The most stunning comeback in recent times in media has been the reemergence of walled gardens. Only a few years ago, many pundits attributed AOL’s early 2000s flameout to its insistence on creating a Disneyfied version of the Internet while the masses yearned to be wild and free.
The past five years has offered a refutation of that idea, as Facebook, in mimicking AOL’s walled garden approach, carved out a huge and profitable chunk of the Internet for itself.
In so doing, Facebook has created a proprietary identity graph—a massive database which lets advertisers design campaigns based on detailed information about each individual user. Advertisers love it, especially on mobile.
But Facebook’s isn’t the only proprietary identity graph out there. There are identity solutions from a dozen or more different vendors, but therein lies that risk that to truly create your own comprehensive graph, you’ll have to cobble something together from disparate sources. A more workable answer is for vendors to provide interoperable solutions.
The balkanization of ID graphs
An identity graph tries to tie together signals from different devices to a single, universal ID, so that users can be recognized as “the same person” across the many devices and platforms that we consumers use every day. With a universal ID in place, all of the data collected about a user can be tied back to one place. Typically identity graphs are anonymized through hashing and encryption to provide security and privacy protection. Since consumers these days flit back and forth between three or four devices every day, this is a tall order.
If you run your own social media network, like Facebook, then you have a great way of normalizing to a universal ID, collecting all of this data, and putting it in one place. But in recent years, some companies have been able to do something similar on the open web.
ID graphs from Oracle, GroupM, the Digitrust push for a universal ID and the AppNexus/MediaMath/LiveRamp consortium, for instance, mimic Facebook’s ID graph, with some success. While each of these identity solutions represents itself as a more “open” alternative to the walled gardens, each operates on the basis of a proprietary approach. As these consortiums grow larger, their boundaries will begin to overlap. Over time, the ongoing proliferation of new “universal” IDs from each identity solution will resurface some of the very problems they are designed to solve. If there are too many IDs to authenticate, operating between consortiums will quickly become unsustainable for marketers.
So far there are promising signs moving against the threat of balkanization, with high-profile platforms like The Trade Desk moving to standardize their ID to operate with the Advertising ID Consortium. But there still remains a bright line between the growth of that consortium and the efforts underway with Digitrust. A truly universal ID remains elusive.
The best solution for marketers
One thing marketers don’t need is another walled garden. A great solution for the market right now would be to create interoperability between these disparate universal IDs. Portability would also be helpful for marketers that are keen on working off their own homegrown ID graphs instead of relying on others.
Let’s be honest, Oracle and GroupM aren’t competing with each other as much as they are competing with Facebook. If marketers are frustrated by these alternative solutions, then Facebook wins out.
Stepping back a bit, what this is is a big test of game theory. Will adtech players band together to offer an alternative to Facebook, or will they continue to scrap with each other, strengthening Facebook’s hand? I’m rooting for the industry to wake up, but that doesn’t mean it’s going to happen.
Originally published 05/16/18 in MarTech Advisor