The standardization of data clean rooms

With the impending demise of third-party cookies, the ad tech industry is busy looking for alternatives. We all know that first-party data is a vital starting point, and that enriching this with second-party data can provide the extra scope required to gain insights into audience-brand relationships. One way to leverage this approach is by using data clean rooms.

While there’s a great deal of buzz about these clean room environments, there’s also a considerable amount of confusion about their definition. However, IAB Tech Lab is looking to change this by publishing a set of standards by the end of the year.

So, let’s take a look at the current state of data clean rooms, and where advertisers hope they could go in the future.

What are data clean rooms?

Data clean rooms are safe environments where companies can combine their aggregated (rather than individual-level) first-party data to gain insights, while still retaining stringent controls and complying with data privacy regulations.

They work by taking a company’s first-party data and examining how those records match up with those from other participants. This enables advertisers to detect overlap, where they may be repeatedly showing ads to the same audience, as well as any duplicated reach on different platforms, such as Twitter, Google, and Meta. Crucially, none of the combined data leaves the clean room, making it compliant with data privacy regulations.

Why are data clean rooms important?

As you add more channels, it becomes harder to model attribution, which in turn means it’s more difficult to find out whether you’ve reached the desired audience, and to understand ROAS (Return On Ad Spend). Most digital advertising uses some form of behavioral targeting, which profiles users and tracks their activity across different websites and devices. But this strategy has conventionally been powered by third-party cookies.

However, web browsers are phasing out these cookies due to concerns about user privacy. Data clean rooms are therefore becoming increasingly important because marketers are looking for a different way to gain the data needed to assess ad effectiveness, while avoiding falling foul of new data privacy laws.

Are all data clean rooms the same?

No, there are several different approaches. There are multi-party centralized data clean rooms, such as those provided by Snowflake, where data from multiple companies is merged into a single data warehouse. Each party is denied access to the combined data, but data analytics is carried out inside it and then the results provided to participants.

There are also clean rooms that remove any PII (personally identifiable information), then put each party’s data into a separate “bunker”. Participants can then query all of the datasets without connecting them together. An example of this decentralized approach is InfoSum or Optable, the latter of which integrates with existing infrastructure to allow parties to collaborate with partners who aren’t participants in the clean room themselves.

Finally, there are single-party centralized clean rooms. For example, Google’s Ads Data Hub, which is largely a tool for measurement. An advertiser can input their first-party data, and doesn’t receive any of Google’s data, but they can use Google’s intelligence to gain insights.

Arguably, customer data platforms are a type of single-party data clean room since they allow advertisers and publishers to query multiple first-party sources in a single location.

So, why do data clean rooms need to be standardized?

Limitations arise because there isn’t a clear or singular definition of what a “data clean room” actually is. This means that there are no standards via which clean rooms can communicate with one another. For example, InfoSum might use a set of protocols which are internally standardized, but without an external schema upon which the industry has agreed, there is no interoperability between different vendors.

Not only would this interoperability enable clean rooms to offer more sophisticated data analytics beyond basic measurement, but this would be achieved while preserving data privacy.

It’s increasingly important for those in the ad tech industry to set up alternatives to third-party data, given the approaching sunsetting of third-party cookies. First-party data is your starting point, but this will likely need to be supplemented with second-party data, so it’s worth doing the research now to check whether using a data clean room is a viable solution for your company.