Kimball's framework was a revelation. Using metric trees to add even more structure to data and streamline analytical operations is an exciting, natural evolution
When I first encountered Kimball’s data modeling framework, I realized that much of the work we were doing - ingesting, transforming and stitching together data pipelines - was essentially about structuring the raw data to make it easier to analyze downstream.
By organizing data into facts and dimensions, Kimball created a structure that allowed us to create meaningful, actionable reports. This framework provided a solid foundation for data engineers and BI developers alike, ensuring that data could be more easily consumed and analyzed.
But once we modeled facts and dimensions, was the job done? Not at all. As the need for data consumption evolves, so do the use cases—moving beyond basic reporting and into deeper analytics and actionable insights.
Business users want to
- Root-cause metric changes.
- Understand drivers of output metrics performance.
- Assess metric performance against targets or budgets.
- Pace and forecast metric trends.
- Simulate scenarios to inform decisions.
- Set metric goals
In theory, with the right facts and dimensions in place, we could manually write and stitch together queries for each of these workflows. But, this is far from efficient.
But as with any framework, the value lies in pushing the frontier of productivity and user enablement. Kimball gave us the base structure, but it didn’t quite capture the full potential of how data could work together to drive insights.
After we’ve structured facts and dimensions, it becomes clear that metrics should be modeled as a primary concept. But even that isn’t enough to fully realize the value of our data.
If metrics are the building blocks, the next step is modeling the relationships between them. This is where metric trees come into play—serving as the missing piece in the evolution of data modeling frameworks.
A metric tree, built on top of your existing data platform, can capture the relationships between metrics, streamline and even automate many of the common analytical workflows, making them easier and more efficient for business teams.
In a sense, metric trees represent the a logical evolution of frameworks like Kimball—continuing to push the boundaries of data standardization and enabling organizations to not just consume data, but to deeply understand and act on it.
