Keep Your Own Map With Reverse ETL
Sean Lynch, co-founder of Census, on reverse ETL's role in the modern data stack
The real advantage of the warehouse first approach is that it lets a company keep its own map of the business instead of squeezing into a vendor's preset map. In practice, that matters when a company has unusual entities like workspaces, subscriptions, or marketplace participants, or needs long tail connectors beyond the default ad and CRM stack. A classic CDP is easiest when the business looks like a standard ecommerce funnel. Reverse ETL is stronger when the data model is the product.
-
CDPs usually bundle ingestion, identity, modeling, and activation into one system. That makes setup simple, but it also means the vendor decides what a user, account, or group looks like, and customers inherit those rules across every downstream workflow.
-
Census was built around the opposite assumption. Data lands in Snowflake, BigQuery, or Redshift, gets shaped in SQL and dbt, then syncs out to tools like Salesforce, Braze, or support systems. That means the same warehouse definition can drive board reporting, sales alerts, and marketing audiences.
-
The broader market kept moving this way. Hightouch also leaned on a schema agnostic model that does not require storing customer data in a fixed vendor schema, while Customer.io positioned its CDP around working with many destinations instead of trapping users in one messaging stack.
Over time, the winning products in customer data look less like closed databases and more like control planes on top of the warehouse. As more teams want real time product, sales, and marketing signals built from the same underlying tables, flexible activation and custom objects become the default, and rigid all in one schemas lose ground.