A project management office is a standing organisational function which concerns itself with project work as a whole and not with any single project. What such an office does varies enormously between organisations. The eighth edition gives the function more attention than earlier editions did, on the grounds that where it sits and what it is entitled to decide will shape every project beneath it.
Perhaps the most useful question one can ask about any such office is how much authority it holds. There are three conventional answers, and almost every dispute about whether an office adds value turns out to be a disagreement about which of the three it was meant to be.
Three degrees of control
| Type | What it provides | Authority over projects | Where it fits |
|---|---|---|---|
| Supportive | Templates, methods, training, tooling, lessons, coaching on request | Low, and advisory in nature | Mature project managers who need consistency and a place to look things up |
| Controlling | All of the above, plus required frameworks, compliance review and assurance | Moderate, through conditions attached to funding or stage approval | Organisations needing consistency across many projects, or answering to a regulator |
| Directive | Project managers themselves, assigned from the office and reporting to it | High, since the office manages the work | Organisations wanting uniform delivery, and settings where project management is the core service |
Trouble generally arises from a mismatch between the type an organisation asked for and the behaviour it receives. An office which was chartered as supportive and then begins rejecting plans has assumed authority it was never given, and project managers will respond by routing around it. An office chartered as controlling but with no connection to the funding decision holds responsibility without leverage, and its requirements will therefore become paperwork.
What a good one does
Certain functions produce value across all three types.
Method and tailoring guidance are held by the office, so that a project manager beginning work has a defensible default and not a blank page. It maintains a portfolio view, which is the only place from which anybody can see that four projects are competing for the same two specialists in March. It runs assurance, by which is meant an independent examination of whether a project is in the state it reports itself to be in. It owns the lessons repository and, more importantly, the practice of putting relevant lessons in front of new work before anybody searches for them. Finally, it develops project managers, since a discipline practised alone does not improve.
Of these, the portfolio view is what most justifies the function. Individual projects will optimise for themselves, and the conflicts between them are visible only from above.
Where offices fail
Measurement of the wrong thing is the common cause. Where an office is assessed on template completion and report timeliness, it will produce complete templates and timely reports, and it will be indifferent to whether projects deliver. Measures which keep it honest are concerned with outcomes, such as forecast accuracy, benefit realisation and the rate at which projects are stopped when they should be.
Some offices become little more than a reporting conduit. Where an office's whole function consists of collecting status from projects and forwarding it upward, it adds a week of latency and no judgement, and everybody involved is aware of this.
A third pattern is method rigidity. Where an office mandates one approach for all work it will force adaptive projects through predictive gates and small projects through the ceremony of large ones. What corrects this is that the office should own the tailoring guidance and not the uniformity, which means publishing what may be adjusted and on what basis, and then reviewing the reasoning given for a departure and not merely the conformity to the standard.
Adaptive delivery changes the shape
Continuous product delivery fits badly with a function built around temporary projects which have a start and an end. Annual funding rounds, stage gates and project closure all assume work that stops.
Offices in these settings tend to move towards funding stable teams in place of approving individual projects, towards the measurement of benefit in place of completion, and towards assurance which samples what is actually being delivered in place of reviewing plans in advance. Very often the name remains the same while the function beneath it becomes something closer to portfolio management.
What the exam expects
Questions on this topic generally describe a friction between a project and the office and then ask what the project manager should do. Answers which score treat the office as a legitimate part of governance and not as an obstacle. In practice this means consulting it early, using what it provides and taking a tailoring departure to it as a proposal with reasoning attached, before proceeding.
One other pattern is worth recognising, and it hinges on the type of office. Where an office is described as providing templates and training it has no authority to reject a plan, and an option in which it does so is wrong for a structural reason and not a procedural one.