Capability Maturity Model
The Capability Maturity Model rates how consistently an organisation runs a process, from ad hoc through to self-improving, so you know which level to aim for next rather than jumping straight to 'best practice'.
Five rungs climb from ad hoc at the bottom to self-improving at the top, one at a time.
Reach for this when…
- Everyone does the process differently and nobody knows the 'right' way.
- Leadership wants 'world class' when the team is barely consistent.
- You need to justify investment in process before capability, not after.
How to run it
- Name the five levels: Initial, Repeatable, Defined, Managed, Optimising.
- Assess the current level for one specific capability.
- Define what 'good' looks like for this context, not the textbook.
- Identify the gap between current and target level.
- Build a roadmap to move up one level at a time.
A worked example
Situation. Yossi Levi's dev shop, Levi Software in Tel Aviv, Israel, missed almost every delivery date - each project was run however that week's lead preferred.
Applied. He assessed the team as Initial and resisted the urge to chase 'Optimising' language from a conference talk. The target was Repeatable: basic estimation and tracking, nothing more, for one team first.
Result. Two quarters later that team hit Repeatable and their missed-deadline rate had halved. He rolled the same low-ambition target out to the next team.
The catch
The model rewards consistency, not good judgement - a team can be perfectly repeatable at doing the wrong thing. And levels get gamed: teams that know they're being assessed produce the paperwork of maturity without the behaviour. Use it to focus effort on one level up, never as a badge.
Skipping a level to impress an audit almost always means faking the evidence for it.
Origin: Watts Humphrey; Carnegie Mellon SEI