PMO Maturity Models: Useful Diagnostic or Corporate Theatre?
An organisation assesses its PMO against a maturity model, comes out at level two, and sets a target of level three by the end of the following year. Eighteen months later the assessment returns level three. Whether anything about delivery improved in the meantime is a separate question, and on most such programmes nobody has been asked it.
That is the honest case against maturity models, and it is not an argument for ignoring them. An assessment measures which practices are present and consistent across an organisation, which is genuinely worth knowing and frequently surprising. The trouble starts when the resulting number stops being a description and becomes an objective, because a level is far cheaper to reach through documentation than through better work.
What an assessment actually measures
Appendix X2.5 of the PMBOK® Guide treats PMO maturity models as a way of diagnosing where capability is uneven, and it is careful not to present a higher level as a good in its own right. Read that way, an assessment returns three things worth having.
It returns an inventory. Somebody has gone through the organisation and established which practices genuinely exist, as opposed to which ones appear in a handbook. Most PMO leaders discover at least one area where they believed something was routine and it turns out to be one team's habit.
It returns a map of inconsistency, which is usually the most valuable output. Where three delivery areas are assessed at different levels for the same practice, the interesting question is why: whether one has a problem the others do not, whether two are doing something sensible that the model does not recognise, or whether a standard was never really adopted anywhere outside the office that wrote it.
And it returns a shared vocabulary for a conversation that is otherwise hard to have. Telling a delivery director that their estimating is weak invites an argument. Showing that estimating practice is assessed at one level in their area and three in another, with the evidence attached, moves the discussion onto ground where something can be checked.
None of those three tells you whether projects are succeeding. A maturity model has nothing to say about whether the practices it counts are helping, because it was never built to answer that. An organisation can be consistently, repeatably, well-documented at doing the wrong things, and the assessment will record it as progress.
How it turns into theatre
The mechanism is simple and almost impossible to resist once a level becomes a target. Assessments run on evidence, evidence is produced by people, and producing evidence is a cheaper, faster and more controllable activity than changing how several hundred people work. Within a few months the improvement programme is a documentation programme with an improvement label, and everybody involved can see it.
The second symptom is practices adopted at the level the model describes and abandoned at the level of the work. A lessons learned process exists, is documented, has an owner and a template, and is performed at the end of each project by one person filling in a form nobody reads. The assessment records it as present. On the ward, so to speak, nothing has changed.
An NHS trust ran a maturity assessment across its estates and digital programmes and set a target level as part of a board commitment. Two years and a considerable amount of effort later the target was met. During the same period the trust installed a pneumatic sample transport system in one wing, commissioned it, and recorded the project as delivered against every gate the improved process required. Portering staff continued to walk samples to the laboratory, because the carriers did not fit the newer sample racks and nobody had owned the interface between the estates project and the pathology department. Every practice the model counted had been performed. The one thing that would have caught the problem, somebody accountable for the outcome rather than the delivery, was not something the model asked about.
Using one honestly
Two disciplines keep an assessment useful. The first is to treat the output as a set of findings and never as a score to be improved. Publish what was found, area by area, with the evidence and the disagreements intact, and resist the summary number entirely if you can. Where a board insists on a number, put it alongside the outcome measures it is supposed to relate to, so that a rising level unaccompanied by any change in delivery performance is visible to everybody looking.
The second is to tie any improvement to a named decision that will get better. Raising a practice from one level to the next should come with a sentence saying which decision will be made differently as a result, by whom, and how anyone will know. Where that sentence cannot be written, the uplift is a cost with no stated benefit, and the honest response is to leave the practice where it is. Judging which controls genuinely earn their cost, and which have accumulated because a framework mentioned them, is the same proportionality question that runs through a structured PMP exam preparation course, and it applies to a PMO's own practices as much as to a project's.
For a PMP® candidate, what to hold onto is that governance and process exist to support outcomes, and that a model describing good practice is not evidence that good practice is producing anything. In a situation where an organisation is pursuing a capability target, the useful thing to ask is what decision or outcome is expected to change, and whether anyone has said how it will be measured.
An afternoon of watching is the cheapest test an assessment will ever get. Pick three practices the last assessment recorded as present, and go and watch them happen. Sit in the lessons session. Read the last four risk escalations end to end. Ask the person who fills in the template what they do with it afterwards. Where the practice is real, you will find somebody who can tell you what changed because of it. Where you cannot find that person, you have found something the assessment could not.
Interested in going further?
Reading a capability assessment for what it genuinely evidences, and holding that line when a board has already committed to a target level, takes a clear head about what process can and cannot demonstrate. Omega's PMP® Exam Preparation works through governance and improvement decisions in terms of the outcomes they are meant to serve.
Maturity models and the rest of the PMO appendix are covered at greater length in the PMBOK® Guide Eighth Edition.
Ad · Amazon affiliate link.
References
A100: Which PMO Model Fits Your Organisation?
A098: PMO Value: What Should a Project Management Office Actually Deliver?
A099: Customer-Centric PMOs: A Better Test of Relevance
A093: Choosing a Project Governance Model
A053: AI Governance for Project Managers
PMP and PMBOK are registered marks of the Project Management Institute, Inc.
