What is scenario planning?

Scenario planning develops several internally consistent accounts of how the future could unfold, in order to change the mental models of the people who will have to decide. It is not forecasting, it cannot be scored, and the case study always quoted alongside it is a participant account rather than evidence of effect.

What it is for#

Wack was explicit that the purpose was changing what managers believed was possible rather than improving prediction. A scenario exercise succeeds when a team that meets an unfamiliar situation has already thought inside a world where it happened, so the response is recall rather than invention.

That is a claim about readiness, and the reason the method is not evaluated the way a forecast is. A forecast carries a probability and can be scored against what happened. Scenarios are deliberately not ranked by likelihood, which is the source of their value and the reason nobody can tell you whether they work.

The Shell story, and what it does and does not support#

The story is that Shell's scenario group had rehearsed an oil supply disruption before the 1973 shock arrived, and that the company therefore responded faster than competitors. The first half is well documented. The second comes from accounts written by people who ran the scenarios, and it has never been benchmarked against how comparable firms actually responded.

It is the most quoted case in the foresight literature and it is a participant account. Cited as origin and intent, it is sound. Cited as evidence that scenario planning improves performance, it is doing work it cannot carry, and this research treats it the second way nowhere.

What a scenario exercise actually involves#

• Name the decision the scenarios are for. Exercises without one produce interesting documents and change nothing.
• Separate what is predetermined from what is genuinely uncertain. The first is usually larger than people assume.
• Build two to four accounts that are internally consistent and meaningfully different, each with a mechanism rather than a mood.
• Work out what each would mean for the decision, and what early sign would indicate you are in that world.
• Keep the early signs, and look at them. This is the step that is skipped, and without it the exercise has no output that anybody returns to.

Where it sits against forecasting#

They answer different questions and the difference matters for what can be claimed. Forecasting produces a probability that resolves, so calibration training can be measured and shown to work. Scenario planning produces preparedness, which cannot be scored, so the method has a sixty-year record of use and almost no evidence base. That is no argument against it, and a reason to state plainly which of the two you are doing.

The AI version of the exercise#

Generating scenarios is now cheap, which threatens the part that did the work. The value was never in the documents; it was in a group arguing about what was predetermined and what was not, and arriving somewhere they had not started. A set of plausible scenarios produced in seconds and read rather than argued over delivers the artefact without the rehearsal, and the convergence problem applies here as elsewhere, since the scenarios that arrive are the ones that currently sound plausible. See does AI make strategy generic.

What this does not establish#

That scenario planning improves decisions or performance. No controlled evaluation exists, the canonical case is a participant account, and the method's own logic makes it difficult to evaluate, since its stated output is a change in the mental models of the participants.

Key sources

• Wack, P. (1985). Scenarios: uncharted waters ahead. Harvard Business Review.

Related SuperSkills research#

• Models of big picture thinking
• What is calibration training
• Big Picture Thinking