PMOs Die Because We Build Reporting Offices — Replace Them with Decision Infrastructure
The report did not fail to reach the board; the board failed to become a place where consequence was owned.
The funeral is built into the design
The pattern is now too familiar to excuse.
A new leadership team arrives and declares that the programme estate lacks control. A programme management office is established with ambition. It will provide insight, challenge, coherence and a single view of change. Within six months, the office is assembling reports. Within a year, it is chasing updates, reconciling traffic lights and policing templates. By the second year, senior leaders complain that it is bureaucratic, delivery teams call it overhead, and the people inside it have responsibility for the quality of information but little authority over the decisions that information demands.
Then the sponsor changes. The PMO is reviewed, reduced, renamed or disbanded.
The organisation congratulates itself for removing bureaucracy. The same unresolved choices, conflicting priorities and weak accountabilities remain. A few months later, someone proposes a new office to restore control.
We must stop pretending this is a lifecycle problem. It is a design failure repeated with remarkable discipline.
PMOs do not die because organisations no longer need governance. They die because we build them as reporting offices, deny them influence, and then blame them for the irrelevance we designed into them.
What should replace them is not a leaner reporting function, a more polished dashboard or another central team with a new title. What should replace them is decision infrastructure: a deliberate system that connects evidence, authority, timing and consequence across the programme estate.
If we are not prepared to build that, we should build nothing.
Stop calling reporting control
The most damaging word in the PMO vocabulary is control. It permits us to confuse the observation of work with the direction of work.
Control exists only when information reaches someone with the authority and obligation to act, while action is still useful.
Yet most PMOs are commissioned around outputs:
- a consolidated status report;
- a common planning standard;
- a risk and issue log;
- financial reconciliation;
- board papers and minutes;
- compliance with the chosen delivery method.
These outputs may be necessary. They are not the purpose. When we treat them as the purpose, the office learns to optimise production. It becomes faster at collecting updates, better at formatting exceptions and more disciplined at recording decisions that other people have postponed.
That is not control. It is the administration of distance between leadership and consequence.
A PMO becomes overhead when the organisation asks it to produce evidence but gives it no role in turning evidence into choice.
We have made the wrong bargain
The conventional bargain appears sensible. Programme leaders make decisions; the PMO supplies information. The separation protects accountability. Nobody wants an unelected support office quietly taking ownership away from sponsors and executives.
But the bargain is false because supplying information is not a neutral act. To integrate a programme is to decide which dates matter, which dependencies cross boundaries, which assumptions are credible and which variances deserve attention. Every consolidated plan contains judgement. Every status summary compresses disagreement. Every red rating challenges someone’s preferred account of reality.
We demand this judgement from the PMO and then deny that it is judgement. We tell the office to challenge, but punish it when challenge disrupts the narrative. We ask for independence, but place careers and budgets under the leaders being reported. We ask for one version of the truth, but allow each programme to negotiate what the truth means.
The result is institutional timidity disguised as process.
By April 2016, organisations have invested heavily in portfolio tools, delivery methods and assurance regimes. They can move data between systems, standardise governance calendars and distribute board packs across locations with ease. Yet the central problem remains stubbornly human: who must choose, by when, on what evidence, and with what consequence if they do not?
No tool answers that question. No template can carry the authority we refuse to assign.
A programme estate that looked controlled
Consider a composite portfolio of fourteen programmes with an annual change budget of £180 million. Its PMO employed twenty-three people. Every month it produced a 186-page portfolio pack. Ninety-four per cent of projects submitted on time. Financial forecasts reconciled to within one per cent of the ledger. The traffic-light definitions were clear, and the board met every four weeks.
On paper, this was a mature control environment.
Underneath, six programmes depended on the same group of forty-two technical specialists. Four business units had committed benefits from reducing the same operating costs. Two supplier contracts assumed that internal decisions would be made within ten working days, although the relevant committee met monthly. None of these contradictions appeared as a portfolio choice. They appeared as local risks.
