In this article, Bharat Baldewa, senior director, Merkle, shares what a data clean room is and how it works. He also discusses why it is important now more than ever.
 The concept of a data clean room is not new, but for an organization’s advertising and data needs, it is becoming more relevant than ever. Growing privacy considerations are driving rapid change in the way that marketers access, collect, and share customer data to support analytics and keep up with developing legislation. The number of regulations that govern the handling, storage, and usage of customer data is on the rise. Furthermore, actions from walled gardens, such as Google’s plan to block third-party cookies by 2023, are creating a worrisome sense of urgency for organizations to plan ahead.
Cookies are heavily relied upon today to give marketers a tracked view of their customers. With these changes, we need new ways of identifying customer data and using it in a privacy-safe manner. All-in-all, data sharing approaches need to be reconsidered. Data-savvy advertisers understand that true advertising and, more generally, business efficiency can be achieved by maximizing the power of first-party data, second-party partnerships, and third-party data attributes.
Maintaining strong compliance with consumer privacy regulations and achieving a high level of business efficiency can seem like conflicting objectives. To align these two goals in this ever-changing landscape, organizations must explore and establish private data clean rooms in which they can share and analyze data of mutual interest. This can help them keep their advertising focused and addressable while developing a strong privacy-first core.
Learn More: Connecting Customer Data in Today’s Data Platform To Deliver Personalized Experiences
What Is a Clean Room and How Does It Work?
A clean room is a privacy-safe data environment that is typically set up by a third party. It allows marketers to share and analyze first-party data with partners in a privacy-safe environment, with controls for how much data is exposed to other parties. It also enables marketers to run queries on the data, pull aggregated reports on campaign performance, onboard and enrich audience data with your third-party partners, and facilitate attribution reports that don’t risk exposing data. And all is kept within an isolated environment.
There is no personally identifiable information (PII) in these environments, and an enterprise or a user cannot reverse engineer the IDs to PII. This can significantly reduce the amount of data subject to regulations, such as the General Data Protection Regulation (GDPR) in Europe and the California Consumer Privacy Act (CCPA) in the US, and increase opportunities for brands to share data in a privacy-compliant manner.
Architecture of a Clean Room
At a high level, the architecture of a clean room starts with the data sources. The diagram depicts the various sources of data flowing into the clean room, from CRM data to ad server data, etc. But the data can also flow in from various identity connections. These connections can come from demand-side platforms (DSPs), as well as walled-garden platforms, such as Facebook and Google. The translation layer allows for identity resolution of the ingested data. The key here is that data is stored and can be used only as de-identified records. When an audience is selected and needs to be pushed back through and utilized, the translation layer is again used to re-key and push out to the appropriate platform.
The Challenges
 When building a clean room, there are some common challenges that are typically encountered along the way. The first challenge is that first-party data often can’t leave the local storage environment. Additionally, measurement must happen across multiple entities and storage buckets, making analysis a somewhat tedious task. Joins and queries need to share common identifiers, such as a hashed ID since PII is not used in the environment. Also, some regulations prohibit exposing certain first-party data to other parties.
Learn More: As Long As Data Management is Broken, So Are Most Marketing Tools
Using the Clean Room for Second-Party Collaboration
There is an increased interest in using second-party data as the new third-party data. Second-party data is first-party data that two or more parties decide to share for mutual benefit.
Brands and publishers need to develop partnerships for standard, safe ways of sharing data with second parties as well. Brands will need to think of media owners more like partners rather than vendors for data sharing, expansion, and activation. The overlap of data can then be used to understand what users are engaging with, buying, etc., to generate insights, build look-alike models, and support advanced measurement. As an example of how the clean room is best used, think about the overlap between a car manufacturer and a popular auto publisher. The data that these two organizations could share would be mutually beneficial to learn more about their shared ideal customers while complying with privacy regulations.
Walled gardens have also set up clean room environments to share marketing performance data with advertisers in an anonymized environment while still exerting strict controls on usage. Advertisers’ first-party data is then added to the clean room to see how it matches up with the accumulated data from the platforms. From there, advertisers can see how the different data sets match up and use it to evaluate the performance of their campaigns within those walled gardens.
In summary, clean rooms are powerful tools to help organizations share and bolster data while complying with privacy regulations by anonymizing a person-based identity across various data sets and sources to apply analytics use cases.