Seller-defined audiences (SDAs) are meant to be a blessing for publishers as the importance of privacy increases. They allow publishers, data management platforms (DMPs) and data providers to scale first-party data within an OpenRTB bid request without tracking a single user across sites.
But what about buyers? Are publishers’ SDAs really giving them what they need to reach their target audiences?
Most likely not, because a significant portion of SDAs fail to include signals that are required by the IAB Tech Lab.
Noncompliant SDA signals
Signals, both contextual and user-based, are at the center of SDAs. Contextual SDA signals assign topics that reflect the website’s content, while user signals assign interests to users based on their historical behavior on publishers’ domains. A single bid request can contain multiple SDA signals.
According to the IAB Tech Lab, a compliant SDA signal contains the name of the entity that produced the signal for the publisher, specifies the topic/segment of the content’s focus or user’s interest and refers to the taxonomy that decodes these segments.
Despite these clear specifications, RTB House found that 33% of contextual SDA signals and 75% of user SDA signals fail to meet the necessary requirements.
While all noncompliant contextual SDA signals had their provider listed, on 99% of occasions they failed to specify the taxonomy. And one quarter of instances didn’t identify the segment.
User SDA signals, meanwhile, had problems across the board. Almost half (48%) of noncompliant signals failed to include provider information, while 89% were missing segment information, and 91% didn’t specify the taxonomy.
It’s worth noticing that, conceptually, SDA narrows the use cases it can be applied to, as it doesn’t have mechanisms for retargeting and cross-site frequency capping. Thus, the absence of aforementioned data points hurts the overall usability even further.
In addition, SDAs don’t have mechanisms to prevent fingerprinting. Based on RTB House’s research, roughly 85% of user SDA signals observed in Q4 2022 were accompanied by other user identification (such as cookies, external IDs or mobile identifiers). This violates specifications and risks bad actors using these signals to enrich cross-site graphs with behavioral data through fingerprinting.
Why should buyers trust noncompliant signals?
The SDA solution may not be a silver bullet, but it has garnered a fair amount of attention, particularly among publishers. Publishers enjoy controlling their inventory and audience labeling. It enables them to independently decide what their articles are about, assign users based on their behavior and use either standardized taxonomies or their own.
There’s less enthusiasm, however, from buyers. Buyers are calling for a better way to verify publishers’ reliability in creating the signals. SDA specification is, after all, merely a set of recommendations for publishers on how they should pass along information on their own inventories and audiences. And buyers find the IAB Tech Lab’s annual certification insufficient as a continuous proof of quality. They shouldn’t have to base their bidding decisions on trust and one annual check.
Whether SDA lives up to the hype all depends on IAB Tech Lab’s desire to address the solution’s issues and if buyer adoption increases. The scale is noticeable but low.
Until more buyers and publishers opt in, all that publishers can do is make sure their SDA signals are compliant and reliable. Right now, publishers are falling short of that mark.
“The Sell Sider” is a column written by the sell side of the digital media community.
Follow RTB House and AdExchanger on LinkedIn.
