Warehouses Becoming the New CDPs

This shift means the customer data layer is moving from a packaged app into the warehouse itself. Traditional CDPs bundled data collection, identity stitching, audience building, and activation in one tool. Reverse ETL vendors and PLG tools now let teams keep raw data in Snowflake or BigQuery, define segments in SQL, and push those audiences into Salesforce, Braze, ads platforms, and support tools, which recreates a big part of the CDP job in a more flexible way.

• Segment’s original power came from being the universal event router. A developer dropped one snippet into the app, then turned on destinations like Mixpanel, Intercom, or a warehouse. That made Segment the easiest way to collect events and fan them out everywhere.
  1 sacra 4 sacra
• Reverse ETL changed the center of gravity. Instead of storing customer truth inside the CDP, teams ingest data into the warehouse with tools like Fivetran, model it with dbt, then use Census or Hightouch to sync user traits, accounts, lists, and scores back into downstream apps.
  2 sacra 4 sacra 5 sacra
• What CDPs still keep is the event pipeline. Even in this warehouse first world, Segment remains strong for real time event tracking and delivery. The overlap is biggest in targeting and audience sync, not in the developer instrumentation layer where Segment built its moat.
  1 sacra 2 sacra 3 sacra

Going forward, more customer software will split into two layers. The warehouse will hold the canonical customer model, while apps compete on who best turns that model into messages, sales actions, and product workflows. That favors tools that plug cleanly into warehouse data over tools that require teams to rebuild customer truth inside a separate system.