Does overreliance on programmatic hide a more significant problem?
Programmatic is the default method for placing media. Search and Social, too, are set programmatically. Estimates are that 80% or more of CTV and OTT advertising is placed programmatically. Media today is exchanged through DSPs, SSPs, DMPs, and additional letters, making up the alphabet soup of the modern media ecosystem.
Last week, I argued that programmatic advertising as it’s conducted has moved advertising away from the audiences with which it should be concerned and towards the inventory that is being transacted, that is, impressions. I suggested that programmatic advertising has shifted focus from reaching audiences and moving those people to take action to finding countable phenomena and associating them with assumed outcomes. The automated, set-it-and-forget-it nature of programmatic advertising has shifted the focus away from media planners' strengths of figuring out how to reach audiences where they are and, by that context, inspiring them to take action and put the focus on what machines do best, which is making transactions by associating ad serves with assumed outcomes such as clicks on ads.
Over the years, agencies have directed clients' digital ad spending through the holding companies’ trading desks. This let the holding company benefit from taxing the inventory on top of the ad tech tax paid to each platform along the daisy acquisition chain. This creates an incentive to maximize programmatic use while still holding on to FTEs against which the compensation models for agencies are often structured. It also creates a disincentive to solving some of Programmatic’s other seemingly intractable problems.
But more significantly, does so much reliance on programmatic hide a more profound challenge of online advertising in general? Does all this dependence on multiple data layers and computing power to spend media dollars more facilely, or the frequent attempts to address the problems that the various layers of technology used to support the system lend themselves to (fraud, transparency, etc., see the most recent effort of the ANA to address here) suggest an inherent weakness in online advertising relative to other media, particularly as it relates to display advertising? It could be that online display advertising suffers from a far more profound existential crisis than robot armies replacing humans because robots move faster. Some digerati might say that linear media benefits from legacy culture and systems regarding how much is still spent despite the dearth of heavy technology it initially had supporting it compared to online. Others will say that linear’s now regular loss of audience and spending should accelerate, and the only reason it hasn’t is because too many marketers are stuck in their ways. However, to paint the entire industry segment still confident in linear’s usefulness with the “stupid” brush seems more like a defensive reflex than a considered response.
It could be that online advertising is not intrinsically good at doing what we are asking it to do in the way we have been asking digital advertising to do it. The truth is, Display, as it evolved in the programmatic era, hasn’t been good at Advertising, capital ‘A.’ Because its utility was primarily to ease the movement of dollars from an advertiser to a media owner, with the facilitating platforms taking a cut coming and going, we asked it to reach more relevant audiences that would be more productive for the advertiser but rarely, if ever, considered the audiences. We wanted the numbers to meet expectations and the technology to work as promised, but we stopped thinking about what any of that meant to achieve tangible business goals. We designed the system to do one thing – easily handle high-volume transactions – only to later retrofit the idea of reaching the people necessary – mostly anemic segmentation data -- to meet an objective that went beyond “BUDGET divided by CPM = IMPRESSIONS.” We got further away from the audience with every layer of tech we injected into the system. Again, we asked it to do something it wasn’t designed for. And no matter how many turkeys you throw out of the helicopter and watch plummet to the earth, at some point, you must admit that turkeys don't fly. Putting little parachutes on or strapping them to hang gliders only obfuscates their innate non-fly-ability. As CTV and OTT grow, their power and effectiveness will eventually reflect what linear used to yield, so long as those channels can solve the problems plaguing Display.
A way to side-step the historical weaknesses of programmatic is to use some of the linear fundamentals for CTV and OTT channels. This means looking NOT at how well the system says a buy meets the numerical goal of impression delivery against targeting criteria, but how well the media is planned against REACHING the target. Sure, the data for doing this in the previous era became more brittle as the number of media points of contact multiplied and the number of people engaging each point specifically got smaller. However, the fundamentals for predetermining audience and impact are still sound. If you don’t REACH an audience, it doesn’t matter how specific the targeting criteria you give a system to satisfy.
If the system works properly, shouldn’t you reach that audience BECAUSE of the targeting criteria? Next time, I’ll address some pervasive myths about targeting and why more targeting often does not yield better outcomes.
And for those of you not old enough to remember... "As god as my witness, I thought turkeys could fly." http://www.youtube.com/watch?v=lf3mgmEdfwg
