dbt Freemium

SQL-based transformation framework that gives your warehouse version control and tests.

dbt is a transformation framework that turns plain SQL SELECT statements into tested, documented, version-controlled models inside your warehouse. dbt Core is open source and free; dbt Cloud starts around $100/mo for a team of five developers.

What is dbt?

Before dbt, analytics SQL lived in BI tool folders and scheduled query editors, untested and invisible to anyone outside the analyst who wrote it. dbt put that SQL in a Git repo, gave it a dependency graph, and let you run dbt test in CI.

Each model is a .sql file with a SELECT. ref('stg_orders') wires models together, and dbt works out the build order from those references. Works with Snowflake, BigQuery, Redshift, Postgres and Databricks.

Product Features

  • Ref-based dependency graph: dbt resolves build order instead of you maintaining numbered SQL scripts.
  • Built-in tests: unique, not_null, accepted_values, relationships — plus custom tests written in SQL.
  • Materializations: views, tables, incremental models that only process new rows.
  • Documentation site: auto-generated with column descriptions and lineage you can click through.
  • Jinja templating: macros and loops make 40 near-identical models into 1.
  • State-aware runs: rebuild only models affected by a change, which is the difference between a 4-minute and a 40-minute job.

Product Highlights

The real shift is cultural. Once a data model is a pull request, a reviewer can ask “why does revenue subtract refunds twice?” before it hits a dashboard the CFO reads.

  • dbt Core runs locally or in any scheduler, no vendor lock-in.
  • Lineage graph answers “what breaks if this column changes” in seconds.
  • Sources let you test raw tables you don’t own.
  • Python models are supported for ML feature prep alongside SQL.

Use Cases

  1. Building a staging layer that cleans raw event data before BI touches it.
  2. Defining revenue metrics once so finance and product stop arguing about numbers.
  3. Running dbt tests in CI to block a deploy that breaks a uniqueness constraint.
  4. Documenting column-level definitions for compliance or new analysts.
  5. Incremental event models that process only yesterday’s rows.

How much does dbt cost?

dbt Core is Apache 2.0 and free, including all testing and docs features. dbt Cloud Developer is free for one seat; Team is around $100/mo for five developers plus CI and the hosted IDE. Larger enterprise deals are quoted and usually cost more than the warehouse for small teams.

Who should skip dbt?

Companies with one analyst and ten tables. The framework’s value scales with model count and team size; with neither, you get ceremony. Teams with no SQL fluency will struggle, and anyone expecting dbt to handle ingestion is looking in the wrong place — that’s Airbyte or Fivetran’s job.

Quick Summary

  • dbt transforms warehouse data using version-controlled SQL models.
  • dbt Core is free and open source; Cloud is about $100/mo for five devs.
  • It adds tests, docs and lineage to SQL that would otherwise be unmanaged.
  • Best for teams with 20+ models and multiple analysts.
  • It does not ingest data — pair it with an EL tool.