dbt Centralizes Business Logic

Diving deeper into

Julia Schottenstein, Product Manager at dbt Labs, on the business model of open source

Interview
dbt will have a very central role in that shift because we describe the underlying logic of what the application needs to do.
Analyzed 4 sources

dbt matters here because it turns a warehouse from a place that stores data into a place that can reliably drive live product behavior. The key asset is not the table itself, but the code that defines how revenue, usage, customer health, or training completion should be calculated, tested, documented, and refreshed. Once that logic is written in dbt, other tools can read from it, sync it into apps, or expose it to customers without redefining the same rules in every downstream system.

  • In practice, this lets analytics teams build workflows that used to require application engineers. One example is customer facing training dashboards powered from warehouse data refreshed in near real time. Another is reverse ETL, where dbt models define segments or account scores once, then tools like Census push those outputs into Salesforce, Braze, or support systems.
  • dbt sits in the middle because it owns the transform step. Fivetran and similar tools load raw data in. dbt cleans and joins it into business ready tables and metrics. Reverse ETL and app layers then consume those outputs. That makes dbt the place where a company encodes what counts as an active user, a qualified lead, or an at risk customer.
  • This role also explains why warehouses and data clouds both partner with and compete against dbt. Snowflake and Databricks can add native transformation features, but companies with multiple warehouses or BI tools do not want to rewrite their business logic in each vendor stack. dbt's vendor neutral control plane pitch is that the logic should live one layer above the infrastructure.

The next step is more production software running on warehouse data, especially internal tools, customer dashboards, AI features, and operational systems that can tolerate minute level freshness rather than millisecond latency. As warehouses get faster and open table standards make multi cloud setups more common, the control point with the most leverage will be the layer that defines and governs business logic, which is where dbt is expanding.