Front-End Discipline: Why Major Projects Go Wrong Before Construction Starts

16 Aug

In a recent interview with Construction Briefing, Darren Mort, president of Bechtel’s Infrastructure global business unit, made a quietly radical claim: most major infrastructure projects are already in trouble by the time the first contractor arrives on site. The cause is not engineering complexity, supply chain volatility, or labour shortage — though each plays a role. It is a failure of front-end discipline.

“Those things are normally expected and planned for,” Mort said of engineering definition, stakeholder requirements, regulatory obligations, cost, schedule, and delivery responsibilities, “but all too often projects rush to execution without having some of those fundamentals in place.”

Bechtel is not a peripheral voice. It is one of the few delivery organisations in the world that operates at the scale of national infrastructure, semiconductor manufacturing, and energy transition programmes simultaneously. When its leadership says the centre of gravity of project risk sits before construction starts, the industry should listen — because the financial consequence of getting it wrong is no longer measurable in months of delay. It is measured in years of lost capacity and billions of stranded capital.

The Structural Pattern

The pattern Mort describes is not new, but it is becoming more acute. Programmes are growing in scale, complexity, and political visibility. Customers expect not only delivery against scope, but a defensible answer to how a project will be delivered — the governance, the incentives, the risk allocation, the readiness to operate.

And yet, the industry’s default behaviour is to compress the front end. Definition is treated as a phase to be moved through, not a discipline to be invested in. The result is a predictable failure mode: scope is locked before it is understood; suppliers are engaged before their capability is tested; operational readiness is deferred to a handover stage that arrives too late to influence the design.

Across our own work in semiconductor, data centre, life sciences, and energy programmes, we see the same compression. The owner’s organisation is rarely structured to hold a multi-year definition phase. The supplier base is rarely qualified against the actual operating envelope of the asset. Risk registers are inherited from prior programmes rather than built from the specifics of the current one. By the time the first concrete is poured, the room for recovery has already narrowed.

“All too often projects rush to execution without having some of those fundamentals in place.” — Darren Mort, Bechtel

What Disciplined Front-End Actually Looks Like

Mort points to Western Sydney International Airport as a working example. Before full implementation, Bechtel spent nearly two years working with the customer to establish the foundations of the programme. The asset is now expected to complete almost seven months ahead of schedule.

Two years of definition is not a luxury. It is a control on cost. Industry analysis consistently shows that megaprojects with weak front-end definition overrun by 30 percent or more, while projects with disciplined early-stage investment cluster tightly around plan. The marginal cost of an additional month of definition is small. The marginal cost of an additional month of recovery, post-mobilisation, is an order of magnitude larger.

What separates disciplined front-end from theatre is structural, not procedural. Three conditions matter most:

1. An Empowered Owner Role

Front-end definition cannot be delegated to a procurement function or a programme management contractor. It requires an owner-side leadership role with the authority to commit the organisation to scope, sequence, and supplier choices — and the seniority to refuse to advance the programme until those decisions are stable.

In most organisations, this role does not exist by default. It has to be created, staffed, and protected from the institutional pressure to declare progress.

2. Clear Division of Responsibility

Mort’s language is precise: “an empowered role and a clear division of responsibility.” The two are inseparable. Empowerment without clarity produces conflict; clarity without empowerment produces inertia.

A defensible RACI for a major programme is not a spreadsheet. It is a contract between the owner, the engineering lead, the supplier base, and the operations function — written before mobilisation, tested against failure scenarios, and revised as the programme matures.

3. Incentive Models That Reward Delivery

The third condition is the one most often skipped. Commercial structures designed to minimise contractor margin will reliably produce minimum-compliance delivery. Structures designed to reward early problem disclosure, integrated planning, and operational readiness produce a different outcome.

Getting this right requires the owner to spend commercial design effort at the front end — on incentive geometry, risk allocation, and gain-share mechanics — rather than treating the commercial model as a procurement artefact.

The AI and Power Dimension

Mort also pointed to the surge in power demand driven by AI data centres, particularly in North America. This is the operating reality for the next decade: data centre programmes are now competing for the same engineering, supplier, and grid-connection capacity as semiconductor fabs, battery facilities, and renewable energy build-out.

In that environment, the projects that secure capacity early — in supplier commitments, in grid interconnection, in regulatory approvals — will deliver. The projects that defer those decisions into execution will compete for residual capacity at a premium, or fail to secure it at all. Front-end discipline is no longer just a delivery question. It is a market-access question.

Implication for Owners

The conclusion is uncomfortable for owner organisations whose internal incentives reward the appearance of momentum. A two-year definition phase is hard to defend to a board that wants to see ground-breaking imagery. It is even harder to defend when the alternative — rushing to execution — produces visible activity, even as it locks in the conditions for later failure.

The owners that will set the standard over the next decade are those who treat front-end definition as a capital decision in its own right: staffed by senior delivery talent, governed at the executive level, and resourced to produce a programme that is genuinely ready to execute. Everything downstream — supplier certainty, operational readiness, risk management, fractional support — is downstream of that one decision.

Mort’s point is not that megaprojects are inherently uncontrollable. It is that the controls have to be installed earlier than the industry is comfortable installing them. The discipline is available. The question is whether owners will pay the front-end cost — in time, in talent, and in organisational patience — to capture the downstream return.