PMO setup, done honestly, starts with a question most providers skip: should this office exist at all? A programme management office earns its cost only when the portfolio is large enough that decisions — sequencing, funding, killing, resourcing — genuinely exceed what a good operating cadence can carry. When it is justified, I build the right-sized version: portfolio governance that produces decisions rather than status, a programme cadence with real gates and escalation thresholds, benefit tracking a CFO will accept, and the minimum reporting that supports all three. When it is not justified, I will tell you so and build the lighter thing instead — an operating rhythm and clear owners, at a fraction of the cost.
The craft comes from running programmes where the governance had to hold, not from a methodology binder. The quality programme that moved 95% to 99% across 2,000+ campaigns, the billing transformation that cut a two-month cycle to fifteen days across 75 entities and 2,000+ employees — these were multi-workstream programmes with sceptical stakeholders, hard baselines and benefits that had to be proven, run inside a global network where slippage was visible immediately. The formal grounding is there too — Google-certified in both project management fundamentals and agile process management, alongside Lean Six Sigma — but certifications describe the toolkit. The evidence is that the programmes landed and the gains held.
The design bias throughout is smallness. The best PMO is the smallest one that governs: a handful of gates that actually gate, a portfolio review that kills weak projects while they are still cheap, benefit baselines set before work starts, and reporting thin enough that delivery teams barely feel it. The office is built to be handed over — an internal owner runs it, the playbooks are theirs, and my exit is part of the design. What I will not build is a template factory: a PMO that measures its own health by compliance with its own paperwork has already failed, expensively and quietly.