Third-party data

Third-party data is information about people or audiences collected by a company that has no direct relationship with those people, then aggregated and sold or licensed to advertisers, often as ready-made segments.

The IAB Tech Lab defines a third party as a party with no direct relationship with the individual involved (IAB Tech Lab, Data Clean Rooms Guidance, v1.0). Segments such as "in-market for an SUV" are built from browsing across sites, purchase records, surveys or models, and sold through DMPs, DSP marketplaces and data exchanges. In OpenRTB, segments travel in the user.data array.

The weakness is opacity. Buyers often cannot tell how a segment was built, how old it is or how much of it is modelled. Industry groups responded with the Data Transparency Label, a "nutrition label" for audience data released as a beta in 2018 by the IAB Tech Lab and marketing trade groups. It has four parts: provider details, audience construction, an audience snapshot and source information (DataLabel.org, IAB Tech Lab GitHub). Third-party cookies and privacy laws now limit how such data can be collected and used.

Example: A DSP marketplace lists "Outdoor enthusiasts, 12 million IDs". Its label shows a 90-day lookback, cross-device expansion and that 60% of the segment is modelled from a seed of 4.8 million. The buyer decides the modelled share is too high for a launch that needs precise targeting.

Related terms: second-party data, DMP, contextual targeting, cookie sync, lookalike audience.

Authoritative source: IAB Tech Lab and partners, DataLabel.org "Data Transparency Label" (beta, 2018; source on GitHub, accessed 9 November 2026).

Read the full guide: DSP, DMP and external data in explainable RTB strategies.