There is a particular kind of confidence that comes from a well-constructed feasibility model. The inputs are sound. The assumptions have been stress-tested. The numbers hold across reasonable scenarios. That confidence is not unfounded. For most of the past few decades, the relationship between rigorous analysis and reliable outcomes was stable enough to justify it. When projects failed, the causes were usually identifiable: execution problems, timing, a specific external shock that could be named and learned from. The model was not the problem.
That relationship is loosening. Projects that look viable on paper are becoming harder to deliver in practice, and the gap between the two is appearing with enough regularity that it deserves a structural explanation rather than a case-by-case one.
The explanation begins with what feasibility models assume. They assume that key variables behave within a known range. Energy costs fluctuate, but not in ways that fundamentally alter the cost base mid-project. Materials can be sourced at prices and timelines close enough to the modelled ones to manage. Labour is available within expected constraints. Financing conditions follow cycles that experienced practitioners can read and plan around. These assumptions do not need to be perfect. They need to be stable enough that deviations can be absorbed without cascading through the entire structure of the project.
That stability is becoming less reliable across multiple variables simultaneously. And that simultaneity is the critical point.
Energy costs are more volatile and more exposed to political decision-making than pricing models typically account for. Materials are subject to supply chain disruption and geopolitical pressure in ways that can move lead times and costs sharply and with limited warning. Labour availability is uneven across regions and sectors in ways that do not resolve on project timelines. Financing conditions can shift quickly when inflation and broader economic uncertainty interact. Each of these can be modelled individually. Together, they introduce a different kind of uncertainty: not variability within a range, but variability across multiple interacting variables at the same time.
This is where traditional feasibility begins to struggle, and the reason is structural rather than technical. Most models treat risk as something that can be isolated and adjusted. Increase the contingency in one area. Revise an assumption in another. The architecture of the model remains intact even as individual inputs are refined. What is emerging is a form of risk that does not isolate cleanly. Energy affects materials. Materials affect timelines. Timelines affect financing. Financing affects viability. These variables are not independent of each other, and when they move together, small changes can produce disproportionate effects. Costs do not just increase. They compound. Delays in one part of the system cascade through the rest in ways that the original model, built on an assumption of relative independence between variables, was not designed to capture.
The result is a growing gap between feasibility on paper and feasibility in delivery. Not because the modelling is incompetent, but because the model correctly describes a system that no longer quite exists.
The practical response is visible in how more sophisticated project development is evolving. Larger contingencies are being built in not as a hedge against identified risks but as structural acknowledgement that unidentified ones are more likely than before. Development is being phased to reduce exposure at any single point. Supply agreements are being secured earlier, accepting higher short-term cost to reduce longer-term vulnerability. Location decisions are being re-evaluated on the basis of energy and infrastructure access alongside demand, because a project in the right market with unreliable power or constrained materials access is a different proposition than the model suggests.
These are not simply enhanced risk management techniques. They reflect a shift in what feasibility means. In an efficiency-driven environment, feasibility is fundamentally an optimisation exercise: find the most efficient configuration of cost, time, and return, minimise slack, maximise output. In a constraint-driven environment, feasibility becomes an assessment of robustness. How does this project perform across a range of conditions, not just the expected ones. Where are the points of fragility. What happens when multiple variables shift at once.
There is also a less discussed dimension to this. For years, precision was the mark of credible feasibility work. Tight models, clear assumptions, definitive outcomes. In a stable environment, that precision was rigour. In a more variable environment, precision applied to an incomplete model produces confident answers to the wrong questions. Judgement becomes more important. The ability to read how variables interact, not just how they behave individually, becomes more valuable than the ability to model any single one of them with greater accuracy. Experience in navigating constraint, rather than experience in optimising within stability, becomes the differentiating capability.
Feasibility has not become guesswork. But it has become something harder than calculation. It is now an assessment of whether a project continues to work when the conditions under which it was conceived are no longer the conditions under which it must be delivered.
That is a higher standard. It is also, increasingly, the only honest one.
