Often enough, when the topic of data leakage comes up, someone—whether they’re coming from a buy-side or a sell-side point of view—will say: “Hold on. Is this data leakage stuff real, or is a bogeyman?”
Understandable response. This is digital media: Wherever there’s just the perception of a problem, you’ll find someone trying to sell you something to fix the problem. But people in the know will tell you data leakage is real. And if you need someone to spell out how it works, Digiday’s article from last week “Confessions of a Programmatic Ad Buyer” does that. The agency buyer in question explained: Instead of going directly to publishers for audience data, they’ll buy some of the publisher’s impressions on the open market, then buy some third-party data and layer it in to build their own audience segments. They’re getting the data they want for cheap, without paying the publisher anything extra for it.
Sounds frighteningly easy, but there are some downsides to that method. There’s plenty of third-party data available from vendors, but not all of those sources are of the same quality. If you’re able to buy this stuff for cheap, there’s probably a reason for it. So the agency is left with piles of third-party data, plus some data they’ve pinched from publishers—but the publisher isn’t in the room to help advise or vouch for the insights in that data. The agency is building out these lookalike models, testing them, and running campaigns when the tests perform well.
It’s obviously possible to operate that way, but it doesn’t sound very efficient. The buyer in Digiday’s column understands that. They admitted third-party data is error-prone, and that they would willingly pay publishers for data instead. There’s a better way for buy-side folks to get pub data—but pubs aren’t taking the right precautions to push buyers down the right path. Pubs aren’t hiring specialists to manage their data, they’re relying on vendors for security point solutions, and they’re not creating the infrastructure where they can conduct secure data deals.
Let’s Get Legit
All of those points about the “right way” to do things—it’s interesting to hear an agency buyer say it, because Scott Messer from Leaf Group said exactly the same thing at the Miami PubForum last year. Scott’s session focused on how if publishers are afraid of marketers stealing their data, they can and should get in front of the ball, provide the framework for legit data deals to happen, and act as a real partner.
As a publisher, you know your audience better than anyone else. You can save marketers and agencies a lot of guesswork by being involved in a conversation about what’s in your data—and you can put a suitable price tag on that partnership. Buy-siders can build lookalike models—but what’s better, having a user that seems to be the target user, or having a user you know is the target? As Scott had said in his session: Accuracy matters, and if you have a login to tie these data points to, you’ll have a much easier time being accurate.
Putting yourself forward and offering an above-ground option is one way of cutting off data thieves at the pass. The other is to clamp down on security. Ad tech companies are awash in audience data as a result of header bidding. If you don’t want interlopers pinching your data, you need to have conversations with your tech partners about what they’re doing to prevent data leakage at the point of an inventory transaction, or anytime.
It’s easy to take a fatalistic attitude toward data leakage, but the problem itself isn’t really a force of nature. If marketers and agencies are cutting around publishers to lift their data, that means they want the data. Pubs have to ramp up security efforts to make it harder to cut them out—and then they should take control of the discussion. Second-party data is a hot market right now. Data can be a veritable additional revenue stream for pubs, if they can make sure they’re the number-one source for that data.