Seller-Defined Audiences: Explained

Tim Cross 31 March, 2022 

Last month IAB Tech Lab rolled out ‘Seller-Defined Audiences’, a specification created within Project Rearc, Tech Lab’s incubator for solutions to replace third-party cookies.

As the first addressability solution to come out of Rearc, there’s a lot of weight on Seller-Defined Audiences’ shoulders. While it doesn’t act as a wholesale replacement for third-party cookies, Tech Lab hope it will be a useful tool for enabling first-party data based targeting on the open web.

A lot of the concepts introduced within SDA aren’t new – in fact Tech Lab says it’s built on existing, widely adopted industry specifications. And it bears more than a passing resemblance to APIs proposed within Google’s Privacy Sandbox.

But SDA is significant in offering an alternative which operates across different browsers and devices, not just Google’s Chrome browser and Android mobile OS. And the fact that it’s out of Google’s hands may alleviate some of the antitrust concerns associated with the Sandbox.

The Basics

The basic idea of Seller-Defined Audiences is that it creates a standardised way for publishers or data companies to create targetable audiences which buyers can easily understand and bid against.

Publishers defining and selling audiences isn’t a new concept, but usually this is done through proprietary tools or private deals. Sending first-party data through the bidstream carries privacy risks.

But SDA is designed to allow publishers and data owners to group audiences into targetable cohorts – users who share similar characteristics. When an individual is shown an ad, information about which cohorts, if any, they’re part of can be sent in the bidstream. And buyers can use this information to decide if they want to bid for that impression, and if so, how much.

Since it uses first-party data, each individual publishers builds these cohorts based on the behaviour they observe – meaning the same individuals will sit in different cohorts on different websites. But the cohorts are standardised, allowing buyers to target the same cohorts across different publishers.

The Technical Detail

As mentioned, SDA leans on a number of established specifications in order to function. Three of the most important are IAB Tech Lab Audience Taxonomy, IAB Tech Lab Data Transparency Standard, and IAB Tech Lab’s Transparency Center.

The put it simply, the Audience Taxonomy is a standardised way of categorising audiences, based on known information about them. IAB has over 1,600 different attributes within this taxonomy, which when combined can be used to build cohorts of similar users.

So under the SDA specification, publishers use any first-party data available to assign these attributes to individual users, and then place them into cohorts. Once these cohorts are big enough (undefined as yet, but big enough they data can be considered anonymised), they can be sent in the bidstream using existing objects within OpenRTB.

The quality of this data is signalled through the Data Transparency Standard, which labels how and when the data was collected. Since this information is self-attested, Tech Lab has developed a separate compliance programme, which verifies which sellers are accurately labelling their data, and therefore trustworthy in their attestations.

The Transparency Center meanwhile supports the Transparency Standard by allowing these data labels to be uploaded to a centralised resource, which Tech Lab members can scrutinise before making buying decisions.

Thus, buyers are able to see which cohorts an individual is part of, and the quality of the data which informed that cohort, without specific data being transmitted through programmatic pipes.

The Pros and Cons

IAB Tech Labs believes its new specification has a lot going for it.

Aside from the obvious benefits of enabling interest-based targeting in a privacy-safe way, Tech Lab says SDA stands out by enabling a standardised way of labelling and buying against first-party data cohorts across different browsers and devices.

The fact that it uses existing specifications should make adoption relatively straightforward. And Tech Labs says that it should make for a fairer ecosystem, where quality of data informs buying, and publishers with quality data are rewarded.

One possible downside though is that not all quality publishers will be equally able to create useful cohorts based on user activity. Publishers with narrow content themes for example may find it hard to pick out useful patterns in browsing behaviour within their properties.

And while it’s an interesting prospect, many publishers have already forged ahead with their own in-house tools which rely on signed-in user data. SDA may compliment these tools for these publishers, but are likely to play second fiddle to their own buying platforms, meaning in many cases buyers won’t necessarily opt to target against these cohorts anyway.

Follow VideoWeek on Twitter and LinkedIn.


About the Author:

Tim Cross is Assistant Editor at VideoWeek.
Go to Top