Conventional wisdom and recent history suggest that targeting trumps context. But what if this is based on a false success metric?
One of programmatic TV’s primary strengths may well be the implied ability to buy audiences, but I wonder how well we will be able to control for creative context. Is it even necessary? The short answer is “yes.” In the digital space, almost every demand-side platform (DSP) was built around the ability to deliver specifically defined audiences. These are data-rich solutions. Gone are the days of using a site’s content or focus to draw conclusions and comparisons about an audience relative to your own and then deciding that the two are similar enough to advertise there. That’s how we currently target a lot of traditional media, from radio and TV to magazines and newspapers.
No, this is fact-based. The goal is perfect audience delivery efficiency. The trades, depending upon the universe size, usually involve the potential for massive frequency but limited control over the context in which ads will appear. Sure, in digital, you can blacklist sites and content types, but that’s based on an infinite amount of inventory. In TV, filters could narrow the available inventory pretty quickly.
The Wrong Context
So when I think of how this would work on TV, I get nervous. What if ads for Hershey’s Kisses and Carl’s Jr. ran during “The Biggest Loser”? Not so great. I think we can agree that there are situations wherein ignoring context could do a lot of damage to an advertiser’s brand.
So, what do I get in exchange for perfect audience delivery? In theory, I’d get better and deeper engagement, as well as more sales. In the digital space, you might buy this on a cost-per-thousand (CPM) basis, but for programmatic TV, it’s clear that you’d want to hold out for a cost-per-action (CPA) deal if you can get it.
But if you can’t do that, how good are the metrics that you can deliver at proxying for your real goals? My guess is that they can only help sugarcoat the already atrocious metrics that the entire industry uses as indicators of success. Targeting, in other words, needs to be able to reliably elevate performance above the noisy baseline that comprises bot traffic, unviewable impressions and the rather odd propensity of 2% of the web’s population to click on ads.
That doesn’t really excite me. What excites me is sales. I actually do believe that perfect audience delivery should result in more sales. But I’m willing to bet that it’s part of a complex attribution story – an imperfect measurement tale from our most measurable and accountable medium.
This gives me pause. When I ask for the same kind of rigor from our least measurable and accountable medium – TV – what am I going to get?
The Final Phases
I’ve been around this industry long enough to know that the cadence of the argument for programmatic TV will go something like this, in order: cost efficiency, audience delivery efficiency, brand value, increased sales. We’re about halfway through the second phase – audience delivery efficiency – right now.
The last two phases will be a dance around a measurement problem that a lot of people will pretend has been addressed by the lessons from digital, but which I think are flawed and probably not portable.
For example, I seriously doubt much TV inventory will be sold on CPA beyond what’s already sold on a per-inquiry (PI) basis. Margins on PI inventory are probably already so squeezed that I don’t think the cost of a technology overlay, coupled with the premium that would come with advanced targeting, could justify the expense. And it’s for this reason that none of the targeting bells and whistles would be brought to bear here, and rather treated as superfluous to final performance as measured by sales. I think brand campaigns will benefit from advanced targeting, but it won’t stand alone.
Programmatic TV performance comes back to sales and, therefore, advanced attribution. The crazy thing I think you’ll find in that analysis is that context matters a lot – likely more than targeting alone.