accept, an egregious level of unexplainable and unauditable probabilistic modeling, which is simply untenable.
What is an outcome, anyway?
Ad tech notoriously latches onto vague terms, some carrying tremendous weight. Outcomes is one of those, not coincidentally alongside its synonyms performance and conversions.
In my almost two decades in the digital ad space, I’ve heard nearly every measure referred to as an outcome, from “hard outcomes” like sales and store visits, to “soft outcomes” like reach, viewable impressions and clicks.
If the word outcome is completely open to interpretation, isn’t it unwise to center it as the driving force that fuels our industry’s measurement and buying decisions?
Goodhart’s law explains the many failings of our industry: “When a measure becomes a target, it ceases to become a good measure.” If nearly all of the measures we refer to as outcomes have stopped becoming good measures once they became targets, why would that be any different with the vague umbrella term as well?
What is media quality, anyway?
It’s easy to knock outcomes. But let’s not assume that the case for quality has been proven, even though most agree about its importance.
After all, quality is also often thrown around as a meaningless buzzword.
Let’s narrow the scope to just focus on media quality, setting aside creative and audience/data quality.
The most succinct definition I’ve been able to come up with for media quality is “the collective sum of attributes and characteristics of ad inventory that are not tied to an individual user or device, which indicates the relative value to an advertiser given the predicted likelihood of that ad driving a desired outcome.” That desired outcome can be a proxy to a short-term and/or long-term objective, but the lift must be truly incremental.
Media quality can simply be thought of as the prominence of an ad placement, the potential for it to capture attention, and whether the context of the ad exposure could affect user receptiveness. (I explored media quality across a number of lenses here.)
Are outcomes and quality mutually exclusive?
Can one have quality without outcomes? Yes and no.
Quality is defined by its correlation to outcomes, but that doesn’t mean that every media exposure needs to be deterministically correlated to an outcome. Instead, probabilities across large groups of ad exposures, users and outcomes can be triangulated to find various proxies.
Can one have outcomes without quality? Also, yes and no.
It depends on how an outcome is defined. MFA sites, outstream video misdeclared as instream and even nonviewable impressions have notoriously been given attribution credit for outcomes justifying many billions of dollars.
Outcomes validate quality. But because outcomes are so loosely defined, they’re easily gamed by low-quality inventory. So quality should be our primary focus.
Quality, with a splash of outcomes
However, outcomes remain critical. No respectable business leader will blindly permit marketing dollars to be spent without some proof it’s contributing to a business outcome.
As cookies disappear, attribution models will buckle under pressure from a dearth of ground-truth data. Marketers will reembrace proper measurement and experimentation like marketing mix modeling and randomized control trials to ensure they are driving incremental lift to business outcomes.
But marketers will run these experiments on about 5% of any given budget at most. Their planning, buying and bidding will then be driven by media quality signals that are the most reliable proxies to those measurable outcomes.
So which do we think is a better aspirational North Star for a healthy industry and functioning market: the Outcomes Era or the Quality Era?
And for those who want to have it both ways, how can we ensure that the systemic attribution hacking produced by our current obsession with outcomes doesn’t get replicated?
We stand at a crossroads. Do we continue down a path that all but the most entrenched powers agree has been a failure? Or do we embark on a new path aligned to time-tested marketing fundamentals?
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow Erez Levin and AdExchanger on LinkedIn.
For more articles featuring Erez Levin, click here.
