Most PMOs fail for the same reason: they start with templates instead of decisions. Leadership asks for "more structure", someone writes a 40-page methodology, project managers comply on paper, and six months later the PMO is seen as overhead. Delivery hasn't improved.
The PMOs that work start from the opposite end. They ask which decisions leadership is making badly or too late, and build the smallest system that fixes them. Here is the 90-day playbook I use: in a US managed services provider it took on-time delivery from 35% to 78% and brought over-budget projects from about 90% down to 22%.
Weeks 1–2: Diagnose before you design
Interview 8–12 people across leadership, delivery, sales and finance. Pull whatever data exists: project lists, timesheets, budgets, ticket queues. You are looking for answers to five questions:
- How does work enter the company, and who decides what starts?
- Where do projects actually slip: scoping, staffing, dependencies, vendors, change control?
- What does leadership see each week, and how long does it take to produce?
- Who is over capacity in the next 8–12 weeks?
- Which commitments (contracts, audits, customers) can't be missed?
Score the current state across governance, intake and portfolio, delivery and capacity, reporting and data, risk and compliance, and automation. The output is a baseline and a ranked list of problems. Not a methodology.
Weeks 3–5: Design the minimum effective PMO
Design only what the diagnosis justifies. In almost every scaling company, that means:
- A lifecycle with 3–4 stage-gates where someone with authority decides to continue, change or stop. Gates that never stop anything are theatre.
- One front door for new work, with a simple scoring model (value, urgency, effort, risk) and a check against capacity.
- A sales-to-delivery handoff gate before contracts are signed, so scope, timeline and staffing are agreed by the people who will deliver them.
- Decision rights: who approves scope changes, budget changes and priority conflicts.
- 6–8 KPIs: on-time delivery, budget variance, utilization, margin, open risks, change volume. Fewer than you think.
- Four or five core templates: charter, plan, RAID log, change request, closure. Nothing else until these are used.
Weeks 6–10: Implement on live projects
Roll the model out on real projects, not a pilot that nobody depends on. Start the weekly portfolio review with leadership in week six, even if the data is imperfect; the meeting is what creates the pressure to fix the data. Build the executive dashboard in the BI tool the company already owns. This is also the right moment to switch on the first AI agent, usually the weekly status report, because it removes the most-hated task from project managers and makes the new cadence cheap to sustain.
The weekly portfolio review is the PMO. Everything else (templates, tools, dashboards) exists to make that one meeting fast and honest.
Weeks 11–13: Measure and hand over
Compare the KPIs with the week-two baseline and share the result with leadership. Coach the project managers and whoever will own the PMO, write a short playbook, and agree what happens next: a permanent PMO lead, a fractional Head of PMO, or a quarterly health check.
Five mistakes that kill PMOs
| Mistake | What to do instead |
|---|---|
| Starting with tools | Fix intake and decision rights first; tools amplify whatever process you already have. |
| Copying an enterprise methodology | Design for your size. A 100-person company needs a page of rules, not a binder. |
| Reporting without decisions | Every dashboard should end in a decision someone owns. |
| Ignoring sales | Most delivery problems are created at signature. Put a handoff gate before it. |
| Making PMs the reporting department | Automate status reporting so project managers manage projects. |
What good looks like after 90 days
- Leadership sees portfolio health every week without anyone building slides
- New work is scored and checked against capacity before it starts
- Slipping projects are visible weeks earlier, with an owner and a recovery plan
- Project managers spend measurably less time on reporting
- On-time delivery and budget variance are trending in the right direction, and you can prove it
If you want a quick read on where your company stands, take the free PMO Maturity Check.
Want this in your company?
I run this exact playbook as a fixed-scope 90-Day PMO Launch, with your team and in your tools.
Book a 15-minute call