Measuring What Predicts Programme Outcomes
A PMO that reports green across the portfolio while three programmes are quietly failing has not produced a report — it has produced a sedative.
The Comfortable Illusion
Every PMO I have worked with or observed produces a monthly portfolio report. The format varies, but the content is remarkably consistent: a RAG status for each programme, a milestone tracker showing what is on time and what is late, a risk summary listing the top five risks by programme, and perhaps a financial summary showing spend against budget. The report goes to the programme board or the portfolio governance forum, where it is reviewed, discussed, and filed. And in organisation after organisation, programmes that were reported as green for months fail catastrophically, apparently without warning.
The problem is not that PMOs produce bad data. The data is usually accurate. The problem is that the metrics PMOs report are lagging indicators of programme health — they tell governance forums what has already happened, not what is about to happen. RAG status is a judgement call made by the programme manager, and programme managers have powerful incentives to keep their status green. Milestone tracking tells you that a milestone was missed after it has been missed. Risk registers capture risks that have been identified, which is a different set from risks that exist. The standard PMO reporting model produces a feeling of oversight without the substance of it.
What Would Predictive Metrics Look Like?
If the purpose of PMO metrics is to inform governance decisions — and if it is not, then the metrics serve no purpose at all — then the metrics need to be predictive, not retrospective. They need to answer the question that governance forums actually need answered: Which programmes are in trouble, and what should we do about it?
In my experience, there are five categories of metric that predict programme outcomes far more reliably than RAG status and milestones.
Decision velocity. How quickly are decisions being made? A programme that is escalating decisions to its board and getting them resolved within the governance cycle is healthy. A programme where decisions are being deferred, delegated back, or lost between governance meetings is in trouble. The pattern of deferred decisions is the single most reliable early indicator of programme failure that I have encountered. It signals that the governance structure is not functioning — either because the right people are not in the room, or because the decisions are too difficult for the current governance arrangements to handle.
Scope stability. How much is the scope changing, and in which direction? Some scope change is healthy — it indicates that the programme is learning and adapting. But a pattern of scope additions without corresponding scope removals indicates that the programme is being loaded with additional requirements without the time or budget adjustments needed to deliver them. Scope creep does not appear in a milestone tracker until it is too late. A simple count of scope changes by category — additions, removals, and modifications — tracked monthly, provides an early warning that no retrospective metric can match.
Resource churn. How stable is the programme team? Programme delivery depends on knowledge, relationships, and momentum, all of which are destroyed by turnover. A programme that is losing and replacing key people every quarter is a programme that is rebuilding its capability rather than applying it. Resource churn is rarely tracked as a programme health metric, but it predicts delivery problems with remarkable accuracy. The threshold is not complicated: if more than a quarter of the programme team has changed in the last six months, the programme is at risk regardless of what its RAG status says.
Stakeholder engagement. Are the people who need to be engaged actually engaged? This is harder to measure, but it matters more than most quantitative metrics. A programme whose senior stakeholders attend governance forums, read the papers, make decisions, and visibly sponsor the work is a programme with organisational backing. A programme whose governance forums are attended by delegates, whose papers are not read in advance, and whose sponsor has not visited the programme team in three months is a programme that is losing its organisational mandate. The PMO can track this through simple attendance and participation records — who attends, who sends a delegate, who reads the pre-read, who contributes to decisions.
Dependency health. Are the programme’s external dependencies being managed, or are they being hoped for? Most programmes depend on other programmes, on business-as-usual functions, on third parties, or on decisions that sit outside the programme’s control. The standard approach is to log these dependencies in a register and review them periodically. The predictive approach is to track whether each dependency has a named owner on both sides, whether the owners are communicating, and whether the dependency is on track. A dependency with no owner on the providing side is not a managed dependency — it is a hope.
Why PMOs Resist Predictive Metrics
If predictive metrics are more useful than retrospective ones, why do PMOs not adopt them? The answer has less to do with capability than with organisational dynamics.
Predictive metrics are inherently confrontational. A metric that says this programme’s decision velocity has collapsed is an implicit criticism of the programme’s governance. A metric that says stakeholder engagement has declined is an implicit criticism of the sponsor. A metric that says resource churn is at dangerous levels is an implicit criticism of the organisation’s ability to resource its programmes. PMOs that report predictive metrics will, inevitably, create discomfort. And PMOs that create discomfort are PMOs that are challenged, marginalised, or disbanded.
A PMO that reports green across the portfolio while three programmes are quietly failing has not produced a report — it has produced a sedative.
The standard RAG-and-milestone model persists because it is comfortable. It gives governance forums the feeling that they are informed without the discomfort of actually being informed. It allows programme managers to control the narrative. It allows sponsors to assume that oversight is happening. It serves everyone’s interests except the organisation’s.
The PMO’s Choice
The metrics a PMO chooses to report are not a technical decision. They are a positioning decision. A PMO that reports retrospective metrics is positioning itself as a reporting function — it collects and presents information without interpretation or challenge. A PMO that reports predictive metrics is positioning itself as a decision-support function — it analyses information, identifies patterns, and tells governance forums what they need to hear rather than what they want to hear.
The second position is more valuable, more difficult, and more dangerous. It requires a PMO with analytical capability, political skill, and executive backing. It requires a governance culture that values challenge over comfort. And it requires programme managers who are willing to be measured by leading indicators rather than by their own assessment of their status.
Not every organisation is ready for this. But every organisation that is serious about improving its programme outcomes needs to move in this direction. The alternative is to continue investing in reporting infrastructure that consistently fails to predict the failures that matter most.
The metrics a PMO tracks are a mirror of the organisation’s maturity. An organisation that is willing to look at decision velocity, scope stability, resource churn, stakeholder engagement, and dependency health is an organisation that wants to know the truth about its programmes. An organisation that is content with RAG status and milestone tracking is an organisation that prefers the comfort of the illusion. The PMO’s job is to make the case for truth — clearly, persistently, and with the political skill to survive the discomfort it creates.