Key Takeaways
- Of more than 16,000 projects in the Oxford megaproject database, only 0.5% came in on budget, on time, and with the promised benefits. Benefits are the side of the triangle most capital projects never close out against.
- A McKinsey review of one energy company's $20bn-plus capital portfolio found more than half of its potential net present value lost to underperformance, with root causes only diagnosed after delivery.
- Independent Project Analysis (IPA) found that only 22% of oil and gas megaprojects were successful; 64% of the failures had serious and enduring production attainment problems, averaging less than half of first-year planned production.
- Benefits realise over decades, not at commissioning. In Energy, Minerals and Resources, the conventional handover at first production is not the end of the gap. It is the start of it.
- Without a benefits realisation discipline tied back to capital allocation, an Energy, Minerals and Resources organisation has no empirical basis on which to know whether its sanctioning decisions are getting better or worse.
Energy, Minerals and Resources (EMR) capital projects are measured against the wrong scoreboard. Sanction decisions are made on the basis of a promised strategic outcome: production volume, reserve replacement, a decarbonisation milestone, a market position, an M&A synergy. Delivery is then governed against cost and schedule. The two measurement systems do not meet.
A project can land on cost, on schedule, and still fail to deliver the strategic benefit that justified the capital. This is the genuine gap between capital strategy and capital delivery. It survives even when every other governance discipline upstream works.
This article makes the case that benefits realisation, governed across the full asset life rather than ending at commissioning, is the missing accountability loop in EMR capital governance. It draws on the Project Management Institute's Benefits Realization Management (BRM) Practice Guide, IPA's industrial megaproject data, and recent EMR cases where the strategic case quietly invalidated while the asset itself was delivered.
What "Delivered" Actually Means in Energy, Minerals and Resources Capital Projects
In most sectors, project completion is a defensible end point. The software ships, the office opens, the road carries traffic. The benefit case and the handover broadly meet in the same place.
EMR is different. An LNG train is commissioned in a year; it delivers its strategic benefit over a 20 to 25 year operating life through offtake revenue, portfolio positioning, and geopolitical hedging. A mine reaches first production within months of mechanical completion; its strategic case rests on a multi-year ramp to nameplate capacity, then decades of steady-state operation against a commodity price that nobody at sanction can forecast. A decarbonisation project commits capital today against a regulatory regime, a policy environment, and a customer demand profile that will move several times before the asset retires.
The standard capital governance handover at commissioning, the point at which the project organisation disbands and the asset transfers to operations, is therefore not the moment the gap closes. It is the moment the gap becomes invisible. The project completes against cost and schedule. The benefits case continues to run for two decades or more, often with no owner still actively accountable for it and no system measuring it against the original promise.
This is the structural condition that the wider literature on optimism bias and benefit shortfalls keeps pointing at. Bent Flyvbjerg's analysis of more than 16,000 large projects found that 47.9% landed on budget, 8.5% on budget and on time, and 0.5% on budget, on time and with the promised benefits. The collapse from one column to the next is the gap this article is about.
What the Industrial Track Record Actually Shows
Inside Energy, Minerals and Resources specifically, the most defensible evidence comes from IPA, whose benchmarking criteria treat success as cost growth under roughly 25%, schedule slip under a year, and production "more or less, as planned". By that yardstick, only 22% of oil and gas megaprojects met all three tests. Of the 78% that did not, 64% experienced serious and enduring production attainment problems in the first two years after first oil or gas, averaging less than half of first-year planned production. Those projects also carried 33% real cost overruns and 30% schedule slip.
Those are not cost or schedule numbers. They are benefits numbers. They describe assets that were built, commissioned and handed over, then failed to produce what the sanction case promised they would produce.
