Scenario planning
A structured way to think about the future when prediction is impossible. You don't forecast a single number — you build three to four distinct, internally-coherent stories of how the world could unfold, and stress-test strategy against each. The output isn't a probability distribution; it's a set of named worlds the team can argue with.
Where it comes from
In the late 1960s, two men inside Royal Dutch Shell — Pierre Wack and Ted Newland — were given an unusual brief: study the long-term oil market. Forecasting was the standard tool of the day, and forecasting in 1968 said the world would keep buying cheap oil from a stable Middle East forever.
Wack was sceptical. He had read Herman Kahn's nuclear war scenarios and noticed something: Kahn's value wasn't in predicting which war — it was in giving generals a vocabulary to think in. Wack wondered whether the same trick would work on oil executives. Instead of a forecast he built a small set of plausible futures, each with a name, and walked the Shell board through them.
When OPEC announced the 1973 embargo, Shell was the only major that had already imagined the world it was now living in. Production planning, capital allocation and refinery configuration had been pre-decided against the shock scenario. The competitive position the company gained over that decade is the reason scenario planning is taught in business schools today.
Wack codified the practice in two 1985 Harvard Business Review articles — "Scenarios: Uncharted Waters Ahead" and "Scenarios: Shooting the Rapids" — which remain the canonical primary sources. Peter Schwartz, who succeeded Wack at Shell, took the practice into the rest of the world with The Art of the Long View (1991). Adam Kahane later adapted it for political and social conflict in Transformative Scenario Planning (2012).
What it actually is
Scenario planning is a strategy tool, not a forecasting tool. The distinction matters.
A forecast says: "Oil will be $80 in 2030." A scenario set says: "Here are four futures — Boom, Bust, Walled Garden, Open Roads. In each, oil trades in this band, the regulator does this, our customers behave like that, and our company should do the following."
The forecast invites a single decision. The scenario set invites a conversation about what to do in each world — and a commitment to recognise which one we're in as it unfolds.
Three things separate scenario planning from cousins like sensitivity analysis and Monte Carlo simulation:
- Stories, not numbers. A scenario is an internally consistent narrative with a beginning, middle and end. The 50-row Excel that labels itself "Scenario A" is not a scenario — it's a model.
- Decisions, not probabilities. Wack famously refused to assign probabilities to scenarios. Probabilities collapse the conversation back into expected value. The point is commitment in each world, not the blended forecast.
- Names. Each scenario gets a vivid, world-shape name — "Retreat", "Bridge", "Bloom". The name is what survives the meeting. "Most likely / unlikely" leaks the answer; "Bloom" travels.
Why it works
It works because it solves the two failure modes of business forecasting at once.
The first failure mode is being precisely wrong. A point estimate projected three years out has roughly a 0% chance of being correct, and yet boards routinely allocate capital against one. Scenarios refuse the false precision.
The second failure mode is uselessly hedged. The honest forecaster who says "anywhere from 5% to 25% growth" has produced a useless artefact — nobody can plan against a range that wide. Scenarios refuse the mush.
What scenarios do instead is commit to a small set of distinct worlds, each plausible, each described in enough detail that the team can ask "if we wake up in this one tomorrow, what do we do?". Strategy then crystallises:
- No-regret moves — moves that win in every scenario. Do them now.
- Big bets — moves that pay off only in some. Sequence carefully or buy options.
- Lose-everywhere moves — moves that fail in every world. Drop them.
The third category is the one most companies skip and the one scenarios catch best. A strategy that loses in every plausible future is, by definition, betting on an implausible one.
When it's the right tool
Scenarios fit when the dominant uncertainties are structural and unquantifiable — political, technological, behavioural. They're the right tool for multi-year capital allocation, market entry/exit decisions, regulatory strategy, and contrarian theses arguing the consensus path is unstable.
They're the wrong tool for tactical 3–6 month decisions (sensitivity analysis is faster), for quantifiable risk (Monte Carlo will out-perform), and for teams that won't revisit. Static scenarios decay into theatre within a year.
