Meta is starting to talk more about data clean rooms as one answer to post-cookie advertising, but it has some reservations about the technology that is among the industry’s most-hyped products.

Dennis Buchheim, VP of Meta’s advertising ecosystem, said he is worried that ad tech leaders at an IAB event on Thursday might “throw tomatoes” at him for saying so, but data clean rooms are “not fully secure, they’re not fully private, in all cases.” The sentiment around clean rooms, coming from the second-largest internet ad company in the U.S., might be considered blasphemy by ad tech firms and publishers that have wholeheartedly embraced the concept.

Buchheim shared a slide during his talk, labeling clean rooms as “identifiable data,” and the slide compared that to “privacy-enhanced technology,” which relies on aggregated data, meaning it's based on groups of data that obscure identity.

Related: Why brands are turning to data clean rooms

The question of what is identifiable data and what’s anonymous is not a minor one. The FTC recently said it will pursue companies that claim to collect “anonymous” data when that data could be easily linked to identity. Clean rooms offer keepers of data a secure environment to host and query that data, which can still be personal, but the entities are responsible for obtaining consent from consumers to use it.

The clean room space is heating up: Google and Amazon, Meta’s top rivals, are developing their own data playgrounds, as are top publishers including Disney and NBCUniversal. There are many different flavors of data clean rooms, but essentially they all allow a publisher to share data with advertisers about their audiences so they can target ads and measure ad outcomes like sales.

Clean rooms are considered a more secure way for platforms and brands to play with data sets, without technically sharing that data with each other, while also managing consumers' consent. Clean rooms can help organize what data publishers share, with what partners, and for what purposes.

The technology is becoming so widely adopted because cookies are going away on web browsers, with Google promising to kill cookies on Chrome by the end of 2024. Apple already blocked cookies on Safari, and last year, it implemented App Tracking Transparency policies that hobbled advertising IDs on iPhones. The future of ad targeting, online data-sharing and identity is being debated throughout the industry. Hardcore privacy advocates want to severely limit data transmissions, while marketing tech leaders worry about the death of personalized advertising. Brands are worried about losing data signals that help them find new customers and juice their sales through targeted online advertising. No ad platform has demonstrated the pressures of the changes more than Meta, which serves 10 million-plus advertisers struggling to adapt to data-sharing restrictions.

Meta, which owns Facebook, Instagram and WhatsApp, is developing a different answer to clean rooms, by pursuing privacy-enhanced technology, using encrypted and aggregated data to hide consumer information, but still returning useful calculations to advertisers.

Meta’s presentations at IAB Tech Lab’s event offered a good glimpse into how it thinks of clean rooms. Meta’s privacy-enhanced technology [PETs] is on track to achieve ad measurement and targeting through aggregated and anonymized data. Meta is working on private lift measurement and attribution, for instance, using aggregated data to determine outcomes from ad campaigns. PETs is “where we see the growth opportunities,” Meta’s Buchheim said.

Meta is talking more with top ad agencies, brands and data service providers about how it could embrace clean rooms in the future, according to top ad leaders, many of whom spoke with Ad Age on the condition of anonymity. Meta has clean room-adjacent services like its conversions API—application programming interface—which is very much a middle-ground technology on the way to PETs. Meta has prodded advertisers to plug into the conversion API, to keep measuring ad campaigns as Apple applies more anti-tracking policies. The conversion API is a pipeline for advertisers to share conversion data, like when a consumer took an action on their website, directly with Meta’s servers. This way, Meta can share which ads were most effective.  

However, brands have to trust Meta’s calculations through platforms such as the conversions API, and not all brands trust Meta to judge itself.

“It’s a problem,” said a top exec at a data services provider, speaking on condition of anonymity. “Think about it, you’re Unilever, you gather a ton of information about your customer, it’s very valuable to you. You don’t want to push that into Meta, here’s all the information on my customer. You want a way to match your audience with the audience in Facebook or Instagram, but you don’t want to have to push that into a Meta environment.”

If Meta offered some clean room integrations, brands could control their data and audit the calculations to grade the reliability of Meta’s results.

Ultimately, Meta could need media holding companies and data clean room providers as partners to create a more neutral data setting. The industry is moving in that direction quickly with the prominence of data clean room players like Snowflake, InfoSum and Habu, and the fast adoption from major internet ad platforms, publishers and retailers.

“Every major platform is working toward clean room solutions,” including TikTok, said an agency exec speaking on condition of anonymity.

Google and Amazon are both moving full speed ahead, with ad clean rooms and their cloud services, which are going to be used by brands to study their consumers and analyze ad campaigns.

“Meta is the most far behind,” said one media holding company exec, speaking on condition of anonymity. “They recognize that and they’re also thinking, not just how to do what’s next, but how to do what’s two steps ahead.”

“But that puts Meta in a tough spot,” the exec said. “Because they’re not in the driver’s seat to push the use of data for media buying, media optimization and media measurement, for probably a year or two.”