Data clean rooms were first popularised by the likes of Experian and Acxiom as a way of those organisations matching their data to your CRM data. These ‘Safe Haven’ solutions would require you, the advertiser, to upload your offline data assets into the third-party’s data clean room, where it would be matched on a unique identifier, such as email or home address.

As we moved into 2020 however, the main names being associated with data clean rooms were the walled gardens; Facebook, Google and Amazon. While Facebook’s and Amazon’s data clean rooms are suspected to be in development, Google is already making its available to advertisers in the form of its Ads Data Hub. ‍

Google Ads Data Hub‍

Prior to launching Ads Data Hub, Google allowed log level data, containing a user ID, to be exported from their DoubleClick platform, and uploaded into a DMP to enable you to monitor the performance of your campaigns. However, in April 2018 Google removed this feature, citing privacy concerns and GDPR.

Google’s Ads Data Hub is positioned as a privacy-safe environment in which you can run analysis of your data with Google’s data sets (such as ad exposure and campaign information from YouTube). While Ads Data Hub does not allow data to be exported, it does allow you to upload your customer data into Google’s environment. This upload of customer data enables you, as an advertiser, to match your data to impression level data across Google’s advertising solutions, enabling you to measure campaign performance. 

However, in doing so, you are accepting a number of trade offs. Firstly, and as we’ll explain in more detail below, you are sacrificing control of your data to Google. Secondly, you are locking your data to Google's own ID (which you don’t have control of). And finally, you are unable to export data back out for further analysis, creating a black box within Google’s environment. ‍

Re-imagining the data clean room for a privacy first world‍

The fundamental challenge with Google’s approach, as well as with the legacy data clean rooms offered previously, is the requirement for your rich first-party data to be processed/merged with other data sets and locked into their, interoperable, ID’s. Having to work this way means you lose sight and, ultimately, control over your data and how it is used. 

InfoSum’s vision as the future of data clean rooms is to provide a privacy-safe environment where insights can be unlocked across multiple data sources and multiple parties, without any data having to be locked away. This is why InfoSum has developed the world’s first data clean room that never requires you, or any party you wish to work with, to share and lose control of data.‍

InfoSum’s data clean room is built on our proprietary Bunker technology. Bunkers are completely isolated cloud instances that are created and controlled by the data owner, you. You place your data into the Bunker and automatically map it to our Global Schema (our version of a data dictionary which drives our ETL process). Once the data is uploaded, it never moves again and only you can ever access the Bunker. 

Instead of sharing raw data between parties, analytic permissions are granted between parties. These permissions never grant other user access to the data within your Bunker, they simply enable them to use the data set in their analysis (and vice versa where permissions are granted both ways). All analysis is entirely anonymous and driven only by the intersection of the two data sets. This means you can only ever learn about individuals which you have in common. 

Finally, because we have built differential privacy concepts into every step of our data clean room process, it is impossible for a single individual to ever be re-identified through analysis. This ensures you are always putting the privacy of your customers first. ‍

Data onboarding with a data clean room

Traditional data onboarding solutions require you to share your raw data with them to be matched against online identifiers and flattened to a third-party ID. InfoSum’s approach to data onboarding means data sharing is no longer needed.. Instead, your offline and online data sources can be uploaded to Bunkers. An identity vendor, such as zeotap can also upload their data to a Bunker and grant you permission to measure the intersection between your data and their identity graph. 

This intersection can then be further analysed, segmented and sent for activation. InfoSum is making possible end-to-end data onboarding without ever having to share any data. A world first.‍

Building data alliances with a data clean room

As we covered in a recent blog, data alliances can be valuable ways of augmenting customer knowledge and increasing the scale of addressable audiences available for you to activate against. However, they have previously been hampered by a requirement for the various data sets to be centralised in a third-party vendor to be analysed ‘as one’. 

With InfoSum’s data clean room, data collaboration and, therefore, partnership, becomes safe, nimble and quick. Each data set can be uploaded to their own isolated Bunker and then analysed as a whole, without requiring the data to be centralised or commingled at all. 

These unified audiences can then be made available to you to match your customer data against, without you having to share your data with the various parties in the alliance