What is a data lakehouse for marketing?

Summary

• A data lakehouse for marketing holds customer, transaction, behavior and engagement data in governed tables on one data platform, so analytics, machine learning and campaign activation can work from the same data.
• Databricks defines a data lakehouse as "a data management system that combines the benefits of data lakes and data warehouses." Its lakehouse uses Delta Lake for ACID transactions and schema enforcement, and Unity Catalog for governance and lineage.
• A lakehouse is the storage and modeling layer of a CDP, not the whole CDP. In Databricks' description of a composable CDP, data collection and data activation are separate layers that can each use a different product.
• CustomerLake is the Databricks agentic CDP, embedded natively in Databricks. Databricks announced it in Private Preview on June 16, 2026.
• Unity Catalog governs access to the tables marketing teams query, using privileges, attribute-based policies, and row filters and column masks.

What is a data lakehouse for marketing?

A data lakehouse for marketing is a lakehouse used as the home for customer and campaign data. Databricks defines a data lakehouse as "a data management system that combines the benefits of data lakes and data warehouses." For marketing, that means customer events, transactions and attributes sit in governed tables that analysts, data scientists and marketing tools can all work from.
The lakehouse is the foundation layer of a marketing data stack. In Databricks' description of a composable CDP, it covers data storage and modeling, while data collection and data activation are separate layers.

What customer data goes in a marketing lakehouse?

Databricks' CustomerLake materials refer to "customer, product and operational data" and to "campaign and engagement data." Common sources include:

• CRM records such as accounts, contacts and interaction history
• Ecommerce and transaction data, including orders and cart events
• Digital behavior from websites and apps
• Campaign and engagement data from email, advertising and messaging tools
• Consent and preference flags that define who can be targeted

What capabilities should a marketing lakehouse have?

Capability	Why it matters	Where Databricks documents it
ACID transactions	Updates land atomically, so readers don't see partial writes	Delta Lake supports ACID transactions
Schema enforcement and evolution	New customer attributes can be added without necessarily rewriting downstream logic	Delta schema evolution
Fine-grained access control	Row filters and column masks limit what each user sees	Unity Catalog
Lineage and auditing	Shows where data came from and how it was used	Unity Catalog lineage and the audit log system table
Access to data in place	Reaches customer data in other systems without migrating it	Lakehouse Federation
Identity resolution	Links records for the same person into one profile	CustomerLake Agentic Identity Resolution
Audience activation	Syncs audiences to email, advertising and engagement tools	CustomerLake managed reverse ETL connectors

How does Databricks approach the marketing lakehouse?

Databricks builds the lakehouse on Delta Lake, "an optimized storage layer that supports ACID transactions and schema enforcement," and Unity Catalog, "the unified governance layer for data and AI built into Databricks." For data that lives elsewhere, Lakehouse Federation provides "governed, read-only access to external data through Unity Catalog foreign catalogs."
On top of that foundation sits CustomerLake, an agentic CDP embedded natively in Databricks. It brings Customer 360, identity resolution, audience building, campaign automation, activation and personalization into the lakehouse. Profile Agents turn raw customer data into business-ready profiles, Campaign Agents help marketers build audiences and automate campaigns, and Genie lets marketers ask questions and build audiences in natural language. Native managed reverse ETL connectors sync audiences and signals to martech, adtech and customer experience tools. Databricks says CustomerLake inherits security, lineage and access controls from Unity Catalog. It launched in Private Preview.

How do a lakehouse, a composable CDP and an embedded CDP relate?

Databricks describes a composable CDP as a pattern that lets an organization pick a product for each layer of the CDP: data collection, data storage and modeling, and data activation. In Databricks' example, the lakehouse is the storage and modeling layer, and other products handle collection and activation.
Databricks positions CustomerLake differently, as an agentic CDP embedded in Databricks, so the CDP functions run in the same platform as the data.

What are common marketing lakehouse use cases?

RFM segmentation. A Databricks blog shows recency, frequency and monetary (RFM) metrics "derived from transactional data residing in the Databricks Lakehouse" segmenting customers into groups with straightforward machine learning techniques.
Audience activation. With CustomerLake, teams sync audiences and signals from Databricks to the marketing, advertising and customer experience tools they already use.
Natural-language audience building. Marketers use Genie to ask questions, get recommendations and build audiences from governed customer data without writing SQL.

FAQs

Does a marketing lakehouse replace a CDP?

Not by itself. In Databricks' description, a composable CDP has collection, storage and modeling, and activation layers, and the lakehouse covers storage and modeling. You add the other layers with separate products in a composable CDP, or use CustomerLake, which embeds CDP functions in Databricks.

How is customer data secured in a lakehouse?

Unity Catalog manages access with privileges, attribute-based policies, row and column filters and workspace bindings. It also tracks lineage automatically and records access in an audit log system table.

What is a composable CDP?

A pattern in which an organization picks a product for each CDP layer (collection, storage and modeling, activation) around a shared data foundation such as a lakehouse.

Is CustomerLake generally available?

Databricks announced CustomerLake in Private Preview on June 16, 2026. Ask your Databricks account team about current access.