From cost center to profit center: how Mitratech built a data platform on dbt
Mitratech's data team turned a stream of one-off customer data requests into a productized dbt offering, cutting onboarding time by 90%.
Overview
Mitratech's original DataShare was built for a single product with heavy customization, limiting scale. Covers redesigning it into a generalized platform for minimal-effort product integration via a 3-phase strategy: refactoring, pilot validation, and scaling.
What you'll take away
- 01
Mitratech turned dbt into an orchestration layer, not just a transformation tool: a DataShare package of macros, with no models or tests, provisions Snowflake shares through dbt run-operation.
- 02
Separating configuration from code let CloudOps, non-developers, onboard a new customer as a data-entry task instead of an engineering request, running a single dbt command.
- 03
The results: CloudOps time per request dropped 97%, and onboarding a brand-new product to DataShare fell from six weeks to one day.
Session speakers
Abhra Gupta
MitratechBrian Kasen
Mitratech