How can I bring marketing activation closer to the data layer?

Summary

• Run activation where the governed data lives. Instead of exporting audiences to a separate stack, build audiences, decide next-best actions, and push to channels on the same governed customer data, so there is no extra copy to secure and no drift.
• Keep identity and profiles at the data layer. When Customer 360 profiles are built in the governed store, activation reads the same trusted data your analytics and AI use, not a stale export.
• Governance travels with the data. Access rules, PII protection, purpose limits, and consent are enforced on the underlying data, so activation stays inside policy.
• Databricks CustomerLake is an agentic CDP embedded in the lakehouse that works this way. It is in Private Preview, so plan access and scope with your Databricks account team.
• Marketers work in natural language through Genie: build audiences and run campaigns without filing a data-engineering ticket, while activation reaches downstream messaging, email, ad, and web channels without manual data movement.

How can I bring marketing activation closer to the data layer?

Bring the activation work to the data instead of shipping the data out to activation tools. Marketing activation moves closer to the data layer when audience building, decisioning, personalization, and the push to channels all run on the same governed customer data where it already lives, not on a copy exported into a separate platform. That collapses the gap where data goes stale, governance is lost, and marketing waits on engineering. The practical test is simple: does activation read the same governed profiles your analytics and AI use, under the same access rules, with no extra copy to maintain?

What "activation" and "the data layer" mean here

Activation is the step where customer data becomes action: you assemble an audience or segment, decide the next-best message and channel, and send it to the systems that reach customers (messaging, email, ad, and web or app channels). The data layer is the governed store where customer data actually lives: profiles, events, and transactions. In many stacks these are far apart. Activation happens in a separate tool that needs its own copy of customer data, so the audience a marketer activates is an export that can drift from the source and sits outside the governance applied to the original. Closing that distance is what "closer to the data layer" means.

What to look for

Capability	Why it matters
Activation on data in place	Running audience build and activation on the governed data avoids a second copy of customer data to secure and keep in sync.
Identity and profiles at the data layer	Customer 360 profiles built in the governed store mean activation uses the same trusted data as analytics and AI, not a stale export.
Governance that travels with the data	Access control, PII protection, purpose limits, and consent enforced on the underlying data keep activation inside policy.
Natural-language audience building	Marketers assemble segments and campaigns without a SQL request to engineering.
Reach to downstream channels	The platform pushes audiences to messaging, ad, and web channels without manual data movement.
Closed feedback loop	Campaign results land back on the same data, so the next decision improves.

How Databricks approaches this with CustomerLake

Databricks CustomerLake is an agentic CDP built natively in the lakehouse, so activation runs where your customer data already lives. It is in Private Preview, so treat it as preview and plan access and scope with your Databricks account team rather than as a shipping product.
The design keeps activation at the data layer. CustomerLake operates on data in place: CDP work runs on the governed data in the lakehouse, with no copy into a separate activation system, and governance is inherited from Unity Catalog, which enforces access control, PII protection, purpose limits, and consent. On top of that governed data it works through two agent modules. Profile Agents build governed Customer 360 profiles using agentic identity resolution that unifies scattered records into trusted profiles and improves over time. Campaign Agents let marketers build audiences, run campaigns, and personalize message and channel in natural language through Genie, using governed customer context to recommend next-best actions. From there CustomerLake sends audiences to downstream marketing and advertising channels without manual data movement, and where trusted customer data lives in other systems it can be reached in place through Lakehouse Federation. For the concept and design, see Introducing CustomerLake and what defines an agentic CDP.

A closed loop on one copy of data

Because activation sits on the governed data, the full cycle runs without exports: build the audience from governed profiles, activate to the channel, and let campaign results land back on the same data so the next decision is better informed. Keeping the loop on one governed copy is what makes activation measurable and repeatable rather than a one-way hand-off to a disconnected tool. See why unified data underpins agentic advertising.

FAQs

Do I have to copy customer data into a separate tool to activate it?

No. With an in-place model, activation runs on the governed customer data where it lives, so there is no separate export to build, secure, and keep in sync.

Is CustomerLake generally available?

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

How do marketers build audiences without engineering?

Through Genie, a natural-language interface. Marketers describe the audience or campaign in plain language, so building a segment does not require a SQL request to a data team.

How does governance stay intact during activation?

Governance is enforced on the underlying data through Unity Catalog: access control, PII protection, purpose limits, and consent apply to the profiles activation uses, so campaigns stay within policy.