McKinsey's recent review of an energy major's portfolio makes the same point at a different scale. The firm's analysts examined a portfolio of major capital projects worth more than $20 billion and found more than half of its potential net present value lost to underperformance, with bias in input costs, productivity, work-hours and market assumptions as the shared root cause. The notable point is not the diagnosis. It is that the diagnosis was retrospective. The organisation could see what had happened to its NPV only after the projects had been built. There was no live account of realised benefits running alongside the live account of cost and schedule.
This is the same failure pattern visible in the nickel laterite buildout of the past decade. Koniambo (Glencore / SMSP, New Caledonia) was sanctioned in 2007 at US$3.8bn against a 60,000 tonne per year nameplate. The plant was built and ran for roughly a decade, then went on care and maintenance in February 2024. Glencore's own statement put total funding at US$9 billion since project inception, with no profit ever realised. Vale's Goro plant in New Caledonia and the Sherritt-led Ambatovy project in Madagascar followed similar arcs: built, commissioned, operated for years, and eventually written down and divested without ever delivering the production case that justified the original capital. In every case, the assets were physically delivered. None of them delivered the strategic case.
A fair objection at this point is that the nickel cases are commodity-market failures, not governance failures. The 2023-2024 nickel price collapse, driven by the Indonesian supply build-out, would have invalidated those strategic cases regardless of how robust an internal benefits loop was. That objection is correct, and it marks a real limit on what benefits realisation discipline can do. It does not protect against exogenous market collapse. What it does is give the owner an early, owned, evidenced view that the case has invalidated, so capital stops flowing in years rather than decades. The cost of not having the loop is visible in Koniambo's funding profile: more than ten years of continuing investment after the strategic case was demonstrably under stress.
Where the PMI Framework Locates the Discipline That Closes the Loop
The Project Management Institute formalised this discipline in its 2018 Benefits Realization Management Practice Guide. The framework rests on a deliberately simple equation: Value equals realised benefits minus the cost of achieving those benefits. Capital governance, as practised in most EMR organisations, manages only the second term in that equation.
PMI's life cycle has three stages.
Identify. The benefit case made at sanction. Most EMR organisations do this competently. The business case carries a production target, an IRR, a strategic positioning argument, sometimes an ESG outcome. It clears the investment committee.
Execute. The delivery phase, measured against scope, cost and schedule. This is the territory of front-end loading, contractor selection, programme governance, and stage-gated execution. It is where most assurance effort sits today. Recent posts in this series have covered how front-end loading erodes under commercial pressure and why a project management office that only reports cannot govern the delivery phase.
Sustain. The realisation of the strategic benefit over the operating life of the asset. This is the stage where the discipline is weakest in practice. The frameworks are clear that accountability should transfer into the business and be carried forward, but by this point the project organisation has been disbanded, the executive sponsor has moved on, and the benefit owner, where one was named, often holds no enforceable accountability in practice.
This is the structural asymmetry the framework exposes. Cost and schedule have owners, baselines, status reports and gated reviews. Benefits, when they are tracked at all, are tracked by the asset's operating P&L, which is too aggregated to attribute back to the sanction case and too far downstream to feed the next round of capital allocation.
Why Energy, Minerals and Resources Makes This Harder Than Most Other Sectors
Three structural conditions distinguish EMR from sectors where benefits realisation is easier to govern.
Benefits realise on time horizons that exceed executive tenure. The chief executive who approves the LNG project at FID will not, in most cases, be in post when the project's strategic benefit case is finally testable. The chief financial officer who builds the model is not the one who will live with the realised IRR. Accountability is structurally diffused across multiple incumbents. The post-FID accountability gap is well documented in the academic literature on project sponsorship, and it is unusually acute in EMR because the time horizons are unusually long.
The benefit case can move while the asset is being built. An LNG project sanctioned on a thesis of long-term European import dependency entered a market reshaped by Russia, then by the US export buildout, then by Chinese demand. The IEA's World Energy Investment 2025 tracks $3.3 trillion of global energy capex this year, with the LNG buildout adding capacity through 2028 that will define market structure for decades. A green steel investment sanctioned on a 2030 regulatory regime may complete into a 2032 regime that has changed twice. PMI's BRM framework acknowledges this explicitly through its "Adapt Benefits" guidance, but the discipline of running a live, updated benefits view across a 25-year asset life is rare in practice.
