Data Core · KPI management · Module
Choosing what to measure
Narrowing to the few measures the organisation is actually run on, balanced across what matters rather than across what is easy to collect.
The idea
How it works
Measures are selected by availability far more often than by relevance. What the systems already produce becomes what the organisation watches, which is a decision made by whoever configured the software.
The number is the discipline. A dashboard of thirty measures distributes attention evenly across things that differ in importance by an order of magnitude, which is the same as having none.
Working with it
In practice
- 01
Start from the decision
For each candidate, name the decision it would inform. Candidates with no decision are reporting rather than KPIs.
- 02
Include leading measures deliberately
Lagging measures are easier and arrive too late. At least one measure should move before the outcome does.
- 03
Balance across what matters
Financial measures alone describe the past. Customer, process and capability measures describe whether it continues.
- 04
Cap the number
Five to seven at any level. Adding one should require removing one, or the list only grows.
One level in
The components of choosing what to measure
A component is something that exists afterwards which did not exist before — a deliverable or a mechanism, not an intention.
Adding a measure should require removing one. Otherwise the list only ever grows.
The other modules in kpi management
Definitions
What each measure means precisely enough that two people calculating it independently get the same answer.
LearnTargets and thresholds
What good looks like, at what level, and what happens when the reading crosses it — because a measure with no threshold produces observation rather than action.
LearnThe review
The rhythm in which measures are actually read and something is decided — because where the review is presentation, the whole measurement effort is decorative.
Learn