Semantic Integrity
Business Definitions vs Calculation Logic
A metric is not just a formula; it is a business definition with boundaries.
TL;DR
- • Logic alone does not describe meaning.
- • Definitions must explain scope, exclusions, and intent.
The problem (layman)
- • Measures are created without a clear business definition.
- • AI can read the formula but not the intent.
Why it matters
- • Without intent, AI answers can be correct but misleading.
- • Stakeholders may interpret results differently.
Symptoms
- • Stakeholders ask “What does this include?”
- • Two teams interpret the same measure differently.
Root causes
- • Definitions live in emails or docs, not in the model.
- • No standard for describing business meaning.
What good looks like
- • Every KPI includes a written business definition.
- • Definitions are stored alongside the calculation.
How to fix (steps)
- • Add descriptions to measures with scope and exclusions.
- • Create a lightweight metric dictionary.
- • Review definitions during model changes.
Pitfalls
- • Overly technical definitions without business context.
- • Keeping definitions in a separate system only.
Checklist
- • Business definition in measure metadata.
- • Scope and exclusions documented.
- • Definition owner identified.