Skip to content

KPI catalog & storytelling

The theory of a KPI does not depend on the tool. Each KPI here is defined once in a YAML catalog, calculated once by a tested Python engine, and presented in two tools: this story website and an interactive KPI dashboard, both generated from the same catalog.

  • Five business cases - service SLA, maintenance, inventory, customers and data quality, each told as context → question → KPI → finding → decision → how we keep measuring.
  • Anti-pattern gallery - the same data shown wrong and right, side by side.
  • KPI catalog - definition, formula, target, owner, how to read it, common misreadings, SQL and a hand-calculated worked example for every KPI.
  • KPI dashboard - all 28 KPIs as cards with status, trend and target; filters by domain and status.

Why this exists

A dashboard is only as good as the definitions behind it. Most KPI problems are not tool problems: an average that hides a long tail, a fill rate that looks fine while customers get incomplete orders, a forecast error metric dominated by slow movers, a green traffic light with no target behind it. This site shows how each KPI is defined, read and turned into a decision, and the method page lists the chart rules every page follows.

Data

Only public (UCI, CC BY 4.0) or synthetic data. With synthetic data the findings illustrate the method; they are not claims about a real organisation. See data sources.