Semantic Integrity
A Lightweight Metric Dictionary That Actually Gets Used
A simple metric dictionary helps teams align without heavy governance overhead.
TL;DR
- • Keep it short, in the model, and owned.
- • If it’s hard to maintain, it won’t be used.
The problem (layman)
- • Metric definitions are scattered across documents.
- • Teams don’t trust or consult the source of truth.
Why it matters
- • AI needs definitions stored in the model to retrieve them.
- • A concise dictionary reduces metric sprawl.
Symptoms
- • People ask “What does this mean?” repeatedly.
- • New measures appear without documentation.
Root causes
- • Dictionary lives outside the model.
- • No ownership or review cycle.
What good looks like
- • Metric dictionary stored as descriptions and annotations.
- • Owners listed and review dates tracked.
How to fix (steps)
- • Start with top 20 metrics.
- • Add definitions to model metadata.
- • Assign owners and review quarterly.
Pitfalls
- • Over‑engineering the dictionary with too much detail.
- • No maintenance schedule.
Checklist
- • Definitions stored in model.
- • Owners listed for top metrics.
- • Quarterly review scheduled.