How to Set Up a PMO at a Mid-Size Company in 30 Days
A PMO isn't set up with an org chart and a 40-tab spreadsheet template. It's set up by proving, in the first month, that leadership can make a better decision because the PMO exists. Everything else — methodology, governance, a formal office — comes after, and only if that first month convinced somebody.
Before you start: what kind of PMO
"PMO" means very different things. For a mid-size company (50–500 people, 10 to 40 live projects), what almost always works is a supportive PMO with teeth: it gives the portfolio visibility and standards, helps project managers, and has the authority to demand data — not to run the projects for them.
What doesn't work in 30 days: a controlling PMO that imposes a heavyweight methodology from day one. It breeds resistance, takes months to show results, and usually dies in the first budget cut.
The only goal for month 1. That in the first portfolio review with the leadership committee, someone says "I didn't know project X was in that shape" and a decision gets made in the room. That moment is what buys the next six months of mandate.
Week 1 — Inventory and waterline
Don't design anything yet. Find out what's actually there.
- A single project list. Everything consuming budget or people. Name, sponsor, project manager, one-sentence objective, committed finish date, budget. There are almost always 20% to 40% more projects than leadership thought.
- Short interviews (30 min) with each project manager. One question that matters: "what worries you about this project that isn't in any report?"
- Define the portfolio threshold. Which projects enter the PMO's radar (by budget, risk, political visibility). The rest get listed but not tracked weekly.
Deliverable of the week: the inventory. A table. Nothing more.
Week 2 — A minimum standard, not a manual
Define the bare minimum needed for every project to be comparable. Resisting the urge to write a full methodology is 80% of this week's work.
| Element | Minimum standard |
|---|---|
| Statuses | A shared vocabulary: not started · in progress · at risk · blocked · closed. And a rule for each color. |
| Schedule | Every radar project has a plan with milestones and a frozen baseline. No baseline, no variance to measure. |
| Progress | A single criterion (0/100 per milestone is enough to start) that isn't decided by whoever's reporting. |
| Risks | One register per project, with probability, impact and owner. Only the high-exposure ones roll up to the portfolio. |
| Cadence | Biweekly data updates; monthly portfolio review with the committee. |
Deliverable of the week: a two-page document with those rules, reviewed with two or three project managers so it doesn't feel imposed.
Week 3 — Load data and calculate
Now it gets populated. With the project managers, not for them: the PMO supports the first load, but the data is theirs.
- Schedule with a baseline for every radar project.
- Budget and spend to date (internal hours included, even as a rough estimate in the first month).
- Status, high risks, and pending decisions.
With that, you can already compute what actually moves a meeting: CPI and SPI per project, milestone variance in days, and a closeout forecast (EAC) for the three or four large projects. It doesn't have to be perfect; it has to be consistent across projects.
This is where the spreadsheet betrays you. Manually maintaining EVM for 15 projects — with their baseline, spend, and milestones — updated every two weeks, is several hours of work per cycle, and by month 2 nobody has them anymore. This part of the method should be automatic calculation from day one, or the PMO will end up being exactly what it promised not to be: a reporting office.
Week 4 — The first portfolio review
The deliverable that justifies the whole month. A 60–90 minute session with the leadership committee, with a fixed script:
- Portfolio snapshot (5 min) — how many projects, how much budget, how many green / amber / red under the agreed rule.
- Ones that need a decision today (30 min) — two or three projects, with data, options, and a recommendation. Not all 15 get reviewed.
- Trend (10 min) — what's changed since… well, since never; this is the PMO's own baseline.
- What the PMO will do next month (5 min) — concrete commitments, not a vision statement.
If at least one decision comes out of that session that wouldn't have happened without the PMO's data, the mandate is won. If what comes out is a "very interesting, thanks," revisit which project you chose to bring and how sharply the options were presented.
What NOT to do in the first month
- Don't write a corporate methodology. Write two pages of shared rules.
- Don't buy or roll out a heavyweight tool with three months of configuration. You need something that computes EVM and shows the portfolio this week.
- Don't audit the project managers. The PMO shows up to help and to give visibility, not to assign blame — at least until it has earned credibility.
- Don't try to cover all 40 projects. Cover the ones that matter and list the rest.
- Don't build a dashboard with 30 metrics. CPI, SPI, milestone variance, high risks, and pending decisions. Five things.
Month 2 and beyond
With the mandate secured, you build what would have been rejected before: charter templates, a change-control process, demand management for new projects, project manager development, and a serious look at portfolio resource capacity. But all of that rests on the first month having produced a visible decision. Without that anchor, the PMO is a staff function the first tough quarter sweeps away.
PMOvio is the tool for that first month
Portfolio, baseline, CPI/SPI, closeout forecast, and the monthly review ready for the committee — without three months of rollout. 15 days free, no card required.
Set up my portfolio →