By Steve Francolla
The industry is on a path to see the open web shattered into a million walled gardens, as cross-site data centralization at scale is being challenged. Today, many premium publishers see a minority of their audience authenticate, with the vast majority on the lower end of an estimated 5% – 30%. This means that third-party data and advertiser first-party data overlaps will be a fraction of what they are today, grossly reducing the universe of people-based bid opportunities at Demand-Side Platforms (DSPs).
Looking beyond the noisy identity crisis that third-party cookie deprecation has caused, there is positive news — publisher first-party data is 100% transactable. To be clear, the data being referred to is full publisher site traffic, inclusive of the unauthenticated 70-95%, classified into anonymous first-party audiences at the inventory source.
This data is extremely valuable. It is organically recent, accurate, and privacy safe. And, it’s easily transactable — if the industry can agree on corresponding segment IDs. With that agreement, we’re looking at audiences standardized and aggregated across publishers, i.e. a foundation for an independent cohort ecosystem.
The Post-cookie World Requires Collaboration — Now
It should be simple, right? Let’s just pick a taxonomy and emit its segment IDs into the bidstream!
Not so fast…
Cookie elimination and lack of authentication at scale introduces a shift in market dynamics that publishers are aware of. Yes, publishers’ first-party data is siloed, making it hard for advertisers to buy at scale. Yes, publishers are interested in bridging this. But, publishers have recaptured a level of control over their data that they can and should protect.
In addition, if the industry opens a pipe that simply passes the Segment ID into the bidstream, the complexities start to get specific. I’ve heard questions, such as “Who gets to decide the taxonomy for the industry in Germany?” and, from the buy-side, “How will I know how users are being qualified for certain segments?”
Seven Key Issues Solved by the New IAB Tech Lab Seller-defined Audience Guidance
At the Tech Lab’s recent Addressability Roadshow webinar, I joined a panel of experts to discuss “Taxonomy and Data Transparency Standards to support seller-defined audiences” — outlined in this report of the same name. The IAB Tech Lab’s guidance in this report solves seven key issues, from trust in programmatic pipes to working together to form a unified approach.
1. Pipe: In the bid request, where do supply-side platforms (SSPs) and demand-side platforms (DSPs) agree to set and get Segment IDs?
2. Taxonomy: Who decides what taxonomy the industry adopts? This is highly subjective due to regionality and content.
3. Assembly: Who is responsible for segment construction? The publisher, the SSP, an on-page data co-op or a publisher standardization alliance?
The IAB Tech Lab guidance solves these first three issues with bidstream mechanics. In a partnership between the IAB Tech Lab and the Prebid Taxonomy Taskforce, we have developed an industry consensus for the adoption of three existing OpenRTB fields to define where a Segment ID should live, which taxonomy that segment ID is associated with, and who the provider of the data is.
We’ve also made sure that a single piece of traffic can have any number of these three-point classifications. This means that a “luxury car” intender could be defined on-page using four different assembly methods applied by, say, the publisher and three different data provider partners. With this, the buy-side can choose to trust the classification of the publisher or one of the other partners for the impression.
4. Transparency: With a thousand publishers and data providers doing assembly, where does the buy-side look to discover what a segment is composed of?
5. Trust: Over time, how does the buy-side know that a data provider’s segment assembly remains true to what is described?
For these issues, the IAB Tech Lab guidance makes use of its preexisting Data Transparency Standard (DTS), DataLabel.org, and compliance program services. These tools allow publishers and data providers to label and differentiate their data quality according to the DTS spec, make those labels accessible on DataLabel.org to inform buying decisions at scale, and provide an additional layer of buy-side assurance by submitting to a voluntary annual audit of their data label contents. With these systems in place, buyers can see and trust what they’re buying.
6. Fraud: Couldn’t a nefarious party misrepresent traffic for higher CPMs?
The answer to this is yes, it can happen. To counteract that, buyers can choose to only buy data that has been transparently labelled according to IAB Tech Lab’s Data Transparency Standard (DTS) validated via the DTS compliance program, and registered on the DataLabel.org industry repository.
7. Balance: How does a publisher participate in open exchange without cannibalizing its direct business and premium audiences?
At Permutive, we are asked this question all of the time. The answer is that this new world requires the publisher to take on the creation of a first-party audience strategy for the open exchange. There are a few things to consider to start.
Post-cookie, the only actors that can classify a publisher’s audience inventory for pass-through to the new seller-defined audience pipe are the publisher themself, those data provider partners that the publisher allows on-page, and other partners with whom the publisher chooses to provide a first-party user ID.
In the case of premium publishers, many are looking to keep it in-house and avoid exposing first-party user IDs by self-packaging their own taxonomies. They will be using their Data Management Platform (DMP) plus the new Prebid “ortb2” config to declare the resulting classifications on-page. Since the “ortb2” config supports the exact three-point classification structure specified by the IAB Tech Lab guidance, the publisher can assemble their data without sharing IDs. The data is then relayed into the bidstream by compliant SSP partners. Having these pipes in place, the publisher then focuses on packaging a quality of data that is valuable for open exchange but not competitive with their direct audience offerings. With this approach, premium publishers both protect first-party IDs and reduce the threat of a cannibalized direct business.
For publishers with a higher priority on indirect business, there will be many new paths to align with demand. For example, there will be the opportunity to align their bidstream data packaging with that of an industry alliance or to bring data providers into their environment to classify data on their behalf. As the alliance or data provider could be more attractive to one buyer or another, the publisher makes themself available to both. Declaration of data for these use-cases is also supported by the Prebid “ortb2” object.
In the new world, each publisher will need to consider the options around in-house assembly vs via data partner and around the sharing of first-party user IDs vs not. Once those things are decided, the publishers then need to think about what quality of data they are comfortable with passing to the open exchange or what third-party taxonomies they want to affiliate with to attract demand.
We Have the Blueprint — and You Can Get Started
Ultimately, with these seven key issues covered, the IAB Tech Lab’s Seller-defined Audience guidance gets the industry aligned for programmatic to scale safely and with publisher control in mind.
To get started, this proposal is something that Prebid and its Taxonomy Taskforce, in partnership with IAB Tech Lab, are deploying in a proof of concept with end-to-end programmatic participants. And, delivering the right approach to scaled publisher first-party data strategies for the open web is something that edge-based data platforms are uniquely positioned to help with.
The blueprint is there, the pipe is being opened — don’t hesitate to get involved in testing and finalizing the Seller-defined Audience proposal today.
Steve Francolla
Head of Global Partnerships
Permutive