Over five reporting cycles, the PMO recorded thirty-one red items. Twenty-four were carried forward more than once. Nine were discussed at three consecutive boards. Only five were framed as explicit decisions with options, authority and a latest useful date.
The PMO responded as PMOs are trained to respond: it improved the pack. It added a dependency page, refined the resource view and introduced stronger commentary. The report became more revealing. The decisions did not become faster.
Six months later, leadership concluded that the office was too large and insufficiently commercial. Eleven roles were removed. The contradiction between demand and capacity remained, now with fewer people available to describe it.
This is how PMOs die. Not in one dramatic failure, but through a long sequence in which the office is made accountable for visibility while leadership retains the right not to act.
The report did not fail to reach the board; the board failed to become a place where consequence was owned.
Replace the office with decision infrastructure
Decision infrastructure is not a department. It is the set of obligations that makes governance real wherever work is performed.
It has five parts.
Evidence that is sufficient, not endless
Leaders must define what evidence is enough for a decision. The demand for more detail is often caution, but it is also one of the safest ways to defer accountability. A decision paper should state the material facts, uncertainty, options and consequences. It should not promise certainty that the programme cannot possess.
The test is simple: will another week of analysis change the choice, or merely make delay look responsible?
Authority that is named before conflict
Every material class of decision must have an owner before the decision arises. Scope, funding, risk acceptance, resource priority, commercial compromise and benefit accountability cannot be left to whatever forum happens to discuss them.
When authority is shared, the mechanism for breaking disagreement must be explicit. “Joint ownership” is too often a courteous description of nobody being able to decide.
Time treated as a governing variable
A choice made after the last useful date is not governance. It is history.
Decision infrastructure records latency. It shows how long an issue has waited, which options are closing and what each period of deferral costs. A board should see not only the recommendation but the price of returning next month.
Consequence connected across boundaries
Programmes fail at the joins. A project can be green while the portfolio is impossible. A supplier can meet a contractual milestone while the business is unready. A programme can protect its budget by transferring cost to operations. Decision infrastructure makes those transfers visible.
The unit of analysis is not the report-owning team. It is the outcome and every commitment required to produce it.
Memory that preserves judgement
Leadership changes. Programme directors move. Suppliers replace account teams. Without a reliable record of why choices were made, each new arrival reopens settled matters or inherits assumptions without understanding them.
The governing memory must record options considered, evidence available, authority exercised and consequences accepted. Minutes alone rarely do this. They prove that a meeting happened; they do not always preserve the reasoning that should guide the next decision.
Decision infrastructure holds these obligations together. A PMO may operate parts of it. It must never be mistaken for the whole.
Abolish the powerless PMO
We should be explicit about what must end.
A PMO with no mandate to challenge an implausible forecast should not exist. A PMO that can escalate an issue but cannot require a decision owner should not exist. A PMO whose performance is measured by submission compliance while decisions age unseen should not exist. A PMO that reports portfolio conflicts without a route to resolve them should not exist.
These offices provide comfort at too high a price. They allow leaders to feel informed while preserving ambiguity. They allow programmes to perform compliance while competing for the same resources. They allow assurance to verify the presence of artefacts without testing whether governance changes outcomes.
We do not need to improve these offices. We need to refuse their premise.
“The reporting PMO is not a weak version of the institution we need; it is a different institution serving a different appetite.”
The appetite is for evidence of oversight. The need is for decisions with consequence.
The serious objection
There is a strong case for restraint. PMOs are support functions. Sponsors and programme directors must remain accountable. Giving a central office too much influence can create a parallel hierarchy, slow delivery and encourage analysts to substitute process for judgement. A forceful PMO can become an empire that challenges everything and owns nothing.
This objection is correct about the danger and wrong about the remedy.
Decision infrastructure does not make the PMO sovereign. It makes accountability enforceable. The office does not decide the business strategy; it requires the strategic choice to have an owner. It does not accept programme risk; it prevents risk from circulating without acceptance. It does not allocate every specialist; it exposes when the combined demand is impossible and places the trade-off before the authorised leader.
The boundary is authority over decision integrity, not authority over the decision itself.
A well-designed function can reject incomplete evidence, expose contradiction, name latency and escalate absent authority. It cannot replace the sponsor. Indeed, it makes sponsorship more real by removing the shelter of vague information and collective discussion.
Build from the decision backwards
We have traditionally built PMOs from methods forwards. We choose standards, templates, calendars, roles and tools, then ask the office to produce governance.
We must reverse the sequence.
- Name the choices that determine success. Which decisions about scope, priority, funding, capacity, risk, commercial position and benefits cannot be left inside individual projects?
- Assign authority and a last useful date. Who must decide, and when does delay begin to destroy options?
- Define the minimum evidence. What must be known, what can remain uncertain, and how will competing accounts be tested?
- Design the route. Which forum decides, how disagreement moves, and who ensures that an issue does not circulate without resolution?
- Create the memory. How will the decision, reasoning and consequence remain visible after people change?
- Only then assign roles and tools. Establish the smallest capability able to keep those obligations whole.
Sometimes that capability will be called a PMO. Sometimes it will be a compact portfolio team, a governance secretariat with real mandate, or a distributed network anchored by a strong integrator. The label does not matter.
The obligations do.
Measure what the old office concealed
If we replace reporting offices, we must also replace their measures.
Submission timeliness, data completeness and forecasting accuracy remain useful, but they are hygiene. They cannot tell us whether governance is working.
We should measure:
- decision age — the elapsed time between an issue becoming material and authorised resolution;
- decision expiry — the number of choices made after important options had closed;
- repeat escalation — matters returned to a forum without new evidence or a new choice;
- cross-programme conflict — resource, dependency and benefit contradictions unresolved at portfolio level;
- action consequence — whether agreed decisions changed plans, funding, contracts or operational commitments;
- assumption survival — critical assumptions carried forward without a current owner or test.
These measures are not attractive because they expose leadership, not merely delivery. That is precisely why they matter.
When the programme board sees that its median decision age is forty-seven days while supplier commitments assume ten, governance becomes a subject of management rather than ceremony. When the portfolio discovers that eight benefits depend on the same headcount reduction, the issue can no longer hide in separate business cases.
What we measure tells the organisation where accountability lives. Measure report production and the PMO will own the failure. Measure decision latency and leadership must enter the frame.
Leadership must accept the replacement
No redesign will survive if executives still want the old bargain.
Decision infrastructure requires leaders to accept choices before certainty is complete, to resolve conflicts among powerful sponsors, and to own the consequences of priority. It requires them to welcome a function that exposes when authority is absent. It requires programme directors to treat challenge as part of control, not as disloyalty.
This is why PMO reform so often stops at structure. It is easier to remove posts than to redesign authority. Easier to buy a new tool than to tell two executives that only one programme can have the people. Easier to demand sharper status than to admit that the status is red because leadership has not chosen.
The next PMO will die like the last one if leadership behaviour remains untouched.
Decision infrastructure begins when leaders make three commitments:
- We will not request information without naming the decision it informs.
- We will not allow material choices to proceed without an authorised owner and a latest useful date.
- We will not call a programme controlled merely because its deterioration is accurately reported.
These are not PMO commitments. They are governing commitments.
Build what the programme actually needs
The age of the ceremonial PMO should end.
We should stop establishing offices because the programme feels large, because assurance expects one, or because a new leader wants a visible symbol of control. We should stop loading them with reporting duties, starving them of mandate and disbanding them when their impotence becomes embarrassing.
We should build the governing capability from the decision backwards. We should keep it as small as purpose allows and as strong as consequence requires. We should place expertise near the work, but never leave integration without a home. We should use tools to move evidence, not to impersonate judgement. We should preserve sponsorship by making it unavoidable.
A PMO deserves to survive only if it makes the organisation more capable of deciding.
That is not a modest reform. It changes what we believe the office is for. It changes who is exposed by delay. It changes the board from a recipient of status into an owner of consequence.
The choice before us is not between bureaucracy and freedom. It is between visible administration and actual governance.
Let the reporting office die. Replace it with decision infrastructure — or stop pretending that the programme is governed.