Building the foundation for a data mesh
EnBW’s data platform supports a wide range of customer-facing analytics, from call center forecasting and customer service optimization to pricing decisions and customer retention initiatives. As adoption grew, the team began transitioning away from Azure Synapse toward a unified Lakehouse architecture on Databricks.
At the same time, the organization was evolving its operating model. Rather than relying on a centralized analytics team, EnBW was moving toward a data mesh approach in which individual business domains own and manage their data products. While this created greater flexibility, it also introduced new governance challenges. Data existed across multiple workspaces, ownership was fragmented, access management relied on manual processes, and users often struggled to identify which datasets were trusted or who was responsible for maintaining them.
“Before, users had to work backward through pipelines to understand where a KPI came from or who owned the data quality,” said Martin Kalusa, Solutions Architect at Databricks and former Data Platform Architect at EnBW. “Those responsibilities weren’t always clear.”
The team also encountered duplicate business metrics across departments, with different teams independently recreating KPIs that shared the same name but produced different results. Without consistent governance and visibility, establishing a singleded
