Contents
Use this when the future is genuinely uncertain — not risky in a quantifiable way, but unknowable — and you need to make commitments that will play out across multiple possible worlds. Scenario planning constructs a small set of divergent, plausible futures and stress-tests your strategic options against each one, replacing the dangerous illusion of prediction with the discipline of preparation.
Section 1
What This Tool Does
The default mode of strategic planning is forecasting. You take the present, extrapolate it forward, and build your plan around the most likely future. This works tolerably well when the environment is stable — when the variables that matter are moving in predictable directions at predictable speeds. It fails catastrophically when they aren't. And the insidious part is that you don't discover the failure until the future arrives and your plan collides with a reality it never contemplated.
Herman Kahn understood this at RAND Corporation in the 1960s, where he was paid to think about thermonuclear war — a domain where the cost of being surprised by an unconsidered future was, quite literally, civilisational extinction. Kahn didn't try to predict what the Soviet Union would do. He constructed multiple scenarios — detailed, internally consistent narratives of how the future might unfold — and then asked what the United States should do differently depending on which scenario materialised. The method migrated from nuclear strategy to corporate strategy in the early 1970s, when Pierre Wack and Ted Newland at Royal Dutch Shell's Group Planning division used it to prepare for the possibility of an oil price shock. Shell's competitors were running single-point forecasts that assumed continued cheap oil. