

- Description
- A compass
- Source
- Slashme · original page
- Licence
- CC0 Public domain
Understanding maturity matrices
A tool to see where you stand, decide where to go, and in what order.
At a glance
- Ordered levels + dimensions + observable criteria.
- Used to diagnose, prioritize, plan and track.
- Pitfalls: chasing the level, optimistic self-assessment, universal target.
- Choose your matrix →
This section describes matrices that place the level of an organization. For the maturity of a product (tool, solution) and how it helps its users, see Product maturity.
Definition#
A maturity matrix (or maturity model) describes, for a given domain, a series of ordered levels, from the most rudimentary to the most advanced, and the observable criteria that make it possible to place an organization on each one. It is used to establish a shared diagnosis, set a target and plan progress.
Origin#
The principle comes from software engineering: the Capability Maturity Model (CMM) was developed from the late 1980s at the Software Engineering Institute of Carnegie Mellon University to assess the capability of software suppliers. It gave rise to CMMI (Capability Maturity Model Integration), now maintained by ISACA, which took over the CMMI Institute in 2016. The most frequently cited staged version distinguishes five levels: Initial, Managed, Defined, Quantitatively Managed, Optimizing.
Since then the format has spread: data management, analytics, security, testing, cloud, DevOps, tool adoption.
Anatomy of a matrix#
- Levels: an ordered scale (often 3 to 6 levels), each described by characteristics.
- Dimensions (or domains, capabilities, themes): the aspects assessed separately.
- Criteria and evidence: what, concretely, makes it possible to state that a level has been reached.
- An assessment method: questionnaire, interviews, document review, audit.
| Level 1 | Level 2 | Level 3 | |
|---|---|---|---|
| Dimension A | criteria / evidence | criteria / evidence | criteria / evidence |
| Dimension B | criteria / evidence | criteria / evidence | criteria / evidence |
| Dimension C | criteria / evidence | criteria / evidence | criteria / evidence |
The grid above is schematic: each cell contains criteria specific to the model.
Two families of models#
| Family | Principle | Examples here |
|---|---|---|
| Staged | One overall level; a level is reached only once the requirements of the previous levels are satisfied. | TMMi, TDWI |
| Continuous (by dimension) | Each dimension, practice or capability has its own level; progress is uneven by design. | DAMA-DMBOK, OWASP SAMM, FinOps, Fabric / Power BI |
Why it matters#
- A common language: technical, business and management speak about the same level, with the same words.
- An honest diagnosis: you know where you stand, without comparing yourself to an unreachable target.
- Prioritization: you deal first with the dimension that limits the others.
- A trajectory in stages: realistic steps rather than a big bang.
- Tracking over time: reassessing makes it possible to measure progress.
- A decision lever: justify an investment by a gap against a well-argued target.
- A product roadmap aid: maturity gaps inform the order of work items, alongside user requests.
How to use them#
- Define the scope: a product, a team, a department?
- Choose a recognized matrix rather than inventing one; adapt it if necessary and say so.
- Assess with evidence, preferably with several people and an outside perspective.
- Set targets per dimension, justified by a stake or a risk.
- Turn gaps into dated actions, each with an owner.
- Reassess at regular intervals and adjust.
Limitations and pitfalls#
- Chasing the level: when the level becomes a goal, you optimize the indicator rather than the result (Goodhart's law: “When a measure becomes a target, it ceases to be a good measure”).
- A level is not value: a high maturity level guarantees neither a good product nor a good financial result.
- Self-assessment is optimistic: without evidence, declared levels are too high (see the Dunning-Kruger effect).
- The “right” level depends on context: aiming for the maximum everywhere is rarely relevant (the FinOps framework says so explicitly).
- A model simplifies: it may ignore sector-specific or size-specific features.
- Bureaucracy: a model applied to the letter can weigh down small teams; adapt it.
- A model ages: cite the version and the date consulted.
My attainment level: how it works#
Each matrix page has a “My attainment level” switch. When enabled, it highlights on the scale the level I have entered, and shows the level per dimension when known. The data is entered by hand in a single configuration file: overall level, level per dimension, assessment date and comment. As long as a value has not been entered, the page shows “not provided”.
Sources and verification#
- CMMI Institute / ISACA — official website — CMMI model
- Carnegie Mellon University, SEI — Capability Maturity Model Integration — origin of CMM / CMMI
- Microsoft Learn — Fabric adoption roadmap: maturity levels — recalls the alignment with the five CMM levels
- FinOps Foundation — maturity model — maturity per capability, don't aim for “Run” everywhere