TL;DR: The phrase ‘data clean room’ is open to interpretation, and misinterpretation, but it’s important to know and distinguish between the different types, writes James Prudhomme, CRO, Optable.
The Inuit have 50 words for ‘snow’, taking in all the many and varied characteristics of the white stuff that most of us don’t tend to think about. The ad tech industry, meanwhile, has 50 definitions of a data clean room, but essentially only one name for all of them.
All data clean rooms focus on enabling collaboration between partners on the basis of audience or customer data. But different types of data clean room are emerging all the time, and as with the Inuit and their snow – ‘aniu’ means snow used to make water, for instance, whereas ‘maujaq’ is snow you sink in – the differences between them are significant.
Being a data clean room vendor, when we hear about someone else operating a data clean room, we try to understand what type it is. Over time, as people begin to understand the distinctions, the language will inevitably become more descriptive – look at the programmatic space, where it took a while before we got to universally accepted terminology like DSP, SSP, exchange, DMP, and all the other Ps.
To get the process started, then, here are the most common types of data clean rooms we see in today’s digital marketplace:
1. Publisher-provided data clean room
The best example of this is Google’s Ads Data Hub. Broadly, this is a platform that is operated within a publisher’s walled garden and allows advertisers to match data with the publisher’s own. The result is insights and activation capabilities that make campaigns better by virtue of starting with real advertiser data.
2. Data vendor-provided data clean room
When speaking to established data vendors who tend to be large companies, we sometimes hear something along the lines of “but we’ve been doing this data clean room stuff for 30 years!” What they mean is that they’ve been in the data-matching business for many years. The core value proposition of a data vendor’s data clean room is to onboard advertiser customer data and make that data better by using the vendor’s data – often drawn from a panel. There are loose notions of privacy, of course, but their definition of a data clean room is more related to security: what they’ve been doing for 30 years is receiving customer data from advertisers in a secure way.
3. CDP-provided data clean room
As more companies lean into customer data platforms as their single source of truth on customers, some CDPs have started commercializing a data clean room approach. In most cases, the main objective is to allow the CDP’s customers to do more with the data they store within it. Leveraging notions like pseudonymisation, the ultimate goal is to use the data clean room approach to establish the CDP as the single source of truth on customers, regardless of where the actual customer data originates from.
4. Storage vendor data clean room
A data warehouse can also be used to store customer data and of course, some of it will inevitably be private. For a storage vendor, offering a data clean room approach really means offering the framework for creating views of data that can be shared with other users of the same storage vendor. This works really well, but another layer of functionality is invariably required – you need applications on top that will query the right data, offer the right reporting, and in most cases, allow you to export data out of the storage layer for activation.
5. End-to-end clean room
This is a complete, end-to-end clean room solution, of the kind Optable offers. There are data clean room ‘apps’ that encompass functions from analytics all the way to direct integration. A complete data clean room platform integrates with existing infrastructure, allowing customers to collaborate with partners who aren’t themselves customers. These can be custom integrations, leveraging APIs and command-line tools. In our case, publishers, advertisers and ad tech vendors can collaborate through integration with Prebid and the emerging Seller-Defined Audiences standard, going all the way to direct ad server integration. This allows us to offer a privacy-preserving solution that covers a complete digital media workflow, integrated for activation, and in turn makes an end-to-end data clean room suitable for use by aspiring walled gardens or embeddable by CDPs. In fact, any company can create its own data infrastructure using a privacy-preserving platform of this kind.
It’s an exciting time to be building privacy tech. There are other variations, and we expect many more – we’ll be sharing these in upcoming publications.
Optable is a data collaboration and clean room solution designed for the advertising ecosystem in the age of privacy. The product was inspired by the radical transformation in how audience data is governed, connected, and activated. Optable takes an end-to-end approach to data collaboration and is integrated with ad delivery platforms for secure activation, making it possible to deploy campaigns directly from a clean room. It is the only clean room solution that offers frictionless collaboration and interoperability enabling customers to safely and securely match audience data with any partner.