← All terms·Cross-media measurement
Reference library · Cross-media measurement
Data Clean Room
A data clean room is a secure, permissioned environment where an advertiser and a platform (or two advertisers) each upload first-party audience data and run joint queries without either side seeing the other's raw records. Matching runs on hashed identifiers such as email or device ID, and only aggregated outputs (overlap, reach, frequency) leave the room. It lets a planner measure cross-platform reach and frequency against a CRM base without handing that base to the platform.
Worked example
An advertiser uploads a CRM segment of 620,000 hashed customer records to a clean room shared with a VOD platform. The platform matches it against its 3.1 million logged-in viewers and returns an aggregated report: 145,000 matched users, average frequency 3.6, and 58,000 of the platform's reached households falling outside the TV campaign's reach. The planner shifts part of the digital budget toward that 58,000-household gap instead of buying more frequency on an audience TV already covers.
How it is used
Clean rooms are used mainly for three things: overlap/incrementality reports across TV and digital, suppression of existing customers from prospecting campaigns, and joint modeling of conversion lift. The recurring mistake is reading a low match rate as low audience overlap: match rates commonly land at 50-70% because of hashing mismatches (login email vs. cookie vs. device ID), not because the audiences don't overlap.
The common mistake
Don't read an unmatched or low-overlap result as proof the audiences don't overlap; it's usually an identity-matching gap between hashed keys, not a real absence of shared reach.