It's The DMP, Stupid
If you want to understand how programmatic audience-buying is transforming the way television advertising is bought and sold, do not look to the litany of acronyms cluttering the advertising technology landscape -- focus on just one -- the DMP, because, you know, it’s the data, stupid. It always was. But increasingly, it will be the platform people use to manage audience data that will determine who has the leverage in the TV advertising marketplace. Based on some recent trends, I’d say it’s beginning to shift to the buy-side. And rightfully so, because for the last 75 years or so, TV audience data has essentially been controlled by the sell-side. That’s finally beginning to change, thanks to new forms of data for targeting and buying television audiences, and new platforms for managing them.
And here’s why. Under the historic TV audience data model -- the one controlled by Nielsen -- TV market currency attributes were pretty simple -- the age, sex, income and a few assorted otehr demographic descriptors that have historically comprised TV ratings. There wasn’t a lot you could spin with those, and it was pretty easy for any sell-side player to understand exactly what inventory they wanted from the buy-side. But by shifting away from conventional demographics, and moving more toward first-party data (information advertisers have directly from consumers) and new forms of third-party data (data from other entities that can better describe the behaviors and attributes advertisers are trying to influence or target), marketers and agencies are finally beginning to create their own black boxes -- proprietary systems that enable them to target TV audiences based on their own secret sauce, which the TV supply-side may never be able to fathom.
That’s a huge shift in marketplace leverage, if you ask me, and a much bigger one than suppliers -- especially premium ones -- normally fret about: The shift toward programmatic exchanges. They worry that putting their inventory into machine-powered exchanges will “commoditize” the value of their inventory, making it indistinguishable from other “me-too” options, and driving their prices down, not up. I think they’ve got that completely wrong, and if they learn how to manage the front-end better -- the data marketers and agencies use to determine what audiences they covet most -- the value of their inventory will go up, not down, assuming that what they always say about it is actually true: that their audiences are more valuable, especially when they are watching highly engaged, network TV-quality programming.
Premiums -- whether they are audiences or the media inventory that are proxies for them -- are funny things, and often in the eye of the beholder, or as the case may be, the bidder. I was thinking about this just yesterday when I sat in on special meeting of a top media sales organization updating their team on programmatic developments. The sales chief asked me how programmatic trading would impact “premium” inventory, so I asked the chief, “define premium.”
The sales executive said, from their point-of-view, it was the price advertisers were willing to pay. I said that was an “interesting” point-of-view, because from an advertiser’s or an agency’s point-of-view, premium was the value of the audience they were reaching, and it often has nothing to do with the price they are paying. That last part is true, because there are all sorts of disequilibrium occurring in the current programmatic marketplace that are distorting the relationship between underlying value of audience and the price a marketer might be willing to pay to serve their ads to them, and they have absolutely nothing to do with rate cards. They have everything to do with data. And as soon as big sales organizations understand that, they will stop fretting over exchanges, and start focusing on where the real action is: data, and data management platforms.
Hey, don’t take my word for it. Ask the folks at Interpublic’s Mediabrands why they’ve invested so much in their exclusive deal with Nielsen, and now Adap.TV, to build a proprietary DMP for targeting TV (and yes, online video too) viewers based not on what any suppliers say about them, or even simply what Nielsen’s TV ratings attribute to them, but on what they mean to Interpublic’s clients brands as current or potential consumers.
Interpublic is not alone, they’ve just been a little more vocal than others. I’m willing to wager that every one of the major agencies is developing their own proprietary DMP, whether they call it that, or even use another acronym to describe it. Some are developing it through their trading desks, as I believe GroupM is doing via Xaxis TV. Yeah, Xaxis TV isn’t actually procuring TV audiences -- yet. It’s utilizing TV metrics to effectively re-target TV-like audiences online. That’s something another innovative DMP, Collective, has begun doing too.
And let’s not forget some other savvy TV native DMPs that are beginning to emerge, especially Dave Morgan’s Simulmedia, Visible World’s AudienceXpress and, of course, Jon Mandel’s PrecisionDemand. They all have slightly different approaches in terms of how they’re cracking and harnessing so-called “big data,” but they’re all about the same thing: correlating the data that sends a valuable, meaningful signal to a brand about a consumer, and then tying it to ways they can actually buy television inventory.
ask me, this is the most exciting -- and disruptive -- development to hit the TV advertising marketplace during the 30 years I’ve covered it, and the amazing part is that few people on the
sell-side actually get that. They’re all worked up over exchanges, and the fact that machine logic may drive the value of their inventory down. When in truth, it will be the logic of other human
beings -- advertisers and media buyers -- using more sophisticated data to give them better intelligence, and a competitive advantage, in understanding what inventory and audiences have the most
value, regardless of how sellers are pricing it.
The "It's The Economy, Stupid" image provided courtesy of Shutterstock.