Most customer data platforms start with a demand: hand over your data, and we will store a copy of it in our system. GrowthLoop, a martech company founded in 2015 as Flywheel Software, built its business around the opposite idea. Its platform runs directly on top of the cloud data warehouse a company already owns - Snowflake, Google BigQuery, Databricks or Amazon Redshift - so segmentation, orchestration, measurement and AI all execute against live, governed warehouse data. There are no ingestion pipelines, no duplicate storage, no vendor-managed copy of your customers.
That architecture is called a "composable" CDP, and GrowthLoop is one of the loudest voices in the category. The pitch is aimed squarely at marketers rather than engineers: instead of filing a ticket and waiting for a data team to write SQL, a marketing manager drags and drops attributes in a visual audience builder and the query runs where the data lives. The company describes the whole system as a "Compound Marketing Engine" - the notion that every campaign and experiment should feed the next, so results compound instead of resetting each quarter.
In April 2025 GrowthLoop leaned into agentic AI, unveiling a set of "Growth Agents" that propose audiences, build journeys, and flag optimizations for a human to approve. It is a deliberately supervised flavor of automation: the agents suggest, the marketer signs off. For enterprises nervous about handing campaigns to a model, that review step is the point.