The beneficiary is diffuse. PMI's framework assumes a named benefit owner with operational authority. In EMR, the beneficiary of an LNG facility is some combination of offtakers, shareholders, host governments and the parent's portfolio strategy. None of those holds the single operational role the framework presumes, which makes the benefit owner harder to locate and easier to leave unfilled. Where that role is not clearly assigned, the framework's accountability structure has nothing to bind to.
What Benefits Realisation Governance Actually Looks Like
For an EMR organisation that wants to close this loop, the structural commitments are not theoretical.
- A benefit ownership role anchored in the EPMO function, not assigned to a named individual at sanction and then quietly orphaned. The reality of capital project timelines means the executive who approves the sanction case will not be in post when the benefit is realised; a project taking seven years to build and ten to ramp will outlast almost any incumbent. The accountability therefore sits with the office, with a standing brief, a documented baseline, and a handover protocol that obliges each successor to inherit the full benefits position. This is closer to how covenant compliance or regulatory reporting accountability is structured over decades, and it survives executive turnover by design. Crucially, this requires a structural shift in corporate power. For this loop to hold, the EPMO must be elevated from an administrative reporting body into a board-empowered governance function. If the EPMO acts merely as a scribe, it will lack the political capital to challenge operational business units or declare a multi-billion dollar strategic case "lost." The function must carry a board-backed mandate to audit and adjust the strategic ledger independently of the operating asset's P&L.
- A Benefits Realisation Management Plan approved as part of sanction, not retrofitted afterwards. The plan carries the measurement system, the cadence of review, and the trigger conditions for escalation.
- A scheduled benefits review cadence through the operating life. Annual at minimum, more frequently in the ramp-up phase. The review has the authority to update the benefit case in light of market, regulatory or technological change.
- Integration with the enterprise project management office so that portfolio-level lessons are captured and fed back into capital allocation, not buried in project-level archives.
- Independent verification of benefits status, on the same principle that cost and schedule status require independent verification at the gates. Reported benefit performance and observed benefit performance are not the same thing, and the gap between them is what assurance exists to surface.
This is not a counsel of perfection. Benefits over a 25-year operating life cannot be forecast with precision; the academic critique of benefits realisation management as a discipline takes that point seriously and is right to. Richard Breese's 2012 critique argues, fairly, that benefits realisation management "is neither a panacea, nor a false dawn", but lies somewhere in between. The argument for the discipline is not that it forecasts benefits accurately. It is that it maintains a live, owned, empirically updated account of whether the strategic thesis still holds, and feeds that account back into the next capital allocation decision.
The Capital Allocation Feedback Loop
This is the closing argument. A benefits realisation discipline does not make capital allocation safe. It makes the failure visible earlier, while there is still time to act, and it builds the empirical record on which the next cycle of capital allocation can improve. Without it, an Energy, Minerals and Resources organisation sanctions every new project against assumptions for which there is no historical record of accuracy. The 0.5% figure that opens this article is not a project management problem. It is the consequence of an organisational learning failure that compounds across decades of capital deployment.
The McKinsey case study referenced above was possible because someone, eventually, did the analysis. The portfolio had run its course; the NPV was already lost. The question for boards is whether the same analysis can be done proactively, while there is still time to act, or only retrospectively when the capital has gone.
The capital deployment wave makes this question sharper, not softer. McKinsey estimates approximately $24 trillion of capital queued for heavy-industrial deployment over the next five years. The volume of decisions is rising. The owner capacity to govern them post-delivery is not. An organisation that cannot answer whether its last cycle of capital allocation worked is, by definition, sanctioning its next cycle blind.
