We’re nearing the final stages of grief and getting around to accepting the third-party cookie’s imminent demise. Now publishers can focus on the good stuff: the opportunity this moment presents for building trust throughout the advertising ecosystem and especially with consumers.
As we inch closer toward a cookieless world, identity solutions have been touted as the perfect panacea. But before going all in on one identity solution over another, wouldn’t it be prudent to weigh all of your options?
You wouldn’t invest all your money in one stock, would you? So why would you throw all of your money at a single third-party cookie alternative and risk your full ability for targeting audiences in the future? Wouldn’t it make sense to invest in a diverse set of identity solutions to see which methods work best for your company?
The range of possibilities runs the gamut. Let’s take a look.
First-Party Data Alone Is Myopic
Advertisers are already on a quest to find new quality data sources. While they’ve long claimed to have an appetite for publishers’ first-party data, this hunger often hasn’t appeared in reality. However, publishers can take this moment to prove their first-party data delivers a real return on investment.
On first take, a first-party data strategy as an alternative to the third-party tracking cookie sounds ideal. But not every publisher has the scale that major advertisers want and need, thus why many publishers have turned to third-party data for targeting for so long.
The truth is, there’s only a handful of large publishers who have the kind of first-party relationships with their audiences that advertisers find desirable—publishers like The New York Times, which boldly announced in May 2020 that it was phasing out the use of third-party data in targeting ads.
These major publishers might be able to drive healthy revenue without cross-site user tracking tools, but many midsize and small publishers are not going to thrive on their first-party data programs alone.
Besides, a publisher’s first-party data is really just a sliver of a users’ profile—the experience and interests of that consumer while on that particular publisher. Advertisers know other solutions provide more robust profiles of their target audiences that fill in interests and behaviors outside a single realm.
When it comes to first-party data, context alone won’t allow advertisers to measure across the entire customer journey, across varying devices, domains and different platforms. Buyers want to understand intent; which messages resonate with specific audiences and which channels are most effective.
So, what good is a first-party data strategy without scale or uniformity across pubs?
The Rise of Data Clean Rooms
The idea of data clean rooms is emerging as a privacy-safe option for publishers, advertisers, and walled gardens to bring together their anonymized data into one secure platform in hopes of solving cross-media attribution by matching customer data with campaign data.
In a clean room, you get detailed advertising impression data, but there are restrictions on what user-level data is output. Basically, what you’ll get is an aggregated look at where your data intersects with others, without direct access to the pooled data.
As things currently stand, Amazon, Google and Facebook still corner the market on identity and as you’re probably guessing, they’re the primary players in the data clean room game. A clean room will allow you to compare data sets to get a sense of who saw your ads and converted, while also looking at inconsistencies to see if ads are being over-served to similar audiences.
Fortunately, it won’t only be walled gardens that can provide this type of service. Procter & Gamble and Unilever are already working on offering their own solutions, but essentially any pub or brand with a large owned audience would be able to provide these services.
Data clean rooms come with their own set of challenges. While you’re given the ability to see the data and insights derived from the data, you won’t be able to remodel it or determine how findings were calculated. You’d be giving up your control of the data to gain insights derived from only a moment in time—and also losing out on multi-channel and multi-touch attribution at scale, across channels.
There are also technical and organizational challenges when it comes to data clean rooms. First, you have to guarantee that your data is secure and then also determine which data goes into the clean room and who will have access to it.
It’s going to be important for the industry to work together—buyers and sellers—to set up use cases instead of allowing the major players to decide the appropriate ones for the entire industry. In the short term, it will require lots of work and cooperation for data clean rooms to be set up to appropriately meet the needs of the advertising ecosystem
What About Universal IDs?
Much of the conversation around the sunsetting of third-party cookies has been focused on what its replacement will be. To be honest, there’s no true replacement for that method of tracking, though identity targeting offers the closest analog to the third-party cookie—but is arguably much more durable and privacy-friendly, with easy access for consumer control.
Many universal IDs are based on deterministic data, proponents believe them to be more reliable than third-party cookies. Absent also are the syncing issues that cause latency in programmatic. In addition, universal ID syncs will also improve match rates in server-side header integrations.
