Changes / ETL / ELT & ingestion / Fivetran
Billing model changemedium impact
Fivetran deletes now count toward paid MAR
Inserts, updates and deletes all count toward paid MAR; previously deletes did not contribute to paid MAR.
Who’s affected
Fivetran pay-as-you-go customers (already in effect since 2026-01-01) and annual-contract customers who signed after March 1, 2025, from their next renewal. Teams with delete-heavy sources (CDC, soft-delete tables, frequent purges or reloads) are hit hardest.
Not affected: Annual-contract customers who signed before March 1, 2025 stay on pre-March 2025 rules, where deletes do not count toward paid MAR.
What it costs
Deletes now add to paid MAR alongside inserts and updates, so billed MAR can rise with no change in your data volume. The record gives no prices or percentages, so the size of the increase depends on how many deletes your connections process.
When to act
Renewal-driven billing changes need time to measure delete volume and negotiate terms before the contract date; PAYG customers are already exposed and should start now.
Blast radius: Monthly Fivetran invoices and budget forecasts; any usage alerts tied to MAR; renewal negotiations. Pipelines and downstream models are unaffected unless you migrate tables off Fivetran.
What should you do? decision tree
- you are on PAYGThe change is already live; pull recent usage per connection and compare paid MAR against pre-2026 months to find the delete-driven increase.
- you are on an annual contract signed before March 1, 2025No change now; ask what rules apply at renewal and get that in writing before signing.
- you are on an annual contract signed after March 1, 2025The new metering starts at renewal; estimate delete-driven MAR now and use it in renewal talks.
- deletes are a large share of your row changesReduce delete churn at the source (avoid full purges and reload patterns) or move those tables to another tool.
- estimated impact is small or you can't measure it yetSet a MAR budget alert, monitor for one or two billing cycles, and revisit at renewal.
Options
- WaitIf deletes are rare in your sources, accept the change and monitor paid MAR for a couple of billing cycles.
- OptimizeCut delete volume at the source: avoid truncate-and-reload and bulk purges, and exclude high-churn tables from sync.
- NegotiateAt renewal, ask for delete-heavy connections to be capped or excluded, or for a list-price discount that offsets the new metering.
- MigrateMove delete-heavy or CDC-heavy pipelines to a tool with different pricing, such as a usage- or capacity-based alternative.
Alternatives editorial picks, not sponsored
The facts straight from the source
Facts above are extracted from the source and checked word-for-word before publishing. The guidance sections are AI-assisted analysis based only on those facts: a starting point, not legal or procurement advice. Spotted an error? Tell us.
Related changes
- FivetranFivetran moves MAR volume tiering from account to connection level
MAR volume tiering is applied per connection instead of per account from March 1, 2025, raising costs for accounts with many equal-sized connections.
Price ↑ - FivetranFivetran signs agreement to acquire Census (reverse ETL)
Fivetran announced a signed agreement to acquire Census, funded by cash and equity; terms were not disclosed and closing is pending.
M&A - FivetranFivetran acquires Tobiko Data (SQLMesh, SQLGlot)
Fivetran announced it has acquired Tobiko Data, the open source company behind SQLMesh and SQLGlot, adding transformation capabilities to its platform.
M&A - FivetranFivetran adds a $5 minimum per connection (1 to 1M MAR)
A $5 minimum base charge applies to standard connections generating between 1 and 1 million MAR per month, starting January 1, 2026 for PAYG.
Price ↑ - FivetranFivetran now bills history-mode repeated updates as paid MAR
Repeated updates tracked in history mode now count toward paid MAR; soft delete and live modes are alternatives.
Billing - FivetranCensus becomes Fivetran Activations, billed on MAR
Activations is billed on inserted, updated and deleted data measured in MAR, with cost per MAR decreasing on each activation's consumption curve.
Packaging