Clean rooms, explained: How they became the buzziest tool in ad tech

The hottest thing in ad tech is also the best gift a 10-year-old can give their parents: a clean room.

It’s one of the industry’s favorite buzzwords, a technological solution that claims to allow advertisers, publishers, and platforms the ability to securely share their own aggregated data without giving too much away to partners.

But don’t take our word for it—ask the market:

• In early November, Permutive, a clean-room provider, raised $75 million.
• Last week, Habu, a clean-room software company, raised $25 million.
• Disney is piloting a clean room alongside Omnicom Media Group, Habu, InfoSum, and Snowflake.

What does a clean room do? Well, if fictional brand Ryan’s Lemonade wanted to share insights with Target, each would be able to put their aggregated data into a clean room to see what the other knows about audiences they have in common—what else they might buy, how often they buy, their demographics, and interests (like what else is in their shopping cart).

• Clean rooms can be used to both gain insights into audiences and as a tool to measure campaign performance, depending on the partners involved.
• Instead of proxies and guesses to estimate audience insights, a brand would literally be able to look under a retailer’s first-party-data hood, all while being privacy compliant.

It’s Switzerland for data, a neutral intermediary where all sides can set parameters as to what information is actually seen. Clean rooms allow advertisers to “expose certain data aspects that they are comfortable sharing” explained Matt Kilmartin, Habu’s cofounder and CEO, to Marketing Brew. “You can determine what you actually want to do with different partners, and what your partners are comfortable doing as well.”

Right now, there are primarily two different kinds: Walled gardens, like Google and Amazon, operate clean rooms that give brands a look inside their own ecosystem. Software companies and data managers like Habu, InfoSum, and Permutive operate the second kind, acting as a conduit between parties that want to share their data.

When a brand partners with a publisher (Google, Amazon, a retailer with a media network, for example), a clean room can give both parties a closer look at campaign performance data like reach and frequency figures without giving away that sweet, sweet personal first-party customer data they often can’t share. In return, advertisers get an “aggregated output” without individual audience identifiers, explained Therran Oliphant, executive director of marketing technology consulting at media agency PHD.

This “output” can reveal audience segments, campaign insights, and look-alike audiences, which can then be shared with a publisher, a demand-side platform, or an ad network to inform a campaign. Or, if you’re a retailer with an ad network, brands can leverage this output when buying ads.

• If you’re a marketer running ads in a walled garden like Google, then, naturally, you can optimize these insights within Google.

Most companies offering clean rooms can run a machine learning model (um, fancy computer math) over someone else’s data set without ever seeing it. This anonymized, modeled data can give advertisers a good indication of who their audiences are and what they’re doing without the personalized data of the audience being revealed.

“Indirectly, they allow for targeting,” Oliphant said. “With a clean room, you get those attributes, but not by individual ID.”

Privacy players

Why now? Privacy legislation in the US is about to get serious. New consumer-privacy laws in California, Colorado, and Virginia take hold in 2023, limiting how advertisers and brands can share consumer data. Stricter data-privacy standards in Europe have also affected how global advertisers collect their data.

As a result, any bad actors under those laws sharing audience data without proper consent could be hit with fines and penalties, raising the stakes regarding how advertisers collect data and who they share that data with.

If you haven’t been paying attention, user privacy is en vogue. Between Apple’s crackdown on mobile ad tracking and the slow, but inevitable, death of Google’s third-party cookies in 2023, there is less personal data going around.

So, brands have had to scramble to find new ways to gain insights that are more privacy-centric.

“Once upon a time, you could have third-party data floating around all over the place and it was easy for individual data owners to link their data up in the cloud,” said Alexandra Bannerman, a senior product marketing manager at Permutive. “It’s not going to be easy to do that anymore. Each data center is going to have their own data in their own little silo and their own little walled garden, and it’s going to be hard to collaborate.”

One solution? Clean rooms.

The big dogs

Google and Amazon each operate their own clean-room solutions, the Ads Data Hub and the Amazon Marketing Cloud, respectively. Facebook has had a few iterations and is beta testing its newest iteration, Advanced Analytics.

Because a walled garden is, well, a walled garden, they can each give advertisers an aggregated, or grouped, look at audience data on their platform and let brands compare it with their own customer data. Like performance metrics, this data can be used to optimize a campaign on these same platforms. Again, it’s all aggregated.

• Google will only give insights into a group of 50 or more users, so advertisers can’t identify individuals.

“The benefit of the walled garden is the simplicity of use...As a user, I am only required to bring my stuff to the table,” said Ira Maher, VP of integrated agency The Basement.

But, but, but: The walled gardens write the rules. It’s still a siloed perspective—advertisers don’t only advertise on Facebook, Amazon, and Google (though….many do).

However, unlike the walled-garden clean rooms, clean-room providers are independent of what’s being shared—they’re just a tool.

“If I’m using an independent party, clearly, the lack of potential bias is an advantage. But I’d have to go to every partner I want to participate in this clean room and coordinate everyone coming together,” Maher added.

Not so easy

Though clean rooms may seem like a more privacy-friendly way to get wet and wild with your data, it’s not that simple. Consumers still need to give clear consent to advertisers before brands are allowed to share their data, clean room or not, explained Richy Glassberg, cofounder and CEO of SafeGuard Privacy, a privacy-compliance company.

“A clean room is not a washing machine, you cannot throw your data in there, throw a tide pod in, and think that it’s okay to use,” he told Marketing Brew. “I would say to people: Be very cautious about being told that a clean room is going to solve all your problems.”

There are also scaling challenges. Currently, clean rooms still operate on a mostly one-to-one scale, partner to partner. If you’re Pepsi and you want to share data with Walmart, Disney, The Wall Street Journal, and Billy Ray’s Taffy and Jerky Emporium, that’s possibly four different clean rooms the brand would have to operate and manage, which could require further staffing.

In other words, Maher said it could take advertisers “10 times as long” to get the data and insights they’re used to getting via tools like third-party cookies.

“That’s a lot of legwork that currently doesn’t exist, principally because instead of doing that leg work, we’ve used legacy technology that’s been somewhat abusive to the perception of consumer privacy.”

Do you work in advertising or media? Is there something we should know or haven’t covered? Drop a tip or reach out to Ryan Barwick at [email protected].