Calixa Collapsing Reverse ETL
Thomas Schiavone, co-founder and CEO of Calixa, on the PLG data pipeline
This reveals Calixa was trying to collapse part of the reverse ETL layer into the application itself, because owning the warehouse connection lets it shape the sales workflow instead of just accepting whatever records another sync tool delivers. In practice, a team could keep Census or Hightouch as the central pipe that moves warehouse data into many business apps, while using Calixa’s native Snowflake connection to define product qualified accounts, custom objects, and rep workflows inside a PLG CRM built around usage data.
-
Census and Hightouch are infrastructure products. They read tables or SQL from the warehouse, map fields, and sync those records into destinations like Salesforce, ad tools, and support systems. That makes them a shared data plumbing layer for many teams, not a purpose built sales application.
-
Calixa was aiming at the layer above that plumbing. The product was built so non technical go to market users could work from warehouse modeled data without writing SQL, and so setup looked more like mapping business objects in a UI than building bespoke pipelines. That is why direct connectivity mattered so much.
-
Using both made sense when the data team wanted one control point for all syncs. Census or Hightouch could remain the company wide sync engine, while Calixa consumed the same warehouse directly for its own opinionated account views and pushed selected objects back out as needed. That is a classic platform plus app split.
The direction of travel was toward warehouse native business software. As more PLG tools connected straight to Snowflake, BigQuery, and Databricks, reverse ETL vendors expanded into activation and audience products, while apps like Calixa tried to absorb more of the operational workflow themselves. The boundary between data pipe and end application was set to keep blurring.