Google is continuing to dispute claims that it served personalized ads to children’s videos on YouTube, an allegation that surfaced last month in a report from Adalytics, an independent ad auditing firm. Now, Google has released a counter-study about ads on YouTube, reverse engineering Adalytics’ process to test the ad platform.
“We verified that no personalized advertising ran on any made-for-kids content,” said Dan Taylor, VP of global ads at Google.
Read more: YouTube kids’ ad placement issue raises brand doubts
For weeks, Google has been on the defensive over Adalytics’ reports this summer, which claimed video ads were showing up in improper settings, including on made-for-kids content on YouTube. The most recent report prompted Fairplay, a children’s advertising watchdog group, to ask the U.S. Federal Trade Commission to probe whether YouTube ads used data in ways that would violate the Children’s Online Privacy Protection Act. Google has been adamant that the ad campaigns are stripped of any personal data that would breach privacy rules when placed on made-for-kids videos.
“This is no different than serving a laundry detergent commercial on a Saturday morning cartoon,” Taylor told Ad Age in a recent interview.
Taylor also said Google met with Fairplay to try to explain its platform. Fairplay executive director Josh Golin told Ad Age the group came away from that meeting unconvinced and still calling for an FTC investigation.
Taylor, who has been Google’s public face during the Adalytics debate, has disputed the accuracy of the reports, calling them flawed. Adalytics had used Google’s ad platform to set up personalized targeting parameters and then looked for ads appearing on made-for-kids channels. To refute the Adalytics report, Google tested six ad campaigns with similar audience targeting settings to the ones Adalytics tried. Google said that even when an advertiser applied personalized audience targeting, such as segments of boating and motorcycling enthusiasts, the ad platform only used non-personalized, contextual signals when ads from those campaigns went to made-for-kids videos.
Google’s report also noted there is confusion over channels labeled as “made for kids,” but those channels can also have videos that are not strictly for kids. For instance, CVS 3D Rhymes and Kids Songs is a channel with a back catalog of videos for all ages, Google said in its report.
“The channels that Adalytics targeted included videos designated by the creator as made for kids as well as videos designated by the creator as ‘not made for kids’ (non-MFK videos),” Google said. “As a result, the campaign’s ads correctly served on both MFK and non-MFK videos within those channels.”
In an interview with Ad Age, Taylor discussed Google’s stance on the ads debate and offered advice to advertisers on how to navigate the situation. The interview has been lightly edited for clarity.
How did Google test its ad platform?
We ran experiments to kind of replicate what we learned from the Adalytics approach to their so-called analysis there. We verified that no personalized advertising ran on any made for kids content when we replicated the campaigns.
What is the confusion here, then?
YouTube made-for-kids channels have a mix of made-for-kids videos and videos that are not made for kids; they’re not exclusively made for kids. So therefor, you can have a YouTube channel that’s made-for-kids that has a video not geared for kids that we can serve personalized ads on.
And from an affinity and audience targeting perspective, the signals that we utilize in order to serve audience-based ads, can be a combination of either specific viewer data or contextual data. In the case of made-for-kids videos, because we don’t use personalized data to serve ads on made-for-kids videos, that included contextual data on those videos.
What do you say to critics who think that no ads should run on kids content? For instance, an ad for a credit card running on a nursery rhymes video.
We’re always welcoming feedback and actively exploring ways to improve the clarity of our policies and our reporting tools, and how we think about those approaches. The way I see it is, from an advertising perspective, this is no different than serving a laundry detergent commercial on a Saturday morning cartoon. And so the ads aren’t personalized to an individual they’re shown in a contextual fashion. Anyone who watches that video would see roughly the same kind of ad contextually targeted.
What are you telling advertisers and ad agencies when they reach out about this?
I haven’t met with too many advertisers and agencies directly over the last couple of weeks, but I do engage really closely with our teams that are, and the general sentiment is that advertising and media practitioners that are working with YouTube, they understand how our tools work, they understand how our targeting solutions work, and we’ve had very productive conversations that I would consider anything but adversarial.
What could you do better?
We are in conversations with lots of folks right now around making sure we can make our policies clearer. We can make our reporting more obvious. We’re obviously continuing to lean in with third-party measurement partners like we always have, companies like IAS and DoubleVerify, and more, that worked with YouTube for years and we’re going to continue to engage with them on topics like this to make sure we’re providing them the tools to provide independent third-party measurement.
Are there policy changes you would make to address any concerns raised in the Adalytics reports? For instance, you could change how made-for-kids content is labeled so there is less confusion.
I’m not familiar with any active changes that we’re planning to make with regards to content creators who have made-for-kids channels on YouTube. We do take this opportunity to think about the way that our reporting works and the way that our policies are structured to make sure that there isn’t any confusion with our advertisers and agency customers.
A lot of measurement partners and advertisers are now turning on stricter “inclusion” lists, which are pre-approved channels that vet where ads run. Is that a smart policy, or does it have disadvantages?
Inclusion lists are certainly something that we see advertisers do and they want specific channels and even specific videos. I think that the more work that you put into that, the less that you can be responsive to changing consumer habits and trends, and things that might be becoming popular really quickly on YouTube. And so we really feel the tools that we provide advertisers right now, to customize the types of audiences they want to reach, including do you want to run on content suitable for families, do you want to reach these kinds of audiences, we think that provides the right amount of customization for most advertisers.
English 
Українська