Most board packs are lists: forty measures, each with a traffic light, none connected to the others. When results disappoint, the meeting spends its first hour establishing what happened and has little time left for what to do.
Start from the P&L, not the dashboard
A KPI tree starts at the line the board cares about — usually EBITDA or cash — and breaks it into the few drivers that move it. Every measure on the tree must explain the one above it. If it doesn’t, it belongs in an operational report, not the board pack.
Separate volume, price and mix
The single most useful split in most businesses is volume, price and mix. Without it, a revenue miss is a mystery; with it, it is a diagnosis. The same split makes forecasts explainable, which is what restores the board’s trust in them.
Five metrics to retire
Revenue per employee, unless headcount is the actual constraint.
Average order value reported without mix.
Utilisation without an agreed definition of capacity.
Traffic-light status with no baseline behind the colour.
Any metric that nobody owns.
A board pack should explain variance, not list it.
From the work · PRV-007
Sample data
Forecast accuracy ·
Wholesale distribution group
Before
71%
+22 pts
After
93%
A driver-based model on volume, price and mix replaced forty spreadsheets; the pack now ships on working day three.
Source
Forecast vs actual, six months before and after go-live.
Illustrative sample data. Provan, its people, clients, figures and sources are fictional and shown for demonstration only.