Data Clean Room

First Generation Data Clean Rooms

Data clean rooms were originally popularised by Experian, Acxiom and LiveRamp as a way for those organisations to match their offline data to a brand's CRM.

The walled gardens then updated the idea for the new digital world and launched solutions such as Facebook Custom Audiences and Google Customer Match which enabled advertisers to match their customer data against the walled gardens internal identity graphs to target their own customers within the social platforms.

To unlock these services , advertisers are required to upload their first-party data into the walled gardens environment. Once uploaded, the advertiser can conduct analysis against the various data sets and the customer overlap.

Decentralisation

Unlike first-generation data clean rooms that require data to be moved into a third-party environment, InfoSum’s data clean room is built on decentralised infrastructure that removes the need to share data with a third-party.

By removing the need to share data, each party retains full control of their data and never risks the commercial value of the data being utilised by a third-party.

Additionally, as no user-level data or personally identifiable information (PII) ever enters the InfoSum Data Clean Room, it is impossible to expose this personal data. By ensuring that personal data is never at risk, the compliance burden on data collaboration is greatly reduced.

Federated architecture

InfoSum Approach

Data is uploaded to a secure and dedicated instance on a cloud server, known as a Bunker. Only the creator of the Bunker can ever access this instance.

The ability to conduct analysis against this Bunker is controlled by the owner, who can grant different levels of permission depending on the nature of the relationship. No permission ever grants access to the raw data.

Queries are powered by anonymous mathematical representations that can safely move between independent Bunkers and measure the intersection between datasets.

Insight engine

Data Transformation

Traditional data clean rooms require considerable data manipulation prior to upload. This extract, transform and load (ETL) process creates a significant delay in unlocking valuable insight from the data.

InfoSum flips ETL to ELT (extract, load, transform) by requiring no change to the original data before uploading. Instead, our normalisation process standardises and maps the data to our global schema automatically.

Human intervention is only required occasionally when a data field does not map to an existing data category or where attribute data is too granular.

This unique approach removes the need for expensive data migration and manipulation processes and allows data to be uploaded and available for analysis in minutes, rather than days.

Global schema

Privacy-by-Design

Privacy has been built into every element of InfoSum’s Data Clean Room, delivering a truly privacy-by-design solution. By utilising a federated architecture, data can remain decentralised and in control of the data owner, while also enabling identities to be matched and analysis conducted - all without sharing any of the underlying data.

All results are at an aggregated statistical level and are designed to drive data insights, planning and measurement. Additionally, differential privacy techniques are applied to all results generated, ensuring that no single individual can ever be identified through the platform.

Privacy-by-design

Unlock true people-based marketing

Download our factsheet to discover how InfoSum’s decentralised Data Clean Room empowers brands to build custom audiences based on attribute level insight to deliver true people-based marketing campaigns.

Redefining the Data Clean Room for the privacy-first era

Data Onboarding

Build data co-ops