How customer pressure pushed dbt past a cautious platform team at Delivery Hero

At dbt Summit 2026, Delivery Hero’s M. Waqas Shahid described how decentralized customer teams, rather than the cautious central platform team, drove adoption of dbt across the company’s large data estate of 300,000 tables, 60 petabytes of data, and 250 dbt projects. Delivery Hero had previously used an internal framework called DataHub, but teams found it difficult to debug and develop with, while early concerns about dbt’s scalability initially led the platform team to delay adoption. Beginning in 2023, teams independently tested dbt, created a dbt-to-Airflow integration called dbtflow, and later moved to Astronomer’s open-source Cosmos when it provided similar functionality. The platform team subsequently focused on CI workflows, scheduling enhancements, governance, and a declarative data-product framework that abstracts dbt from some users while retaining it underneath. Shahid highlighted recurring tensions between abstraction and flexibility, standardization and autonomy, local versus central control, and delivery speed versus stability, with dbt macros, ownership tags, version-controlled development, testing, and native Airflow task execution helping address these issues. He concluded that platform teams should adapt tools deliberately, recognize that organizational readiness matters alongside technical readiness, and continuously evolve their preferred “golden path” by incorporating successful team-led workarounds.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.