By Lana Warner, Senior Director of Partnerships & Strategic Solutions, Lotame
Data clean rooms have been a trailblazer in the new privacy-first, cookie-free paradigm. At first, they were a novel idea that had marketers and media owners excited about the possibilities of data sharing; then it seemed like everyone had built a clean room of their own; and now confusion has set in as the category splits and mutates into various forms.
It’s this confusion that we need to tackle to fairly assess the value of data clean rooms, who can benefit from using them, and whether they live up to the hype. So, let’s explore the different types of clean rooms on the market today and how each can help connect the dots in a fragmented advertising ecosystem.
Data warehouse clean rooms
Rather than being distinct and independent platforms, many clean rooms today are built into data warehouses. For example, Snowflake and Amazon Web Services (AWS) offer clean rooms that allow existing customers to collaborate and compare data sets.
The clear limitation is that both you and your collaborator must use the data warehouse the clean room is built on top of, and few would be willing to port their entire data library for a single project. But if all parties already use the same platform, their ability to handle vast volumes of data makes them ideal for attribution and measurement.
Also, despite being packaged into existing platforms, these clean rooms are by no means plug and play. On the plus side, those with the tech and talent can bring in their own algorithms, query sets, or identity layers to develop highly bespoke functionality.
Walled garden clean rooms
At the other end of the usability scale are clean rooms offered by the likes of Google, Meta, and Disney. These are built into existing media planning and buying interfaces but aren’t exactly data collaboration tools. You can bring in data to enhance retargeting or suppression in campaigns within the host’s ecosystem, but don’t expect to get anything back out.
What these clean rooms lack in flexibility or collaborative capabilities they make up for in simplicity. Big tech and media players put a lot of effort into minimising friction for marketers, with slick interfaces and pre-set functionality that can quickly deliver common use cases.
If you’re planning on running a campaign within a walled garden and you have compatible data ready to go, there’s no reason not to explore their clean room options. For any other purpose, you’ll have to look elsewhere.
Data collaboration clean rooms
Data collaboration clean rooms allow two (or more) parties to port their data into a single environment for data matching. Much of the data transformation and orchestration tends to be handled by the platform itself, making for a relatively user-friendly experience.
The main selling point of data collaboration clean rooms is that the outputs can be taken from the platform and used elsewhere. This makes them ideal for the downstream enrichment, analytics, and activation that walled garden clean rooms cannot offer, while their independence from underlying infrastructure provides more flexibility than data warehouse clean rooms.
There are limits to this flexibility, however, as there’s always the risk that two parties might prefer different clean room vendors and will be unwilling to pay for and learn how to use another. Work towards cross-platform interoperability and single-party payment models is needed for data collaboration platforms to achieve their full potential.
Query clean rooms
Finally, query clean rooms work similarly to data collaboration clean rooms, with the distinction that no data is ported directly into the platform. Each party could keep their data in a data warehouse, a CDP, their own private ecosystem, or anywhere else; the clean room doesn’t ever touch the data directly, instead, it allows users to make queries to find overlaps and matches.
Query clean rooms come closest to the original promise of data sharing and are ideal for those who take a hard line on data portability. But they do have downsides: a query-based structure is less customisable than the hands-on approach of data warehouse clean rooms, while data must be prepared so that it is compatible with the platform, making them inaccessible to those with limited data expertise.
Hopefully, this guide has helped you get to grips with the diverse array of clean room options on the market. Remember, whichever you choose, making the most of a clean room begins with considering your use cases and the data available to you. Otherwise, you risk dumping time and resources into the wrong technology and missing the rich opportunities of data collaboration.
