You are working on a complex dbt model with many Common Table Expressions (CTEs) and decide to move some of those CTEs into their own model to make your code more modular.
Is this a benefit of this approach?
The new model can be documented to explain its purpose and the logic it contains.
Correct Answer: A
Yes, this is a benefit of breaking large CTE-heavy SQL models into modular dbt models. According to dbt and Analytics Engineering best practices, modularity improves clarity, maintainability, and documentation quality. When CTEs remain embedded inside a single large SQL file, their purposes are often unclear, difficult to document, and hard for other developers to reuse. By extracting a logical CTE into its own model, dbt treats it as a first-class resource-meaning it can have its own description, tests, documentation, lineage, and metadata defined in YAML.
dbt's documentation system allows each model to include a description explaining what the transformation does, the assumptions being made, and the expected behavior of the data. This aligns with the Analytics Engineering principle of creating self-documenting pipelines, where transformations are transparent and easier for downstream users to understand.
Additionally, modular models improve lineage visualization in the DAG. Instead of a single model hiding multiple transformation layers, a modular structure reveals how data flows through each intermediate step, helping both debugging and governance. Modularization also enables reusability-other models can reference the intermediate model rather than rebuilding the same logic through duplicated CTEs, supporting DRY (Don't Repeat Yourself) principles.
Therefore, moving CTEs into separate dbt models absolutely provides a documentation benefit and improves the overall engineering quality of the project.
