Your first-party data knows what a customer did on your site. It rarely knows who they are, how old they are, or whether their household is about to buy. Data enrichment fills that gap by appending attributes from outside suppliers—and the same process has been running inside programmatic auctions for years.
Offline, a client sends match keys such as email addresses, phone numbers or name-and-address combinations, usually normalised and hashed. The provider matches them against a reference database and returns fields with confidence grades. Inside a bid request, OpenRTB’s Data and Segment objects have carried provider segments since version 2.0 in 2011, and Prebid’s Real-Time Data modules add more before a buyer ever sees the impression.
What you are actually buying
Enrichment arrives in three forms: batch files for direct mail, real-time APIs for B2B workflows, and clean rooms where two parties join data without moving it. Clearbit built the API model and, when HubSpot agreed to buy it in November 2023, claimed coverage of 20 million companies and 500 million decision-makers. HubSpot later folded the unit into Breeze Intelligence, launched in September 2024.
The input has always been the weak point. In a September 2026 Firmable survey, 222 B2B sales professionals estimated that 32% of their CRM records were inaccurate, incomplete or outdated—the same lists that end up in ad platforms for matching.
Why accuracy should change how you buy
- An independent Marketing Science study of more than 90 third-party audiences from 19 data brokers found gender accuracy averaged 42.3%, below random guessing.
- Truthset’s 2026 report put age segment accuracy at just 13%.
- RTB House sampled about 135 billion requests enriched with seller-defined audiences and found 75% of user signals non-compliant.
That does not make enrichment useless. As match rates fall between platforms—Equativ put the loss at 40% to 70% in 2024—marketers lean on enriched data to stay addressable. Google’s data strength framework even names signal enrichment as an early step before automated bidding can price auctions.
Legal pressure is uneven but growing
GDPR Article 14 requires notice when personal data is processed without being collected from the individual. Poland tested the rule in 2019 with a fine of about 220,000 euros against Bisnode. Italy fined Lusha 2 million euros in 2026 and rejected legitimate interest as a basis for selling enriched contact profiles. HubSpot’s plan to pool customer CRM data into a shared enrichment dataset was withdrawn within days in July 2026 after customer backlash.
The supply side is still moving. In July 2026, Nexxen added Acxiom household and purchase data to Nexxen Discovery, and Acxiom launched an identity accelerator with Salesforce—both without published match-rate benchmarks.
Play it like a buyer, not a believer
Treat enrichment as an unverified input. Demand match-rate and accuracy benchmarks, ask which taxonomy or attribute is being appended, and confirm that the legal basis covers people who never dealt with the supplier. If a vendor will not publish match rates, price that silence as risk.
Source: PPC Land
