Most reporting arguments are not about the chart. They are about the definition behind the number. Revenue, margin, active customer, pipeline, utilization, and churn can all mean different things across teams.
A dictionary creates shared language
A KPI dictionary documents the metric name, business definition, formula, source, owner, refresh cadence, and acceptable use. It does not need to be complicated to be valuable.
Inside business intelligence and semantic modeling, this dictionary becomes the bridge between business teams and technical reporting logic.
Start with the top metrics
Do not document every metric first. Start with the metrics used in leadership meetings and recurring dashboards. The goal is to reduce ambiguity where decisions are made.
What a useful KPI dictionary includes
Each KPI should have a business definition, formula, source system, refresh rule, owner, related dimensions, and known caveats. If the metric appears in Power BI, Tableau, Looker, or Excel, the dictionary should also point to where the logic is implemented.
The strongest dictionaries are not separate documents that become outdated. They connect to semantic models, report documentation, and review habits so metric changes are visible.
KPI ownership prevents silent drift
Metrics drift when no one owns them. A finance team may change a revenue rule, sales may add a new pipeline stage, or operations may redefine completed work. Without ownership, dashboards keep showing numbers that look official but no longer match the business.
A simple owner field gives teams a route for questions and changes. That is often enough to prevent recurring reporting arguments.
FAQ
What is a KPI dictionary? It is a shared reference that defines metrics, formulas, sources, owners, refresh rules, and business context.
Which KPIs should be documented first? Start with leadership metrics, board reporting numbers, revenue metrics, operational KPIs, and any metric used across multiple dashboards.
How does a KPI dictionary improve Power BI? It helps analysts build reusable measures, align semantic models with business language, and reduce conflicting report logic.