The design that performs best at the forecast average can be the design that fails first when the future refuses to behave like the forecast.
Capital planning often begins with expected values: expected demand, expected price, expected energy availability, expected climate, expected utilisation. This is necessary. No organisation can design every asset for every imaginable condition.
The danger is not using an expected case. The danger is quietly turning it into a guaranteed future.
Research on off-grid energy, climate adaptation, waste infrastructure and sustainable supply chains shows why robust decisions treat uncertainty as part of design rather than as a note added after the design is complete. When inputs change, the optimal configuration can change with them.
For leaders, this is not an argument for maximum redundancy. It is an argument for choosing where efficiency should yield to resilience, flexibility and option value.
The Strategic Context
Brivio and colleagues' 2017 work on off-grid power-system design is a clear illustration. Rural electrification projects often begin with imperfect information because historical load data may not exist. Future demand can also evolve as households and businesses gain access to electricity, while renewable-resource availability varies. Their design approach generated multiple load profiles and lifetime evolution scenarios, producing a map of solutions rather than relying on one deterministic demand curve. The research concluded that input uncertainty could materially alter the optimum plant configuration.
Sacchelli and colleagues approached uncertainty from a different field. Their climate-adaptation model for Italian wine farms combined climate scenarios, extreme events, farm economics and alternative adaptation strategies. The work reflects a broader reality: environmental uncertainty interacts with market and operational uncertainty rather than arriving as a separate risk category.
Habibi and colleagues incorporated uncertainty into long-term municipal waste infrastructure planning. Yousefi and colleagues used robust and fuzzy approaches because management goals and supply-chain relationships could not be represented as perfectly known values.
Across these cases, the common principle is simple: when uncertainty is structural, deterministic optimisation can create brittle confidence.
What Leaders Commonly Misread
The first mistake is to confuse forecasting accuracy with decision quality. A forecast can be wrong and still support a good decision if the chosen strategy performs acceptably across a reasonable range of outcomes.
The second is to assume robustness means designing for the worst imaginable case. That can destroy economics. Robustness is better understood as acceptable performance across the futures that matter.
The third is to treat uncertainty as a risk-register problem after architecture is fixed. Some uncertainties should influence capacity, modularity, contracting, technology choice and sequencing before execution begins.
The fourth is to believe that the option with the highest expected return must be strategically superior. A slightly lower expected return may be preferable if it protects downside, preserves flexibility or creates learning before irreversible commitment.
Reframing the Issue
A deterministic question asks: What configuration is optimal if our assumptions are correct?
A robust question asks: Which configuration remains acceptable when our most important assumptions are wrong?
This shifts attention from point optimisation to performance envelopes.
For a power system, that envelope might include demand growth, renewable variability and component replacement. For a manufacturing expansion, it might include volume, product mix, labour availability, energy price and automation performance. For infrastructure, it may include climate extremes, demographic change and maintenance capacity.
The aim is not to predict all futures. It is to identify the small number of uncertainties that can break the investment logic.
Efficiency and Robustness Are Different Design Goals
An asset sized tightly around expected demand can have excellent utilisation and poor resilience. Excess capacity can protect service but trap capital. The strategic design problem is therefore not “more buffer is safer”. It is where to place flexibility.
Flexibility can be physical, such as modular capacity. It can be contractual, such as options to expand or exit. It can be financial, such as staged funding. It can be technological, such as interoperable interfaces. It can be organisational, such as retaining specialist capability during transition.
The strongest design often combines several modest forms of flexibility rather than buying large amounts of idle redundancy.
Consider a hypothetical industrial energy project. Management expects electricity demand to rise 20 per cent over five years. A single large system may have the lowest expected unit cost. A modular system may cost slightly more at first but allow later capacity to be added when demand becomes observable. If demand stalls, the organisation avoids stranded capital. If demand accelerates, the expansion path remains open. The value of the modular option is not captured fully by expected unit cost.
Uncertainty Can Change the Objective
Climate adaptation adds another layer. Under stable conditions, the objective may be cost efficiency. Under increasing disruption, continuity, recoverability or quality protection can become more important.
This means leaders should not only vary model inputs; they should test whether future conditions change the relative importance of objectives.
A business facing water scarcity, for instance, may initially optimise for lowest cost per unit produced. If water availability falls below a threshold, resilience of supply may dominate cost. The governance system should recognise that threshold before the crisis arrives.
Decision Framework
A practical robustness review can be built around five questions.
1. Identify the decision-breaking assumptions
Which three to five assumptions, if wrong, would materially damage the investment case?
2. Define plausible ranges or scenarios
Avoid false precision. Use bounded ranges, directional scenarios or discrete states where probabilities are weak.
3. Test option performance
Do not only recalculate NPV. Test service, capacity, safety, environmental performance, cash exposure and reversibility.
4. Value flexibility
Identify where staged commitment, modularity, contractual options, pilots or deferred decisions reduce irreversible exposure.
5. Set triggers
Specify what evidence would cause expansion, redesign, pause or exit.
The output should distinguish three types of commitment:
Commit now where the option is robust and delay carries material opportunity cost.
Stage where learning has value and future commitment can be preserved.
Defer where uncertainty is high, reversibility is low and the cost of waiting is acceptable.
From Strategy to Execution
Immediate action is to add an assumption stress test to major business cases. Ask teams to show how the recommendation behaves when demand, price, timing or performance moves outside the base case. This is different from adding an arbitrary contingency percentage.
Medium-term capability building requires scenario libraries and operational data. Brivio and colleagues' work is valuable because uncertainty was converted into multiple operating profiles and lifetime scenarios. Organisations can do the same with their own demand, downtime, yield, maintenance and market data.
Long-term positioning means structuring portfolios so that not every initiative depends on the same future. If all growth projects require one commodity price, one regulatory assumption or one scarce capability, the organisation has created portfolio-level fragility even if each individual business case appears sound.
Related article: Strategic Flexibility: Match the Management System to Environmental Turbulence
Related article: Not Every Portfolio Uncertainty Belongs on a Risk Register
Related article: A Transformation Portfolio Must Be Sequenced as a System
Signals to Monitor
Monitor variance between forecast and actual demand, utilisation, resource availability, maintenance effort and unit economics. Persistent variance is not merely forecasting noise; it may indicate that the design assumptions are decaying.
Watch for increasing dependence on one scenario. If an investment only remains attractive under the base case while downside scenarios have become more plausible, governance should reopen the decision.
Track the cost of preserving options. Flexibility itself can become expensive. A modular architecture that is never likely to expand or an unused backup contract can become a permanent tax on the system. Robustness must remain economically disciplined.
Finally, monitor leading indicators that change objectives, not just inputs. Climate exposure, service criticality, regulation or customer tolerance can shift the enterprise from efficiency-seeking to resilience-seeking before financial outcomes visibly deteriorate.
Questions for the Leadership Team
- Which assumption can break this investment fastest?
- Are we designing around an average that rarely occurs in operation?
- What is the cost of being wrong in each direction?
- Which part of the commitment can be staged until uncertainty reduces?
- What option are we giving up by choosing the apparently most efficient configuration now?
- Which trigger would cause us to expand, pause, redesign or exit?
Closing Perspective
The future will not reward the organisation for having produced the most elegant base-case forecast.
It will reward the organisation whose assets, contracts and portfolio choices continue to function when reality departs from the forecast.
Efficiency matters. But when uncertainty is material and commitments are difficult to reverse, the executive task is not to design for the average future. It is to design a system that can survive being surprised.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.