Glossary

A

Advanced Analytics DCR - Environment designed to enable collaborative data analysis where the use of data can be precisely controlled by the data owners.

AI lookalike audiences - Machine learning model that identifies users in the base audience most similar to users in the seed audience.

B

Base audience - Audience data provisioned to a Media DCR by a publisher or data partner in the form of a Datalab. Any Media DCR requires the provisioning of base audience and a seed audience to function.

C

Computation - SQL queries, Python scripts, or R scripts that can be performed on the data.

Confidential Computing - Technology that prevents all involved parties - including Decentriq and the cloud provider - from unauthorized access to data.

D

Data Analyst - Role that can run pre-configured computations and retrieve results.

Data Owner - Role responsible for contributing data and reviewing DCR configuration.

Datalab - Validated collection of datasets that serves as a base audience for Media DCR participation. It contains at minimum a matching table, optionally a segment table.

Demographics Dataset - Maps user IDs to age brackets and gender.

Deprovisioning - Removing provisioned data from a DCR.

E

Embeddings Dataset - Maps user IDs to numeric embeddings for AI lookalike modeling.

Enclave - Verifiable confidential-computing environment where all processing happens.

H

Hashed Email - SHA-256 hash of an email address. Most commonly used as matching ID.

I

Insights - Statistics about the seed audience in terms of base audience attributes.

M

Match Rate - Percentage of seed audience users that overlap with the base audience.

Matching Dataset - Maps internal user IDs to matching identifiers such as hashed email.

Matching ID - Identifier used to join the base and seed audience (e.g., hashed email, phone number, string).

Media DCR - No-code workspace where collaborators can safely combine first-party data to gain privacy-preserving insights and create target audiences.

O

Overlap - Users matching between the seed and base audiences.

P

Provisioning - Making data available in a DCR.

R

Remarketing audiences - Overlap users between the seed audience and the base audience, typically used to re-engage or remarket existing clients from the seed audience provider.

Rule-based audiences - Combine audiences with inclusion and exclusion criteria to build a custom audience.

S

Seed audience - List of users that are matched against the base audience.

Segments dataset - Maps user IDs to segment labels for insights and targeting.