A PMO is not one thing. It is a dozen different jobs that share a name — and choosing the wrong one is the most common reason the function ends up resented rather than relied upon.
Program governance · Project management · Project monitoring office
Ensure initiative business case and value is realised for end client.
A sponsor asks for a PMO. One executive imagines an independent body that will tell the board the truth about a troubled programme. Another expects a reporting function that collects status from workstream leads. A third wants a delivery engine that will actually chase the work. All three are legitimate; none of them is compatible with the others. Six weeks in, the office is producing packs nobody reads and taking blame for delays it has no authority to prevent.
So the first conversation is never about tooling or templates. It is about which of the scenarios below the organisation is actually in, who the office reports to, and what it is empowered to stop. Get that settled and the mechanics are relatively straightforward. Leave it ambiguous and no amount of governance rigour will rescue it.
Most often on the client's side of the table, particularly where a systems integrator is running delivery. An integrator's PMO reports on the integrator's progress; that is not a criticism, it is simply whose office it is. An owner needs its own instrumentation — independent status, independent risk, and someone whose reporting line does not run to the party being measured.
I have also built offices and handed them over. A PMO that cannot eventually be run by the client's own people is a dependency, not a capability.
Almost every PMO scenario is a variation on one of these positions. Naming the altitude first settles most arguments about scope and authority.
Each has a different mandate, a different reporting line and a different definition of success. Identifying which one applies is the first deliverable of the engagement, not an assumption made before it.
Plan, risk register, dependency map, financials and reporting cadence for a single major change such as an ERP replacement, a network build or a plant systems roll-out. The default and most familiar form.
Prioritisation, stage gates, capacity modelling and a single honest view of what is genuinely in flight. Usually the answer when every function is running its own change and nobody can total it.
Coordination across workstreams that report to different executives, where the risk lives in the seams rather than inside any one stream. Carries change management and benefits alongside delivery.
A standing office that holds the digital roadmap, arbitrates between competing initiatives, and keeps the portfolio pointed at the outcomes the strategy promised rather than the technologies that are fashionable.
Independent instrumentation for the buyer: own plan, own risk view, own reading of contractual milestones and change control. Reports to the client sponsor, never through the integrator.
A time-boxed review rather than a standing office. Tests the plan, the estimate, the resourcing and the reported position, then reports plainly to the sponsor or audit committee.
Stop, establish the true position, re-baseline scope, plan and commercials, and restart with a governance structure that will surface the next problem earlier. Frequently the hardest conversation, not the hardest work.
Day 1 readiness and Day 100 delivery across two application estates, two infrastructures and two operating models, with the separation and transitional service arrangements tracked alongside.
The mirror of integration: untangling shared systems, data and contracts to a defined separation date, with transitional service agreements scoped and priced before signature rather than after.
Service levels measured and enforced, commercial baselines maintained, renewals anticipated and supplier performance reviewed on a cadence the contract actually supports.
Fixed external date, non-negotiable scope, and evidence requirements that must survive inspection. Governance built backwards from the regulator’s test rather than forwards from the plan.
Reconnects the delivered solution to the business case: what was promised, what has landed, where the value leaked and what it will take to recover it. Runs six to eighteen months after go-live.
Experienced programme managers, business analysts and PMO analysts supplied against the client’s own governance framework, for a defined period and a defined scope.
Stand the office up, run it through the difficult phase, then train and hand over to the client’s own team against an agreed exit date. Success is measured by how cleanly it can be left.
Four to six weeks to mobilise, then a steady cadence. The sequence is deliberate: authority is settled before process, and process before tooling.
Which scenario applies, who the office reports to, what it is empowered to escalate and what it is empowered to stop. Written down and agreed by the sponsor before anything else begins.
Scope, plan, cost, resourcing and risk as they actually stand rather than as last reported. On troubled programmes this stage is uncomfortable and is the entire value of the exercise.
Forums, membership, decision rights, escalation thresholds and meeting cadence — sized to the programme. More governance is not better governance; it is usually the substitute for authority.
Milestones defined so completion is unambiguous, dependencies mapped across workstreams, risks owned by named people with dates, and change control that a supplier cannot route around.
One version of status, issued on a fixed rhythm, written so a non-technical board member can act on it. Amber means something specific and is used.
The business case carried through delivery rather than filed at approval, so that scope decisions can be tested against the value they were meant to protect.
Documentation, tooling and trained client staff, with a dated handover. An office that becomes permanent by default has usually stopped being useful.
A single schedule across workstreams and suppliers, with dependencies and critical path visible rather than implied.
Owned, dated and actively worked — not a list reviewed monthly and otherwise ignored.
One pack, one version of the truth, at board, sponsor and workstream altitudes.
Budget against commitment against actual, with forecast to complete maintained rather than restated at year end.
Every scope movement costed, approved and traceable to who asked for it.
Defined entry and exit criteria, so go-live is a judgement made against evidence rather than a date defended.
Service levels and contractual milestones measured by the client, from the client’s own records.
Business case commitments carried forward and measured after go-live.
Everything the client’s own team needs to run the office once the engagement ends.
Clients are described by sector rather than named.