One P3O or three? Designing a federated model
A FTSE 250 engineering group has grown by acquisition. It runs one £400m portfolio, three transformation programmes and about 70 projects. Each division has its own PMO: 11 staff in total, four reporting templates, two scheduling tools and no common assurance. The group COO wants 'one P3O' to cut cost by 25% and standardise. Strategy and finance sit centrally. Two regulatory programmes must retain dedicated support for at least 18 months; most projects are small, repeatable upgrades. There is no portfolio office today; the capital committee makes investment decisions from spreadsheet submissions. Project managers report to divisional MDs, who are protective of their local offices. The COO asks for a recommendation in two weeks and warns that divisional MDs will resist losing their local offices.
1. Recommend a P3O model mix and reporting placement. Justify it against permanence, decision support and the 25% cost target.
Read the worked answer
Recommendation: a permanent portfolio office at group level, a permanent hub/centre of excellence, and temporary programme offices for the two regulatory programmes. Do not create a single central office that absorbs all 70 projects. Most projects are small and repeatable, so they need standards, templates and light assurance, not dedicated office staff; a project office should be justified only for a large, complex or high-risk project.
Placement: the portfolio office should report to the COO or a portfolio director with group-wide authority, because its job is to support investment decisions and prioritisation. The hub should sit alongside it under the same head, providing methods, tools, reporting standards, assurance and capability development. Programme offices should report to their programme directors for day-to-day delivery, with a functional link to the hub for standards and assurance. That preserves the temporary, programme-specific support the regulatory programmes need for 18 months without rebuilding 11 permanent local offices.
Why this meets the COO's intent: 'one P3O' should mean one coherent structure, not one team. Cost reduction comes from removing duplicated administration, standardising tools and reporting, and redeploying staff - not from stripping support from the two programmes that carry regulatory risk. A single central office would distance the P3O from divisional delivery, slow decisions and likely fail against MD resistance.
2. Which services should sit centrally and which locally? What should happen to the 11 existing staff and to assurance independence?
Read the worked answer
Central services: standards and methods, common reporting framework, tools, assurance, capability development, portfolio prioritisation and benefits framework. These need consistency and independence.
Local services: schedule maintenance, risk and issue administration, programme reporting, stakeholder logistics and local change control. These depend on daily proximity.
Staff: map the 11 existing staff against the service catalogue before deciding headcount. Retain experienced programme-office staff for the two regulatory programmes; move standards, reporting and assurance specialists into the hub/portfolio office; release or redeploy pure template administrators. Avoid announcing redundancies before the model is stable, because the cost target can be met through duplication removal and tool consolidation. Use secondments and a 6-month transition.
Assurance: do not let the hub's assurance reviewers also run the programme offices they review. If the same people provide delivery support and assurance, findings lose credibility. Separate the roles or use independent reviewers for high-risk programmes. Publish a service catalogue with customers, outputs, service levels and prioritisation rules so divisions know what they are losing and gaining.
3. What are the main risks of the COO's 'one P3O' instruction and how would you mitigate them?
Read the worked answer
Main risks: over-centralisation and loss of business intimacy; resistance from divisional MDs; assurance independence being compromised if the hub reports to delivery; cost cuts landing before value is proven; unclear mandate; and duplication continuing during transition.
Mitigations: co-design the model with divisional MDs and publish a clear mandate from the COO; phase transition over two quarters with service-level agreements; keep local business-partner roles; separate assurance reporting from delivery reporting; baseline current cost and service performance so savings are evidenced; and protect the regulatory programme offices until their closure. Measure decision timeliness, reporting quality, assurance coverage and cost per project - not just headcount.