Project Governance

What project governance means

Project governance is the framework of authority, accountability and decision rights within which a project is directed. It answers four questions that no schedule, methodology or software tool can answer on its own: Who decides? Who is accountable when it goes wrong? Who sees the truth about progress? And who has the authority to stop?

Governance is routinely confused with management. Management is the work of planning, coordinating and delivering; governance is the structure that directs, oversees and — critically — constrains management. A project can be competently managed and catastrophically governed at the same time. Most of the great project disasters of the past fifty years were exactly that: delivery teams working hard inside governance structures that made honest reporting impossible and bad decisions irreversible.

The empirical pattern is uncomfortable. When large projects fail, the post-mortem almost always finds that someone, somewhere, knew — and that the governance structure either filtered that knowledge out or punished the person carrying it. Governance failure is therefore rarely a failure of information. It is a failure of the design that determines what happens to information.

Key questions the topic raises

• Where should decision authority sit, and how should it change as a project's risk profile changes?
• How do boards detect that they are being told what they want to hear?
• What is the difference between oversight and interference?
• How should stage gates be designed so that "stop" is a genuinely available option?
• Why do governance structures that work at £10 million collapse at £1 billion?
• What role should independent assurance play, and when does it become theatre?

The anatomy of governance failure

Across the case archive, three governance failure patterns recur with almost mechanical regularity.

1. The optimism trap: governance that cannot hear bad news. The Berlin Brandenburg Airport opened nine years late and roughly three times over its original budget. The technical problems — fire safety systems chief among them — were real, but the governance failure preceded them: opening dates were announced politically and then defended as facts, while the supervisory board received progress reports that bore diminishing relation to site reality. Each announcement raised the political cost of honesty. By the time truth forced its way through, the board had become the last audience to hear it. This is the optimism trap: governance structures that make delay politically expensive will be supplied with information that conceals delay.

2. Diffused accountability: everyone responsible, no one accountable. The NHS National Programme for IT — discontinued after costs approaching £10 billion — was governed through layers of agencies, boards and supplier consortia in which accountability for the clinical and operational outcome belonged to no identifiable person. When everything is governed by committee, the committee's first product is ambiguity about who can be blamed, and its second is paralysis. Contrast with the structural problem in the Crossrail delay: two parent sponsors, a separate delivery authority, and dozens of contractors, with no single body owning the integrated railway until late in the day.

3. Governance capture: when the governed write the rules. The Boeing 737 MAX investigation is the extreme case: a regulatory delegation regime in which the manufacturer performed safety assessments on behalf of its own regulator, under commercial schedule pressure, produced a certification of an aircraft whose flight-control characteristics were not fully disclosed to operators. Governance capture is not confined to regulators. Internal audit functions that report to the executives they audit, and "independent" reviewers selected and paid by the project director, are structurally captured before they begin.

Designing governance that works

Effective governance at scale shares identifiable design features, visible by their absence in every failure above.

Single-point accountability with matching authority. One named individual must own the outcome and must hold the budget and decision rights to match. The Sydney Opera House illustrates the inverse: after the architect's departure, no single party held coherent authority over design intent, and cost and schedule became variables to be negotiated rather than constraints to be governed.

Gates with teeth. Stage gates only function if stopping or redirecting the project is a real, survivable option for the decision-maker. Where a gate review has never once stopped a project, it is a ceremony. Political and financial sunk-cost dynamics must be actively countered — for example by independent gate reviewers with no stake in continuation.

Honest reporting architecture. Governance quality is determined less by the reports received than by the channels that bypass them. Site visits, direct access to delivery teams, and protected escalation routes are the mechanisms by which boards discover what dashboards conceal.

Proportionality and escalation of scrutiny. Governance intensity should increase with irreversibility and integration risk, not merely with spend. A £50 million decision that commits a programme to a single architecture deserves more scrutiny than a £500 million procurement of a proven commodity.

Governance in the Asian context

Asia's delivery environment adds specific governance dynamics: state-directed programmes with strong central authority can achieve remarkable decisiveness — and can also suppress the upward flow of bad news with unusual efficiency. Conversely, multi-party coalitions and rapid political turnover in some democracies produce governance churn, in which accountability resets with every election. Neither system is inherently superior for delivery; each requires its governance structure to compensate for its characteristic blind spot, a theme developed in our governance category archive and across the Asia Infrastructure pillar.

Featured investigations

• Boeing 737 MAX — governance capture and delegated oversight at its most consequential.
• Berlin Brandenburg Airport — political schedule-setting and the suppression of bad news.
• Crossrail — fragmented sponsorship and missing integration accountability.
• Governance category — further case studies in board-level failure.

Related frameworks

• Leadership Blind Spot Matrix — locating where leadership attention systematically misses risk.
• Project Failure Pyramid — how governance defects compound upward.
• Decision Quality Model — structuring decision rights and decision review.

Frequently asked questions

What is the difference between project governance and project management?

Governance defines who decides, who is accountable and how oversight works; management executes within that structure. Failures of governance — suppressed bad news, diffused accountability, captured oversight — cannot be fixed by better management.

What is the single most common governance failure in megaprojects?

Structures that make honesty expensive: politically fixed deadlines and budgets that cause reporting systems to conceal variance until it becomes undeniable, as documented at Berlin Brandenburg and Crossrail.

Who should chair a major project's board?

Someone with the standing to absorb bad news without career damage and the authority to stop the project — which in practice means a senior sponsor with genuine independence from the delivery organisation's incentives.

Do stage gates actually work?

Only when stopping is a real option. Where gates have never halted a project, they function as assurance theatre. Independent reviewers, protected escalation and explicit "exit criteria" give gates teeth.

How does governance differ in state-directed programmes?

Centralised authority accelerates decisions but can suppress the flow of adverse information upward; decentralised systems surface information but struggle with decisiveness. Effective governance design compensates for whichever blind spot the system carries.

Last reviewed: 1 August 2026 · Author: Ramesh Dixit