Most data clean rooms pitch themselves as being interoperable – except for the biggest ones. And this leads to confusion.
Whereas independent data clean rooms focus on interoperability, walled garden ad platforms have very little interest in it.
And walled garden-based clean rooms are where the money goes, namely, Google’s Ads Data Hub (ADH), the Amazon Marketing Cloud (AMC) and Meta Advanced Analytics.
Walled garden clean rooms have a built-in edge compared to open clean rooms, said Alex Bloore, Goodway Group’s senior director of product and data. In addition to having superior first-party data, the fact is that Alphabet, Amazon and Meta having don’t need to collaborate, which mean they can focus on their own business priorities.
The ability to optimize ad spend across Google’s properties, for example, is more valuable than an advertiser and publisher matching data and running a clean room campaign on the web.
Which begs the open question: Can anyone beyond the Big Tech giants win in the clean room category?
The big two
Google and Amazon’s respective clean room offerings are in the lead in terms of advertiser and agency mind share.
Both follow similar principles, but take fundamentally different approaches.
Google was first out of the gate with ADH and was first the walled garden platform to do audience targeting via a clean room. ADH also has more bells and whistles for advertisers, such as modeled frequency capping tools.
But ADH is also a strange hybrid business for Google, because it’s used primarily by advertisers and agencies but is actually part of the Google Cloud Platform (GCP).
The strength and ubiquity of Google’s advertising business gives GCP leverage to sign up accounts. But since ADH is plugged deeply into Google’s ad business and identity graph, Google can’t be as open as Amazon is with its clean room.
In October, for instance, LiveRamp was announced as the first identity partner synced with ADH. Although that integration comes with strict guidelines – it’s for measurement purposes only and with no audience querying – the partnership is a coup for LiveRamp, Bloore said.
A week after the LiveRamp/ADH announcement, LiveRamp, Habu and InfoSum signed on as the first clean room vendors to partner with Google PAIR, a clean room-style product for third-party measurement. And, back in 2021, LiveRamp was the first identity partner to integrate with GCP.
The point, Bloore said, is that Google’s MO is to take a slower approach with ADH. And when Google does bring on third-party tech companies, it does so with one or a small handful of vetted partners over a long testing phase.
LiveRamp was also the pilot partner for AWS Clean Rooms, which launched in November. But, according to Bloore, AWS has a more open cloud developer perspective. AWS is generally happy to add whichever vendors customers want or need. The priorities of Amazon Advertising aren’t factored in.
“That is their differentiator at AWS and the philosophy that they built their business on,” Bloore said. “I don’t know if you could say that Google has built its business on giving consumers lots of choice.”
The day-to-day function of ADH is to generate more value from Google’s media and advertising empire. AWS, by comparison, is more concerned with deepening it relationships with cloud infrastructure developers and adding queries, since it’s paid by computing load.
Crowding the clean rooms
After Google and Amazon, there’s a second tier of clean room operators that don’t have major clean room revenues, but do have ambition and are not accustomed to being “second tier” anything, including Meta, Disney, Roku and Walmart.
Although Meta Advanced Analytics is an important component of its overall ad platform, it’s less of a growth opportunity because Meta doesn’t have its own cloud infrastructure business.
An advertiser that spends heavily on Facebook and Instagram and needs to test creative in more detail will get real value from Advanced Analytics, Bloore said. But that value only exists for marketers that are largely dependent on Meta’s platform to reach customers.
Disney and Roku also fall into the “tier two” data clean rooms category.
Both are following in the footsteps of Google and Amazon in the sense that they launched walled garden-style clean rooms, but with one major difference: They partnered with third-party tech to do it. Roku is backed by Snowflake while Disney leans on InfoSum, Habu and Snowflake.
Having gone the third-party route, Disney and Roku’s clean rooms are limited to narrow use cases. Disney advertisers, for example, can use its clean room tech to more accurately frequency cap, but only on Disney-owned media. Still, rather than being forced to use the Google DSP to create audiences in ADH, an advertiser can use the Disney clean room and The Trade Desk.
And whereas Google combines YouTube, Maps and web browsing data, Roku’s clean room merges its CTV audience data, its walled garden DSP data (the former dataxu, now Roku OneView) and data from Foursquare, a pilot partner.
Data clean rooms are an important service for TV advertising, Louqman Parampath, Roku’s VP of product management, told AdExchanger at the time.
An advertiser that onboards data for a big holiday campaign, for instance, will have a matched audience they can target and analyze over the course of the campaign. The following year, the brand and retailer will have to do the same thing and get a new anonymous match group.
But within the data clean room, Parampath said, advertisers can retain anonymized audience sets over time to attribute whether the same audiences are more or less receptive the next year, or which kinds of new buyers later became loyal customers.
There is significant demand for this category of hybrid walled garden clean room, as evidence by the growth of Snowflake as an advertising and media player.
Other independent clean room providers are in the mix, too, including InfoSum, Habu, Optable and LiveRamp, but Snowflake has serious scale.
Snowflake invested in Habu and in OpenAP, a data clean room provider co-owned by NBCUniversal, Fox, Paramount and Warner Bros. Discovery, and has enough gravitational force to continue investing in startups that could be stitched together as a first-party network.
Privacy pros
But what about privacy? It’s one of the main selling points of a data clean room.
Snowflake’s ownership stake in OpenAP and Habu provides potential air cover to share and merge data, even when strategic commercial partnerships could theoretically accomplish the same goal. That’s because, for many important media channels – anywhere Apple’s or Google’s advertising data policy applies – privacy is about first-party data ownership, rather than specific tracking tactics.
An important advantage for platform-based clean rooms compared to open clean room tech is that they provide a way for Google, Amazon, Meta, Roku, et al. to more closely track their own users – not as a way to help a third-party company track someone else’s customers.
For example, Google tracks users and combines data from across Maps, Gmail, Chrome, its ad network, search and many other subsidiaries. That’s considered kosher. But if an advertiser, a major publisher and few data or ad tech companies combine data using a cloud-based infrastructure provider like Snowflake, even with the same privacy-preserving tech in place and contractual commitments not to expose profiles, it’s often a no-go.
The privacy magic isn’t in encryption or even the degree to which unbeknownst users are tracked online; it’s whether the data is held in the hands of one company.
From that perspective, a walled garden clean room is appealing.
To be fair, third-party clean rooms that plug into identity solutions are more private than current systems, according to Paul Bannister, chief strategy officer at CafeMedia.
But that argument isn’t going to hold water with European regulators.
“Informed consent will be required and the bar the EU will set for informed consent is going to be so high as to basically be impossible to get,” Bannister said.
If an advertiser, publisher, a few data companies and a cloud provider (Snowflake, say) were to partner on clean room analytics, each company would require informed consent from every individual. That’s tough to scale. But Google can adeptly bundle its services into one identity spine.
“The publisher and DSP integrations are what I’m starting to see slowly scale,” said Therran Oliphant, SVP of data and technology at Essence.
But going through a third-party clean room vendor, a DSP and even just a small subset of publishers is messy if an advertiser isn’t also leaning on a platform solution like Google ADH or Amazon Marketing Cloud.
“Currently,” Oliphant said, “to make this work is resource intensive and cost prohibitive.”
