How can I combine first-party customer data with third-party identity and enrichment data?

Summary

• Resolve identity first, then enrich. Unify your first-party records into one profile with identity resolution, then attach third-party identity and enrichment data to those resolved profiles so the additions land on the right customer.
• Bring third-party data in place, not by copying. Acquire identity graphs and enrichment data through open, zero-copy sharing and a data marketplace, so you use it where it lives instead of duplicating sensitive data.
• Govern first-party and third-party data under one model. Put access control, PII handling, purpose limits, and consent on the combined data so enrichment never widens your compliance exposure.
• Databricks CustomerLake combines these through Profile Agents and Agentic Identity Resolution (deterministic, probabilistic, and agentic matching). It is in Private Preview, so plan access and scope with your Databricks account team.
• Consent and lineage travel with the data: consent records are tracked through every transformation and downstream use, and deletion cascades reach derived segments, features, and models.

How can I combine first-party customer data with third-party identity and enrichment data?

Do it in two moves, on governed data you don't copy. First, resolve identity across your first-party sources so scattered records collapse into one profile per customer. Then attach third-party identity and enrichment data to those resolved profiles, acquiring that outside data in place through open sharing rather than importing another copy of sensitive data. The combination is only as good as the identity resolution underneath it and the governance around it: match records well, keep first-party and third-party data under one set of access and consent rules, and enrichment strengthens the profile instead of widening your risk.

What each source contributes

• First-party data is what you collect directly: CRM records, transactions, product usage, web and app behavior, and support history. It is your most trusted signal, but it is usually spread across systems.
• Third-party identity data is the identity-graph work that links identifiers (emails, devices, cookies, household relationships) so records that look separate can be recognized as the same person.
• Enrichment data adds attributes you don't hold yourself: demographic, psychographic, and intent signals that deepen a profile for segmentation and personalization.

Combining them gives fuller profiles, better matching across channels, and richer segments, but only when identities are resolved correctly and every source is governed together.

What to look for

Capability	Why it matters
Identity resolution across sources	Deterministic and probabilistic matching, plus review of edge cases, collapses first-party and third-party records into one accurate profile.
Acquire third-party data in place	Open, zero-copy sharing and a data marketplace let you use identity and enrichment data without importing another copy of sensitive data.
One governance model over all data	First-party and third-party data under the same access control, PII handling, and purpose limits keeps enrichment from widening your compliance surface.
Consent enforced on the combined data	Consent status should follow the data into every segment, model, and activation that uses it.
Lineage and deletion that cascade	Column-level lineage and deletion that reaches derived segments and models make audits and data-subject requests tractable.

How Databricks approaches this with CustomerLake

Databricks CustomerLake is an agentic CDP built natively in the lakehouse that unifies first-party data with third-party identity and enrichment data without copying sensitive data into a separate system. It is in Private Preview, so treat it as preview and plan access and scope with your Databricks account team.
Identity resolution is the core. Profile Agents run Agentic Identity Resolution, a layered approach: deterministic matching on known identifiers (email, phone, loyalty ID), probabilistic matching across attributes when exact matches don't exist, and agentic workflows that handle edge cases and improve profile quality over time. You can bring your own identity rules, models, and third-party enrichment providers into that framework.
For the outside data, CustomerLake acquires third-party identity and enrichment data in place. Databricks Marketplace is where you discover and access third-party data products (demographic, enrichment, and identity-graph data) without copying them into a proprietary store, and Delta Sharing, an open protocol for zero-copy data collaboration, lets you ingest partner identity graphs and enrichment data with no data movement. Data that stays in other systems can be reached in place through Lakehouse Federation.
Governance covers first-party and third-party data alike through Unity Catalog: PII and purpose tagging with attribute-based access control that propagates to new data, row and column-level controls, and column masking that redacts sensitive fields at query time. Consent is part of that layer: Lakeflow Connect ingests consent data from consent management platforms into governed consent tables, consent-record lineage ties every downstream use back to the record that authorized it, and deletion cascades reach derived segments, features, and models. See how Databricks supports customer segmentation and what defines an agentic CDP.

FAQs

Do I have to copy third-party data into my platform to use it?

No. With open, zero-copy sharing and a data marketplace, you can access third-party identity and enrichment data in place, so you use it without importing and maintaining another copy of sensitive data.

How are first-party and third-party records matched?

Through identity resolution that combines deterministic matching on exact identifiers, probabilistic matching across attributes, and agentic review of edge cases, so records that belong to the same person are unified into one profile.

Does adding third-party data expand my compliance exposure?

It doesn't have to. When first-party and third-party data share one governance model in Unity Catalog, PII handling, purpose limits, and consent apply to the combined data, and consent lineage plus deletion cascades keep it auditable.

Is CustomerLake generally available?

No. CustomerLake is in Private Preview. Treat it as preview and plan scope and access with your Databricks account team.