What is a data clean room?

A data clean room is a controlled environment where organisations analyse data together while restricting access to the underlying records. It allows participants to answer agreed questions without simply handing over their customer databases or exposing the underlying records to each other.

For example, a retailer and a publisher might measure how many customers purchased after seeing a campaign. The clean room can restrict which datasets may be compared, which queries may run and what results participants receive.

Depending on the platform and configuration, safeguards may include:

• Approved queries and restrictions on accessible data fields.

• Minimum group sizes before results can be released.

• Controls over who can run analyses and receive outputs.

• Restrictions designed to prevent results from revealing information about individual people.

• Differential privacy to reduce what results reveal about individuals.

AWS Clean Rooms’ analysis rules illustrate how these restrictions can be implemented.

A clean room can support privacy-preserving analytics, but it does not automatically anonymise data or provide permission to use it. Participants must still establish appropriate purposes, legal grounds and sharing arrangements. The protection depends on the actual controls, including whether repeated queries or released results could reveal information about individuals.