Background
Henkel is a global consumer-goods and industrial company operating across many strategic business units and markets. Its integrated steering-and-reporting platform gives commercial teams a single place to manage account mapping, sales pipeline, opportunities, orders, and market models. We owned the data-engineering layer beneath that platform — the governed serving views and fact tables on Databricks that every business application reads from — making sure the numbers shown to decision-makers were consistent, governed, and safe to deploy across environments.
The challenge
- Marketing and sales data spanned many strategic business units, each with different needs and access requirements.
- Metric definitions (net external sales, product contribution, win rate, market size, growth) had to stay identical across every consuming application.
- The same logic had to deploy safely and identically across development, test, acceptance, and production.
- Access had to be tightly controlled — each business unit and persona seeing only what they were entitled to.
Our approach
- Built governed serving views and fact tables on Databricks with Unity Catalog as the single source of truth.
- Templated the DDL with environment placeholders so identical SQL promotes cleanly across all four environments.
- Implemented fine-grained, role-based access — granting visibility per business unit and persona via directory groups.
- Encoded shared metric definitions directly into the views so downstream apps report consistent numbers by construction.
- Maintained a recurring weekly reporting cadence and supported the application teams consuming the layer.
My role on the project
- Authoring and maintaining the serving views and fact tables and their templated DDL.
- Implementing and reviewing the per-business-unit access grants.
- Encoding agreed metric definitions consistently across the serving layer.
- Running the recurring reporting cadence and supporting downstream app teams.
Architecture
Source data
- Marketing & sales
- Multiple business units
Databricks
- Unity Catalog
Serving layer
- Templated views & facts
- Shared metric definitions
Access control
- Per business unit & persona
Business apps
- Steering & weekly reporting
Outcomes
- A single, governed source of truth feeding multiple business applications.
- Consistent metric definitions across business units and reports.
- Safe, repeatable promotion across environments via templated, parameterized SQL.
- Controlled, auditable access aligned to business roles.
