Reading scenarios and transition drivers

A scenario is a coherent description of what could happen if a set of assumptions holds. It is useful in energy decisions because an asset's operating life may extend beyond the period for which prices, policies and technologies are reasonably visible. A scenario does not remove uncertainty; it shows how outcomes change when premises change. In World Energy Outlook 2025, the International Energy Agency distinguishes exploratory pathways based on specified starting conditions and policies from a normative pathway that begins with desired goals. It explicitly states that its scenarios are not forecasts. Do not infer the probability of a normative pathway or treat a price in that pathway as a local purchase offer. Understand the underlying logic, then connect it to the organization's data, contracts and decision horizon.

Start with the investment question and the date after which commitment becomes difficult to reverse. Choosing a new plant within a month differs from developing an organizational capability that can expand over several years. Identify variables that affect the choice: service-demand growth, fuel and electricity prices, grid emissions intensity, connection timing, technology availability and carbon prices where relevant. Distinguish an external variable, such as a public network's completion date, from an action the organization controls, such as training timing. An internal delivery failure should not become a market scenario that removes responsibility. Avoid a long list of small variables when two explain most decision reversals. Begin with influential drivers, then test whether an omitted factor could overturn the conclusion.

Construct coherent cases rather than combining every favorable input into an optimistic case and every unfavorable input into a pessimistic one without explanation. Expanding electrification can increase demand for network connections and equipment, so opportunities to reduce fuel use may coincide with connection constraints and supply delays. Fuel and electricity prices may also move together when the same fuel influences electricity generation. Explain the causal relationship within each group and distinguish a test assumption from an evidence-supported relationship. When using the endpoints of a price range, add a shock that reveals a different kind of exposure, such as a one-year connection delay. Moving every price by the same percentage can miss that vulnerability. The purpose is to reveal weaknesses in the strategy rather than give each alternative an attractive result on its own worksheet.

This result belongs to a declared decision rule; it does not prove that efficiency is best in every sense. Electrification has the highest value in the orderly case, while efficiency avoids a large sacrifice in the shock case. If the board requires a non-negative present value in every test case, full electrification is excluded before ranking. A positive average must not silently override that condition. The shock might be implausible, or the hybrid might contain an expansion option omitted from the model. In either situation, correct the assumption or value the option with evidence and its cost, then compare again. Do not assign free flexibility merely because an alternative is called hybrid, or add an option whose benefit is already included in the original values.

If probabilities are needed, establish their basis and test how changing them affects the recommendation. In the worked case, assume solely for sensitivity analysis that the probability of the orderly transition is p and the remaining two cases share the rest equally. Expected value becomes 55 + 15p for efficiency, minus 10 + 120p for electrification, and 25 + 60p for the hybrid. Electrification exceeds efficiency when p is greater than about 0.619, and exceeds the hybrid above about 0.583. These are arithmetic switching thresholds, not evidence that the actual probability exceeds them. They show the confidence needed to defend a choice dependent on the orderly future. A strict downside constraint can still prohibit the choice even when expected value becomes attractive.

Connect scenarios to a learning process. For each important driver, specify an observable signal, its source, a review frequency and an authorized response when a threshold is crossed. A binding connection agreement is stronger evidence than an announcement of intended network expansion. A performance test under actual process conditions is stronger than a marketing promise. One unusual month of prices need not establish a permanent shift; require an appropriate period average or an institutional event before changing the investment. Separate a signal for updating an estimate from a signal for stopping unsafe work. Set an expiry date for deferral: waiting for information that arrives only after the procurement window closes can become an undeclared investment decision.

Check each case from three directions. First, is service provision equal across alternatives, or does one look profitable because it fails to meet demand? Second, are common variables treated consistently, or does the preferred project alone receive cheaper electricity without a supporting contract? Third, does the horizon cover relevant commitments and terminal costs, such as decommissioning or contract expiry? Give influential unknowns their own row rather than hiding them in a high discount rate. A discount rate cannot repair an impossible network connection or make an unknown probability known. Present the conclusion conditionally: this alternative remains non-negative in the three cases tested. State that those cases do not exhaust all possible futures, and identify the practical limit this places on the recommendation.