The full decision logic is in the skill-side reference — the operational chuleta. The version on this page is the why; the skill version is the how.
Worked example — an activist breakup decision
Imagine the board of a mid-cap conglomerate facing an activist demanding a full breakup by year-end. The board's instinct is to reject. The CFO wants scenarios.
Focal question: "Should we accept, reject, or partially concede the breakup demand by Q4?"
Two critical uncertainties chosen:
- Proxy advisor support (ISS / Glass Lewis) — Hostile to the activist ↔ Neutral / sympathetic.
- Public market multiple compression — Compression continues ↔ Multiples reverse with rate cuts.
These aren't independent in a strict statistical sense, but they're decisively independent for this decision: each can move on its own calendar, driven by different mechanisms (governance reform vs macro-rate cycle).
The four worlds:
| Multiple keeps compressing | Multiple reverses | |
|---|---|---|
| Advisors hostile | Stand-off — concede partial spin | Capitulate — full breakup |
| Advisors neutral | Boardroom war — proxy fight likely | Quiet truce — reject demand |
Early signals to watch (next 90 days):
- ISS draft report tone — hostile / neutral / supportive.
- Sector multiple movement vs S&P 500 — ±200 bps threshold.
- Top-5 holder voting pattern in comparable recent activist campaigns.
- Activist's tone in media — escalation vs settlement language.
The wind-vane: if signals 1 and 2 both trip toward Capitulate by week 8, the board moves from reject to negotiated partial spin before the campaign forces the issue. The scenario isn't a prediction — it's a pre-committed contingency.
This is how scenarios actually get used in a boardroom: not as a forecast, but as a trigger plan. Six months later, when the world has settled into one of the four quadrants (or, more often, somewhere on the boundary), the team already knows the move.
Common pitfalls
Fake symmetry. The most common failure: producing one optimistic scenario, one pessimistic, one base case, one "black swan". That's a sensitivity analysis with a costume. Real scenarios are each internally coherent and plausibly the future — not Goldilocks-spaced around a midpoint. If your four scenarios can be ranked from "good" to "bad", you've drawn one variable on two axes.
Naming the scenarios after probabilities. "Most likely" leaks the answer before the conversation begins. Use world-shape names that describe the shape of the world, not its desirability.
Skipping the signals step. A scenario without leading indicators is a museum exhibit. The whole point is to recognise which world you're in as it happens, not after.
Forgetting to revisit. Scenarios decay. Every 12–18 months you re-run the exercise, falsify the dead ones, and ask whether the critical uncertainties still are.
Quant-heavy, narrative-thin. A 50-row spreadsheet labelled "Scenario A" is a model, not a scenario. The story is what the team remembers in the boardroom six months later. If you can't tell the scenario over coffee, it doesn't exist.
Related canon
scenario-analysis— the slide artefact that summarises a scenario set: 4-quadrant grid, signals row, strategy callout. Scenario planning is the process; scenario analysis is the deliverable.planning-fallacy— the cognitive bias scenarios are designed to counter: the systematic optimism baked into single-point forecasts.three-horizons— McKinsey's complementary lens. Horizon 1 maps to no-regret moves; Horizon 3 maps to scenario-contingent big bets.monte-carlo-simulation— the right tool when the uncertainty is quantifiable.
Canonical references
- Pierre Wack — Scenarios: Uncharted Waters Ahead — HBR, Sept-Oct 1985.
- Pierre Wack — Scenarios: Shooting the Rapids — HBR, Nov-Dec 1985.
- Peter Schwartz — The Art of the Long View — Doubleday, 1991.
- Adam Kahane — Transformative Scenario Planning — Berrett-Koehler, 2012.
- Royal Dutch Shell — 40 Years of Shell Scenarios — public retrospective.
- Angela Wilkinson & Roland Kupers — Living in the Futures — HBR, May 2013.
- Pierre Wack — Wikipedia (biographical context).
- Case study: How Shell anticipated the 1973 oil crisis — Polytechnique Insights.