Decentralized Clean Rooms, Not just another Adtech Bandaid

On popular demand, I was asked to write more about the recent study from IAB on #datacleanrooms. So here are my views.

First and foremost, the industry is looking at DCRs as “yet another” point solution in the smorgasboard lumascape of AdTech and MarTech solutions. DCRs should not be treated as a band-aid to fix challenges with the existing technologies. It has the potential to totally re-architect the existing ad-tech infrastructure and that is where the biggest cost efficiency and return on its investment comes in. 

I want to bust 5 Myths from the IAB report, basis our experience in working with our partners. That is because, we did not see DCRs as another centralized solution that is being built or set up by individual enterprises for their own data. If the purpose of DCR is to #collaborate, then we need at the least two hands to clap, which Google recently called it as a PAIR. We looked at the application of DCRs for more than a PAIR. If it is for an advertiser, it is meant to cover their entire publisher marketplace. If it is for a publisher, it meant to offer smarter collaboration for enhanced subscription. Essentially, it should enable a trusted environment for a safe and secure marketplace.

1. Set up is NOT lengthy and complex. Centralisation leads to repeated execution of the same process with all the participants who wants to collaborate. That makes the entire process extremely complex. We have re-structured the DCR into two modules. One a private configuration that is installed at your premises. Second is a federated configuration that is meant for, either a PAIR or more than a pair of participants. This ensures that most of the processes are NOT repeated. This also ensures that minimum connectivity with other DCRs are by secured by default. 
2. DCRs are not an expensive add-ons to the existing CDP, DMP or CRM suites. It just sits on the same infrastructure as a private configuration. No additional #cloud costs. No additional infrastructure. It is built as a portal for sharing and collaboration. Partners who wants to collaborate with an enterprise just need to have a similar portal enabled at their premises. Typical set up time is not more than 3-4 hours; and the enterprise is ready for collaboration. Yes, one of the basic requirement is for the enterprise to at the least have a data platform where they store their customer data.
3. Match rates should not be the metric of success. This is like throwing the baby with the bath water. If there is strong match rates and every enterprise data is totally synced up and matched with other enterprises, then there is no need to collaborate. In fact, I would argue that, DCRs should build capabilities to work on the un-matched to help the enterprises grow their customer base in a privacy compliant manner. The power has already shifted to the consumer. It is all about the attributes and behavior they exhibit online. Not the common identifier. 
4. Privacy features cannot be built over time. This should be considered as all or nothing. Privacy is the non-negotiable feature in a clean room. We cannot start with basic encryption and say that we are adding incremental features like #differentialprivacy, salting etc., as added benefits. It should all be built into the basic feature. Data should never leave the native location. We have found a way in which the data can be offered for collaboration in the most compliant manner. By moving the data out to another centralized location, which operates as a clean room we are duplicating cloud consumption costs multiple times. This is not sustainable.
5. Connectivity with other DCRs cannot be an after thought. That an enterprise will first set up their own clean room and then slowly prospect other enterprises one by one. DCRs must be set up or built for a pre-agreed configuration, at the least as a PAIR, with an ability to scale up and add more participants. Enterprises must have a clear view of their collaboration landscape and then consider deploying a DCR solution. And this should include not just analytics, but activation and attribution in their scheme of things. 

Overall DCRs must not be treated as yet another centralized point solution. It totally defeats the purpose. DCRs are the natural next step in the technological innovation where an enterprise can move from the current #web2 infrastructure to the world of #web3 where the control is totally with the consumer. It enables a total re-architecture of the current advertising stack.

That is where this all should begin. As a starting point for a re-architecture.