dbt Turns Analysts Into Owners
Julia Schottenstein, Product Manager at dbt Labs, on the business model of open source
dbt’s wedge is that it turns the person who knows the business questions best into the person who builds the production tables everyone else uses. Before dbt, analysts usually stopped at ad hoc SQL and handed off recurring data models to data engineers. dbt packaged software engineering habits, version control, testing, pull requests, and documentation, into a SQL first workflow so analysts could ship durable tables in the warehouse without becoming full data infrastructure specialists.
-
The analytics engineer sits between analyst and data engineer. In practice, that means writing SQL models that clean raw source data into trusted tables for dashboards, finance reporting, and product metrics, while staying close to business definitions like active user, churn, or pipeline coverage.
-
This role mattered because the modern data stack split ingestion, warehouse, transformation, and BI into separate tools. Teams needed someone who could translate messy business questions into production ready warehouse logic, without waiting for a smaller, more infrastructure focused data engineering team.
-
The strategic payoff for dbt is ownership of the transformation layer where business logic lives. That is why dbt expanded from core modeling into orchestration, catalog, observability, and semantic tooling, because the team that defines clean data assets often becomes the team that defines the company’s canonical metrics.
The category is moving toward broader access, but the center of gravity remains the same, the tool that owns trusted business logic can grow outward into governance and AI interfaces. As more companies want business users to work directly with warehouse data, the winners will be the platforms that let SQL fluent operators publish reliable data assets with less help from specialized engineers.