Seller-defined audiences: An outcome of – and a fuel for – innovation

In the past few years regulation and consumer demand for better user experiences have focused largely on privacy. This movement yielded significant innovation in addressability as publishers aim to take back control of their data. In order to solve ‘signal loss’ challenges, there’s a clear and present opportunity for publishers to create and showcase their powerful audience signals in compliant and future-proofed ways to drive advertiser performance.

Seller defined audiences – a natural innovation

Magnite and others set the tone back in 2021, leading a consortium to test and iterate providing publisher-driven first party segments to buyers. Seller defined audiences’ (SDAs) was one of the outcomes of the IAB Tech Lab’s global Project Rearc initiative. As we begin to scale SDAs they will provide further fuel for innovation to address signal loss.

Seller-defined audiences (SDAs) provide a methodology for publishers to communicate first party audience attributes in an OpenRTB bid without revealing user identities. First party audience signals are mapped to taxonomies and curated into audience cohorts and those anonymized audience IDs are relayed via OpenRTB to downstream providers through existing Prebid header bidding integrations. This enables buyers to bid against those labels through DSPs. The industry currently lacks urgency to change buying practices since the third party cookie world still exists.  However, we fully expect that as third-party cookies go away buyers will turn to publishers for addressable audiences, and SDAs will anchor the solution.

Build it and they will come

The process for building SDAs is similar to building audiences for direct audience transactions – SDAs are just the privacy compliant vehicle for communication of those first party signals on the open market. As a result, accurate audience categorization and segmentation will become increasingly critical in attracting demand when third party cookies disappear. That’s why building these tools now – or working with those that can provide them – will help future-proof audience strategy.

Categorization into a taxonomy

Firstly, publishers need to extract all meaningful content themes, keywords and behaviors from a page visit into a taxonomy to relate those signals to user interests, brands and keywords. This can be done using automated categorization tools that leverage natural language processing. The output is mapped into the IAB Tech Lab Audience Taxonomy of 1,600 demographic-, interest-, and purchase-based attributes, and/or the IAB Tech Lab Content Taxonomy for contextual-based audiences. Based on recent discussions with a cross section of publishers, Magnite estimates that over 80% of publishers already rely on the IAB taxonomies to some extent. However, for publishers that want to leverage their even more unique (and high value) audiences they can map signals to their own taxonomy terms for selling custom audiences via direct deals or PMPs, and in SDAs with IAB Tech Lab’s approval.

Furthermore, the contextual intelligence that comes from improved categorization provides both a complementary and compliant alternative to behavioral targeting. Those same contextual insights can be combined and enriched with behavioral insights and custom event signals such as time-on-page, number of pages visited, and custom taxonomy terms to uncover further unique audiences for selling via however publishers choose.

Audience curation and segmentation

Once data is categorized, it is then ready to segment audiences using automated and/or manual methods. For automated methods this includes machine learning-driven segmentation techniques that build audiences as the signals come in, based on either preset rules determined by the publisher or where the AI sees high engagement in specific areas. For manual methods this means having a comprehensive audience builder where publishers can test and iterate combinations of audience variables, whilst also providing specific audience makeups for direct clients.

Fuel for innovation

In the face of potential ‘signal loss’ publishers are collecting valuable first party data to create controllable and compliant audience addressability, leveraging such audiences to power their direct and PMPs business. SDAs have opened another path for publishers to deliver against their highly valuable first party data, ultimately replacing how buyers previously bought these audiences via third party cookies with a privacy compliant and publisher-controlled method of audience addressability. As a result, by improving the ways we categorize, segment and curate audiences we can fuel improvements in both direct and open sides of our ad monetization, encouraging both standardization and differentiation in audience strategy.

How to get started

Publishers need supply-side partners that can provide the tools and controls to compliantly categorize, curate and activate their audience data now and in the future as the market continues to evolve. Automated categorization provides the basis for audience analytics, contextual intelligence, and more intelligent audience curation and segmentation. Leveraging those insights with machine-driven audience segmentation tools can further unlock audience value. Publishers looking to unlock data value engage SSPs like Magnite to provide comprehensive audience building tools tightly coupled with demand activation. With a single integrated partner, publishers can execute an end-to-end proposition for audience and monetization strategy across all channels.