Jaco van der Laan

From Kimball to Metadata: How Dimensional Thinking Still Shapes Modern Data Architecture

19/09/2025  ·  Data Architecture  ·  Data Engineering  ·  Data Modeling

analytics article base/dimensional-modeling dimensional-modeling kimball
‹ ›

From Kimball to Metadata- How Dimensional Thinking Still Shapes Modern Data Architecture

Reflections on Kimball’s influence and how it integrates with metadata-driven, ensemble, and business-friendly approaches

Summary

==Ralph Kimball’s dimensional modeling framework has left an enduring mark on data architecture.==

While I’ve since embraced metadata-driven engineering, ensemble modeling, and ==dual SCD2 historization==, the clarity and business-focus of Kimball’s approach still influence my thinking — and that of many others.

This article explores Kimball’s lasting impact and how it fits into a modern metadata-first ecosystem.

🆕 Update — New Series
Since publishing this piece, I’ve written a 5-part series on taking Kimball’s dimensional thinking further with metadata-driven, historized architectures:

If you enjoyed this article, this series dives deeper into dual SCD2 historization, PIT tables, automated star schema generation, and the Unified Star Schema (USS)

Kimball’s Legacy: Clarity and Accessibility

When Kimball introduced dimensional modeling, he made data understandable to business stakeholders. Fact tables and dimensions provided a storytelling language for data, turning warehouses from black boxes into something managers and analysts could actually use.

For me, and for many in the field, Kimball established two enduring principles:

  1. Business-first modeling — start with the questions and processes, not just the data sources.
  2. Simplicity in design — dimensional models abstract complexity into forms people can navigate.

Even as methodologies evolve, these principles remain timeless.

Beyond Kimball: Where My Journey Took Me

Over the years, I encountered challenges that Kimball alone didn’t address:

  • Scalability in historization — where dual SCD2 and PIT tables became essential.
  • Metadata automation — generating models, mappings, and diagrams from metadata, rather than manually designing each star schema.
  • Agility — ensemble modeling (Hans Hultgren, DVEE) offered patterns better suited for frequent business and system changes.

Still, the business orientation Kimball championed continues to shape how I evaluate these newer methods.

Fitting Kimball Into the Modern Picture

Today’s modern architecture is hybrid:

  • Ensemble modeling gives us agility and adaptability.
  • Metadata-driven engineering provides automation and governance.
  • Kimball’s dimensional thinking still guides how we design exposure views and delivery layers that business users love.

In other words:

  • Back-end historization may look like Data Vault or dual SCD2.
  • Integration and modeling may be ensemble or metadata-driven.
  • Front-end delivery often still echoes Kimball, because dimensional clarity works.

Why Kimball Still Matters

For me, Kimball remains the bridge to the business. When I teach, model, or architect, I often find myself returning to Kimball’s language: facts, dimensions, conformed dimensions, and star schemas.

The Universal Star Schema (USS) builds on this tradition, extending Kimball into a modern, metadata-driven, reusable pattern. It shows how foundational Kimball’s ideas still are — even in an age of automation.

Closing Thoughts

Kimball gave our field a shared vocabulary. Ensemble modeling, metadata-first approaches, and automation have expanded the toolkit, but the Kimball perspective continues to ground us in clarity, simplicity, and business alignment.

For me, Kimball is not something we left behind — it’s a layer of thinking that still fits into the broader architecture puzzle.