A project management office, or PMO, is the team or function that gives an organization a single, consistent way to select, run, and govern its projects. At its best, a PMO is the difference between a company that finishes the right work and one that stays permanently busy on the wrong work. At its worst, it is a reporting layer that produces status decks nobody reads.
The difference is not the org chart. It is whether the PMO owns decisions or just paperwork. This guide covers what a PMO genuinely does, the three common types, and a practical path to standing one up without the bureaucracy that gives the function a bad name.
Key takeaways
- A PMO standardizes how projects are chosen, run, and reported so leaders can compare them fairly.
- There are three common types: supportive, controlling, and directive. Pick the lightest one that solves your actual problem.
- Start with the one decision the organization keeps getting wrong, not with a 40-page methodology.
What a PMO actually does
Strip away the frameworks and a working PMO does a handful of concrete things. It defines how a project is proposed and approved, so the answer to "should we do this?" is not just whoever asked loudest. It sets a common way to estimate, plan, and track work, so two projects can be compared on the same terms. It runs the governance cadence where projects get reviewed, funded, paused, or killed. And it gives leadership a true picture of what is in flight, what it costs, and whether it is on track.
Notice what is not on that list: doing the project management for every team. A good PMO raises the floor on how projects are run across the organization. It does not try to personally run them all.
The three types of PMO
Most PMOs fall into one of three models, and the right one depends on how much authority the function needs to fix the problem you actually have.
Supportive PMO
A supportive PMO provides templates, training, lessons learned, and a shared toolset. It has a low degree of control: project teams can take or leave its help. This model fits organizations where projects are generally healthy and the gap is consistency, not discipline. It is the cheapest to stand up and the easiest to ignore.
Controlling PMO
A controlling PMO requires that projects follow defined standards: a particular planning approach, specific gates, mandated reporting. It has a moderate degree of control and works when leadership has decided that inconsistency is causing real harm and wants compliance, not just guidance. The risk is that it tips into box-ticking if the standards are not genuinely useful.
Directive PMO
A directive PMO supplies the project managers and runs the projects directly. It has the highest degree of control because the work happens inside it. This fits organizations with high project volume, regulatory exposure, or a track record of delivery failures that justifies centralizing the discipline.
How to set up a PMO: a practical path
The fastest way to discredit a new PMO is to open with a thick methodology binder. Do the opposite. Start narrow and earn the right to expand.
1. Name the one decision you keep getting wrong. Is it that everything is "priority one"? That projects start before they are funded? That nobody knows who is overcommitted? Your first job is to fix that single failure, not to standardize everything.
2. Build the intake and prioritization first. Most PMO value comes from choosing the right work. A simple, transparent way to score and rank incoming requests beats a perfect tracking system for projects that should never have started. We cover this in depth in how to prioritize a project portfolio.
3. Establish a light governance cadence. A recurring forum where projects are reviewed against criteria, with real decision rights, is what turns a PMO from a reporting function into a steering one. See project portfolio governance for the cadences and decision rights that work.
4. Make capacity visible. You cannot say yes responsibly until you know who is already committed. A basic view of demand versus the people you actually have prevents the silent over-commitment that wrecks delivery. More in resource and capacity planning for project portfolios.
5. Report on outcomes, not activity. Executives do not need a list of tasks. They need to know whether the portfolio is delivering the results it was funded for. We cover what that looks like in PMO reporting that executives actually read.
Signs your PMO is working
You will know the function is earning its keep when leaders start asking the PMO before committing to new work, when at least some projects get paused or killed at reviews instead of limping on, and when teams stop being assigned to four "top priority" efforts at once. If none of that is happening after a couple of quarters, the PMO has become a reporting layer, and it is time to revisit which decisions it actually owns.