September 13, 2023 • 3 min read • By Ronan Shields
The debate sparked by Adalytics Research’s investigations into the transparency of media buys on YouTube has seen its parent company Google vehemently strike back at its allegations.
Central to its defense has been to punch holes in the research outfit’s methodology, and in its latest gambit, Dan Taylor, Google vp, global ads updated an earlier blog post last week outlining his assertion that Adalytics’ has employed a flawed methodology.
Taylor’s update outlined how Google performed its own analysis on YouTube using the same methods purportedly employed by Adalytics.
Here, he denied the claims in the second Adalytics report, which claims to have seen widespread instances of using behaviorally targeted ads on “made-for-kids” content on YouTube, adding that “our systems are working as intended.”
Taylor also told Digiday that despite YouTube issuing refunds in the aftermath of Adalytics’ June 2023 report, which alleged problematic overlaps with Google Video Partners, now have been issued in the wake of its second report.
Digiday spoke further with the Google vp on the Adalytics’ particularly how some of the largest media buyers in the industry have given its researchers there an extended hearing, even as Google mounts a very public defense.
The following conversation has been lightly edited for brevity.
How have your conversations with advertisers been over these past few weeks, and what kind of improvements have they requested?
There’s certainly been some interest in the topic [raised in Adalytics’ studies], and we have seen some interest in understanding of how our tools work on YouTube, how we serve interest-based ads, and how we serve ads on YouTube.
There’s been some questions about ways we can improve campaign reporting and reduce the types of confusion like we saw with the Adalytics report.
We don’t have anything to announce [in terms of updates to YouTube’s buyer controls] at this moment, but we’ve long provided advertisers with extensive guidance on how to use content controls.
We’re using this as an opportunity for dialog with customers on this topic, to understand what might be more useful in the future. But it hasn’t been a major area of confusion from our buyers.
Some argue that mixing elements of behavioral data with contextual data is (effectively) behavioral advertising. What’s your response?
I would definitely disagree with that. Interest-based advertising can be served in a variety of ways, there are a variety of ways to understand where a relevant ad might be served.
You can get an understanding of a general audience that visits a website and you can extrapolate that understanding to kind of a general audience categorization of a given content provider. And when we feel like it can deliver a relevant ad, we can do so even if we aren’t using specific individual user-level data.
Advertisers cite the screengrabs in the Adalytics reports and the availability of its methodology as evidence of behaviorally targeted ads. Can you categorically say this never happens?
The whole purpose of the updated blog post and analysis we provided was after Adalytics updated its report, sharing specifics of how they came to their conclusions. This allowed us to replicate those campaign tests using a similar configuration, and when we did that, we found no instance of personalized ad serving on made-for-kids content.
So, there is no way possible for personalized ads to get through the protections on MFK content?
I can speak to the analysis we ran, and found no evidence of that happening in any way.