Fivetran MAR: The Invoice That Doubled, and the Credits You Still Pay
Monthly active rows surprise you, warehouse writes have their own bill, a custom extractor can beat a connector, and Fivetran still does not replace dbt.
- fivetran
- elt
All stacks · Ingestion & ELT
Managed ELT with schema drift handling and usage-based pricing.
Monthly active rows surprise you, warehouse writes have their own bill, a custom extractor can beat a connector, and Fivetran still does not replace dbt.
Fivetran is managed ELT: connectors that handle schema drift, incremental updates, and a monthly invoice tied to monthly active rows. It is the default “stop writing extractors” purchase in mid-size companies.
Cost and ownership are the production topics — MAR surprises, destination credit usage, and still needing dbt after the load.
For long-tail SaaS APIs, almost always. For one Postgres CDC you already know how to run, Debezium + Connect can be cheaper at scale. Model the MAR and the engineer-hours, not the logo.
Light transformations exist; they are not your metrics layer. Land raw/staging, model in dbt. Mixing business logic into the connector is how you cannot test grain.
Row counts vs source, lag, and schema diffs. A successful sync of zero rows is a silent outage.
One email a week with our latest tutorials. No spam.
Newsletter signup is not live yet. Use the contact form if you want to be notified.