Third-party data
Also: 3rd-party data, third party data, purchased data, brokered data, aggregated audience data, donnees tierces
Data purchased from external aggregators, collected from audiences you don't own. It is bought or licensed rather than gathered through your own direct relationships.
What it is
Third-party data is information about individuals or organizations that you did not collect yourself and that comes from sources with which you have no direct relationship. It is typically bought or licensed from data brokers, aggregators, or exchanges that gather signals across many websites, apps, and offline sources, then package and sell it.
It is distinguished from two related categories:
- First-party data: collected directly from your own customers and audiences (purchases, logins, site behavior).
- Second-party data: another company's first-party data, shared or sold to you directly, without an aggregator in the middle.
Third-party data is defined by the absence of a direct relationship with the person the data describes.
Why it matters
- Scale and reach: it lets you target or model people beyond your existing customer base.
- Enrichment: it appends attributes (demographics, firmographics, interests, intent signals) to records you already hold.
- Declining reliability: browser restrictions on third-party cookies, mobile identifier limits, and privacy laws (GDPR, CCPA) have reduced its accuracy and legality.
- Compliance risk: because consent was not obtained by you, provenance and lawful basis can be hard to verify.
How it is used in practice
- Audience targeting in advertising platforms and demand-side platforms.
- Lead scoring and enrichment in B2B sales and marketing.
- Credit and risk models in finance, using bureau and alternative data.
- Market sizing and competitive benchmarking.
A common technique is a data match: you upload hashed identifiers (emails, device IDs), the vendor matches them against its pool, and returns appended attributes or a lookalike segment.
Concrete worked example
A retailer wants to reach likely luxury car buyers. It licenses a segment from a data aggregator labeled "in-market: premium automotive" covering 4 million cookies or device IDs. The retailer pushes this segment into its ad platform and runs a campaign.
- Match rate: only 55 percent of IDs are still active, so effective reach is about 2.2 million.
- Accuracy: the aggregator inferred intent from browsing, so some users are researchers, not buyers.
- Cost: the segment is priced per thousand impressions (CPM) on top of media cost.
The retailer measures conversions against a control group and finds the segment underperforms its own first-party purchase data, a typical outcome that pushes many organizations back toward owned data.
See also
Frequently asked questions
What is third-party data?
Third-party data is information about people or companies that you did not collect and that comes from sources with no direct relationship to you, typically bought or licensed from data brokers, aggregators, or exchanges. These vendors gather signals across many websites, apps, and offline sources, then package and sell them as segments or attributes.
What is the difference between first-party, second-party and third-party data?
First-party data is collected directly from your own customers and audiences (purchases, logins, site behavior). Second-party data is another company's first-party data shared or sold to you directly, without an aggregator in between. Third-party data is defined by the absence of any direct relationship with the person it describes.
Why has third-party data become less reliable?
Browser restrictions on third-party cookies, limits on mobile identifiers, and privacy laws such as GDPR and CCPA have cut both its accuracy and its legal usability. There is also a compliance risk: since consent was not obtained by you, provenance and lawful basis are hard to verify.
How does a data match with a vendor actually work?
You upload hashed identifiers such as emails or device IDs, the vendor matches them against its own pool, and returns appended attributes or a lookalike segment. It is the standard mechanism behind enrichment and lead scoring in B2B, and behind audience extension in advertising platforms.
What kind of results should I expect from a purchased audience segment?
Expect real shrinkage between the size sold and the reach obtained. In a typical case, a retailer licenses an "in-market: premium automotive" segment of 4 million cookies or device IDs, but only 55 percent are still active, giving about 2.2 million in effective reach, with a CPM charged on top of media cost. Intent inferred from browsing also mixes buyers with researchers, which is why such segments often underperform a company's own first-party purchase data when measured against a control group.