But most important, identity-resolution-powered universal IDs provide a secured, shared ID across publishers and the rest of the supply chain—and one that can be leveraged for accurate targeting, frequency capping, and campaign measurement. The faith placed in third-party cookies for these functions can be replaced with confidence in universal IDs.
In effect, these IDs become a common transport for transferring data across the open web. At the same time, every publisher and advertiser can customize their user profiles with proprietary data—while the IDs are universal, they mean something different to each ID holder.
But wait—why are there… multiple universal IDs instead of just one? Multiple universal IDs sounds like a misnomer, right?
Well, not all universal IDs are built the same. Every identity resolution provider has its own special sauce in assembling ID graphs, and often ID providers partner with each other to expand their bases. Many universal IDs have deterministic data at their core (typically hashed emails) for building profiles; and, some then layer in third-party data (online and offline) to form more robust ID graphs.
However, relying on deterministic data alone is limiting—particularly considering that authenticated or identified traffic is only a small percentage of the traffic making its way across the open web. And what about publishers with small amounts of authenticated traffic—or none at all? Are they doomed in a cookieless web where advertisers demand audience targeting and measurement?
That’s why newer entrants to the universal ID space are using “hybrid models” that augment deterministic data with machine learning. This greatly expands the amount of active profiles and makes more of the open web identifiable—especially as ID providers partner with each other.
Still, it makes most sense for a publisher to integrate a variety of universal IDs to ensure maximum coverage. And within this portfolio of IDs, make sure you have a mix of straight deterministic IDs and hybrid IDs leveraging machine learning to augment portfolios, and identify more of the open web.
As more pubs, advertisers, agencies, as well as SSPs and DSPs adopt universal IDs, the open web will have a better chance at leveling the playing field with the walled gardens, and rivaling their reach and measurement capabilities. Universal IDs could potentially help the industry reach the long fabled right-ad-right-time-right-person concept, which would be better for both buyers and sellers.
Portfolio Strategy For the Win
Given that the alternatives to the third-party cookie all have benefits and shortcomings, the only sane publisher’s approach is to deploy a portfolio strategy—one that mixes first-party data, authentication, and universal IDs (straight deterministic IDs as well as hybrid IDs leveraging machine learning) to supply reach and frequency across campaigns. This method would offer richer profiles for advertisers, but pubs would also have the flexibility to adjust to advertiser needs.
While large premium publishers can lean heavily on first-party data marketplaces, small to midsize publishers could use some support from identity solutions, which can work in tandem with first-party data segmenting to prove the power of publisher data-crunching. Segment performance can easily be gauged leveraging data common to the advertiser via identity. As audience data becomes more of an expected commodity in digital advertising, first-party data marketplaces will truly serve as publisher differentiators and justify premium CPMs.
The World Federation of Advertisers released a cross-media measurement framework that looks a lot like a clean room proposal. You can also view the Privacy Sandbox proposals as something akin to a clean room, with the browser controlling the data outputs.
It seems the buy side—and certain platforms—are betting on clean-room-style technology, but again this is really limited. We’ve heard time and again from publishers using clean rooms and second-party data that scale was a challenge and often they bolstered campaigns driven by second-party… with third-party data-based lookalike modeling.
But publishers can’t just lean on a single identity solution or universal ID—each one’s unique approach (and various ecosystem connections) offers different advantages. For sure, publishers need to ensure they have both straight deterministic IDs as well as hybrid models employing machine learning in their portfolio to maximum reach and coverage.
As we’ve noted before, there’s a lot of pressure on publishers to lead the way forward as the cookie crumbles. And this is a challenge, because instead of a convenient tool like the cookie that can easily be passed among parties, success in the post-cookie ad world requires collaboration and implementing a handful of solutions that perform in different ways.
A portfolio approach is the only way to go, where publishers leverage universal identity solutions to sync their audience profiles with those of advertisers’; clean rooms particularly to ensure measurement for buyers; and first-party data marketplaces to offer differentiation and additional targeting.