EF Education First Scores dbt Models to Rein In Snowflake Costs

EF Education First built a scoring system on dbt and Snowflake to bring unmanaged cloud data costs under control, Adrien Ledoux, who leads the company's data platform team, said at a dbt Summit 2026 breakout session.

The shift from a legacy SQL Server setup, where spend was fixed, to Snowflake's consumption-based compute gave domain teams the freedom to run pipelines on their own schedules - but every dbt model run now carried a real cost. Costs climbed steadily, some pipelines had little downstream usage, and a few expensive projects had no clear owner.

Ledoux's team scored every dbt model and test on execution time, credits consumed, and rows processed per second, normalized as a percentile rank against similarly sized peer models. Top-tier models are highlighted as success stories, mid-tier ones get periodic review, and bottom-tier ones trigger a conversation with the owning domain team.

The results were measurable: some of EF's heaviest pipelines saw a 40 percent efficiency improvement, one of the biggest models dropped from 10 minutes to 2 minutes, and some Power BI workspaces saw up to 70 percent of reports deleted once low usage became visible.

Ledoux said the biggest ongoing challenge is adoption of the Cortex Agent itself, and drew a sharp distinction between usage and value: a report checked a thousand times isn't necessarily valuable, while one checked weekly by the right person can be. His closing warning: a good technical idea does not adopt itself.

Source: dbt Labs Blog