Fivetran expands beyond connectors
Charles Chretien, co-founder of Prequel, on the modern data stack’s ROI problem
The pressure on Fivetran’s core business is pushing it from a connector company toward a broader data platform. The weak point is not that companies stop needing data moved, it is that the most valuable pipelines are increasingly being claimed by the warehouse itself on the database side and by SaaS vendors on the application side. That makes pure extraction and loading less defensible, and makes bundling adjacent workflows like transformation and activation the logical next step.
-
Fivetran’s original edge was building and maintaining trusted connectors so customers did not have to keep fixing broken API based pipelines themselves. That model still matters, but it is costly to maintain and easiest to attack where a few very large connectors drive a lot of sync volume and revenue.
-
The two concrete attackers are warehouses and source apps. BigQuery and other warehouses can handle database replication natively, while vendors like Stripe, Salesforce, and Customer.io can ship their own warehouse exports. When that happens, Fivetran often loses the highest frequency, highest volume workloads first.
-
That helps explain the acquisition path. Census adds downstream activation, dbt adds transformation, and together they turn Fivetran from a tool that only lands raw data into one that can help shape it and push it back into business systems. The strategy is to own more of the workflow around the warehouse, not just the pipe into it.
The next phase is a fight over who owns the control point around warehouse workflows. Fivetran is likely to keep expanding from ingestion into transformation, activation, and potentially the compute layer itself, because the winning product in this market is less likely to be the best standalone connector and more likely to be the default operating surface for how data moves and gets used.