Maarten Van Stiphout

Data Vault. Kimball. Medallion. Domain-driven. Every team building a Lakehouse eventually hits the same fork in the road: how rigorously should we model our data? One camp goes strict, layered Data Vault feeding Kimball-style marts. Robust, auditable, future-proof and often painfully slow to deliver business value. The other goes pragmatic: "just ship it" – fast at first, until the lakehouse drowns in duplicated semantic models, conflicting KPI definitions and unmaintainable sprawl. Domain-driven and data mesh thinking sharpens the tension: should each domain pick its own approach, or bow to central standards? In our work with organizations across the Microsoft Data & AI ecosystem, the strongest teams refuse to pick a side. They match modelling rigor to the value at stake, and they let AI absorb the cost that made rigorous modelling so expensive in the first place.

In many organizations, data quality controls quickly grow into a scattered collection of checks that do little to truly improve data reliability. In this session, you’ll discover how to move beyond that with a powerful, impact-driven approach. The key differentiator? Controls that communicate with each other. By letting controls exchange insights and reinforce one another within a smart framework, we achieve far more than traditional, isolated checks ever could. Combined with efficient data governance, this approach enables real, sustainable improvements in data quality. Join us to explore how communicating controls can elevate your data strategy.