Tugger Toolkit - Power BI Data Model - Session 1.2: Fact Granularity & Real-World Choices
đź“… Wednesday 11 March 2026
⏰ 11:00 AM - 11:30 AM GMT
⏱ Duration: 30 minutes
🎯 Level: Intermediate - Advanced Power BI users
Your data model is built. Now let's make it do something useful.
Session 1.1 gave you the Star Schema foundation. Session 1.2 is where it gets interesting.
Real-world data doesn't always arrive at a neat, consistent level. Jobs have headers and lines. Absences haverecords and daily entries. Cost centres cut across both. The grain you choosedetermines what your reports can and can't do - and getting it wrong meansfilters break, totals double, and your numbers quietly mislead.
We cover the decisions Power BI developers actually face - and the patterns that resolve them correctly.
What you'll leave with:
- Understand fact granularity - What it means, why it matters, and how mixing grains silently breaks your filters and aggregations
- Apply the grain rule - The user's filter defines the model's grain - how to apply that rule to make confident design decisions every time
- Resolve header vs line conflicts - When a single line-level fact is the right call, and when it isn't • Handle the allocation problem - Why header facts can't be filtered at line grain - and the three ways to tackle it
- Build a dual-fact model - How to structure a model with both a header and a line fact - and what each one can and can't do
- Apply custom allocation - How to allocate header-level metrics proportionally to line grain so every filter works consistently
- Navigate real-world trade-offs - Data isn't always perfect - how to make deliberate, informed choices based on what your reporting actuallyneeds
Join us and build Power BI models where every filter works - every time.
New to the series?
Session 1.2 builds directly on Session 1.1. If you haven't caught it yet, we'd recommend starting there - therecording is available now.
