Royal Dutch Shell's planning group, under Pierre Wack in the early-to-mid 1970s, built scenario planning as a deliberate alternative to single-point forecasting. Conventional forecasts had repeatedly failed to anticipate the oil shocks reshaping the industry. Wack documented the method in his 1985 Harvard Business Review pieces "Scenarios: Uncharted Waters Ahead" and "Scenarios: Shooting the Rapids." The premise is easy to forget under deadline pressure. The future isn't one number. It's a range of plausible, internally consistent stories, and a strategy built for only one of them will be caught flat-footed by the others.
How does scenario planning work?
The driver logic starts from the method's own first move. It identifies driving forces and separates the predetermined ones, things that will happen regardless, from the uncertain ones. From the uncertain forces you select the two critical uncertainties. Each has to be highest on impact, meaning it would substantially change the right strategy, and highest on uncertainty, meaning you can't predict which way it resolves. Cross the two and you get a 2×2 matrix of four scenarios. Each scenario is an internally consistent narrative with named early indicators, observable signs that tell you in real time which scenario is unfolding. You don't then pick your favorite scenario and bet on it, because that's the single-point-forecast error the method exists to avoid. You "wind-tunnel" the current plan against all four. That finds the no-regrets moves that hold up regardless, and the decision points where you must commit differently depending on which scenario emerges.
Building it correctly means putting as much rigor into choosing the axes as into writing the scenarios, and axis selection is the step most real exercises rush. Screen the candidate driving forces on impact and uncertainty, and pick the two that score highest on both. Don't pick the two most comfortable to discuss, and don't pick the two the CEO already has a hunch about. Choose the wrong axes and every scenario downstream is beside the point. Once you have four scenarios, name at least one early indicator for each. Without indicators to monitor, the exercise can't tell you which future is arriving.
Advertisement
What are the common mistakes in scenario planning?
Four mechanism-level blind spots recur. First, and most common of all, is scenario paralysis. Organizations produce beautifully written narratives and stop there. They never name the no-regrets moves, and they never commit capital to them. Second comes false precision in the narrative. A vivid, well-written story feels more credible than it has any right to, purely because fluent narrative persuades whatever its actual probability. The AI lessons teach the identical caution about a fluent AI answer, and it applies here too. A compelling story isn't evidence that it's likely. Third, the exercise collapses if you choose the wrong two critical uncertainties. Pick comfortable axes over the two that actually matter and you'll get four superficial scenarios, or four that are essentially the same story wearing different names. Fourth, scenario planning generates strategic options and robustness. It never produces a forecast or a decision by itself. You still have to pair it with a real capital-allocation process and live monitoring against the named early indicators, or it's just an expensive workshop with no operational teeth.