PMO · Project Management Office
Project Management Office
What is a PMO?
- Project standardization, visibility, and resource coordination
- Matures from supportive to controlling to directive
- If the PMO becomes just a report collector, it has failed
The first thing a PMO changes is how the monthly management meeting runs. Before, deciding which project got more people came down to which executive pushed hardest, and a report of '80% progress' meant something different on every team. The development team called design completion 80%; the production team called 80% of allocated effort used. Half of what a PMO does is pin down the definitions of progress and completion in a single sentence, and the other half is putting every project on the same scale using the numbers that result.
PMO authority levels fall into supportive, controlling, and directive types, but what to weigh first in practice is scale. With fewer than 10 concurrent projects, one planning-team manager working part-time beats a dedicated unit; once you pass 20 to 30 and people start overlapping across departments, 2 to 3 dedicated staff begin to make sense. The boundary between PM and PMO must also be clear. A PM is accountable for the deadline and deliverables of their own project, while the PMO is accountable for resources, priorities, and stop decisions across projects. When this line blurs, the PMO ends up acting as the PMs’ boss and only draws resistance.
A PMO gets its real power not from reports but from gates. Control works only if budget and headcount can be reallocated at three or four checkpoints, such as kickoff, design completion, and handover to mass production. Resource planning needs one principle too: a headcount plan built on 100% utilization is effectively a reservation for delay. Unless you leave a 15–20% buffer for vacations, incident response, and rework, the moment one project slips, every linked project falls like dominoes.
The most common misconception is setting up the PMO as an audit body. If you start with control, business units build double books: real schedules for internal team use and separate schedules for PMO submission. Every number on the status board is then accurate, and none of it reflects reality. The test is simple. If, within three months of launch, the time business units spend writing reports has not gone down, that PMO is already on a path to failure. The right order is to lighten the load with templates and data, and in return get standards adopted.
John KimAuthor of B2B Marketing Design and CEO of Nanum Consulting
Soyeon KimExpert in design thinking, social venture startups, and the social economyIt spread in the 1990s with demand for managing large IT and construction projects, and became an established organizational function as PMI’s project management standards gained wide adoption.
Let’s look at a hypothetical case. Company A makes industrial automation equipment, with 480 employees and annual revenue of KRW 120 billion (≈ US$89,780,000). As orders grew, concurrent projects swelled to 27, and its on-time delivery rate fell to 58%. The CEO said, "We only find out which project is at risk when something blows up," and set up a three-person PMO under the planning office.
- When Director Kim in the planning office spent two weeks compiling the contract deadlines and staffing of 27 projects onto a single page, Sales’ count of active projects was off by 4 from the real number.
- The PMO standardized the progress measure to 'accepted deliverables only,' and in the first tally one development lead reversed decisions from two meetings, protesting, "We were at 80% complete, so why does it say 45%?"
- After creating a company-wide monthly status board that reported all 27 projects in three colors (on track, caution, at risk), the CEO decided on the spot to postpone the start of 2 of the 6 at-risk projects.
- It turned out that a lead engineer in the equipment control group had been assigned to three projects at once, booked at 140% utilization, so the PMO shifted one project to outsourced design.
- Six months in, the company added two gate reviews, kickoff and design completion. Of the five reviews in the first quarter, two were put on hold because specifications were not finalized.
- Eight months later, on-time delivery had risen from 58% to 84%, average delay had fallen from 34 days to 11, and PMs’ weekly report-writing time dropped from 6 hours to 2.