Why Fivetran’s Pricing Change is a Wake-Up Call for Data Teams

Hey everyone,

I’ve been seeing a lot of chatter in professional circles lately about Fivetran’s pricing. It seems like the March 2025 and January 2026 updates have sent shockwaves through the community. I’ve been looking closely at the numbers, and the impact is undeniable. We’re talking about a significant spike in monthly costs for many data teams, with some seeing their bills more than double.

So, what exactly is happening?

For those who might not be deep in the weeds of vendor billing, Fivetran has shifted its pricing model. Previously, they calculated costs based on Monthly Active Rows (MAR) at the account level. This allowed for a form of bulk discounting; a large data pipeline could effectively subsidize smaller, less active connectors. Now, they’ve moved to a per-connector model. This means every single connection is evaluated on its own MAR, and the tiered discounts apply only to the volume within that specific connector.

If you have a dozen connectors each processing less than a million MAR per month, you’re no longer benefiting from a cumulative discount. You’re paying the higher rate for each one individually. The result is a pricing structure that can feel exponential for teams with diverse, lower-volume data sources.

Fivetran’s official explanation, as outlined in their documentation, is that this change aligns cost with infrastructure and operational reality. Their argument is that a low-volume connector still requires setup, monitoring, support, and compute resources, and the old model didn’t reflect that. They’re essentially saying they can no longer afford to subsidize smaller workloads at the expense of larger ones.

This brings us to the core question: What is the long-term play here? Is this a sustainable move for a company in the data engineering space, or are they pushing their luck? From what I’m seeing, many data professionals are starting to seriously evaluate their reliance on this single vendor. The conversation is shifting from “How do we manage this cost?” to “What are our alternatives?”

The high switching costs are a real barrier, but the financial incentive to leave is growing. I’ve seen comments suggesting that building an in-house solution could pay for itself in a matter of months, not years. The ROI calculations are starting to look very compelling for teams that have the technical bandwidth.

This situation highlights a critical lesson in modern data architecture: the danger of vendor lock-in. When a single tool becomes central to your data pipeline, you become vulnerable to pricing changes that are outside your control. We’re seeing a renewed interest in open-source frameworks and more modular approaches to data ingestion. Tools like Airflow, Prefect, and dlt are being re-examined not just for their technical capabilities, but for the financial control and flexibility they offer.

The shift also sparks a debate about the nature of data engineering as a commodity. Are we paying for a unique, high-value service, or are we paying a premium for a managed wrapper around standard APIs? As the market matures and competition increases, it’s natural to expect prices to trend downward, not up. This recent change feels like a direct contradiction to that trend.

I don’t have a crystal ball to predict Fivetran’s future, but I can read the room. The sentiment is turning. Data teams are more price-conscious and wary of lock-in than ever before. This pricing update might be the catalyst that forces a healthy re-evaluation of our tool stacks and a move towards solutions that offer more predictability and control.

If you enjoyed this article, then please consider sharing with other data professionals, and subscribing for more insightful, entertaining, and informative newsletters.