The biggest problem with modern day media-buying is that the ad industry has been optimizing the wrong thing: impressions.
Yes, the industry uses a variety of factors to identify the best impressions, including user IDs, profiles,“viewability,” and a number of contextual weights to ensure that brands target the most “premium” inventory.
The problem is that it’s still utilizing an “opportunity to see” model in an era of infinite choice and hyper-fragmentation, when the only thing that really matters is whether someone sees and engages with your ad.
The right metric to optimize for is actual attention. The problem is that until recently, the industry had only crude surrogates for actual human attention or methods that were too unwieldy (ie. neuromarketing) to be practically applied in real-time media-buying, especially the programmatic kind. It’s the reason why Oracle’s Moat unit has become the closest thing to a de facto currency for digital audience attention.
Moat has been a good start, using a variety of heuristics that approximate a digital user’s attention, but last week I took a brief from serial entrepreneur Marc Guldimann, who has developed an even more promising approach.
Guldimann, who founded Parsec, a programmatic media-buying platform that enables advertisers to pay for media on a “cost-per-second” basis, has quietly incubated a complementary platform that enables marketers to “arbitrage” human attention.
The platform, dubbed Adelaide, uses a proprietary method to identify digital audience impressions with the greatest propensity for attention, factoring things like viewability, duration of exposure and adjacent clutter surrounding them, and then lets advertisers or agencies bid for them programmatically.
Adelaide is a self-serve model enabling brands to identify and target the digital ad inventory with the highest propensity to deliver the attentive audiences. Think of a DMP of human attention.
Adelaide has been actively in market for a few months, but based on a series of A/B case studies, it is showing promising results, delivering participating brands CPMs that are nearly half the cost of campaigns utilizing conventional metrics like “viewability” to optimize.
“It’s kind of a no-brainer,” Guldimann says, explaining why actual attention trumps viewability.
“When you optimize for viewability you get the smallest ads on the screen,” he says, referring to the fact that most standardized viewability metrics are based on the percentage of an ads pixels in view of a user’s screen, not necessarily the ones that will stand out and deliver actual attention.
Is the base data server-side data, or is it device-side data?
Thanks John - Adelaide uses both server and client side exposure data.
Thanks Marc. So, can I safely assume that you use a process to extrapolate client-side exposure to server-side traffic.
It would be interesting to see what assumptions underlie the appliction of attentiveness factors based on this system and what ad awareness or "impact" research exists to support said assumptions.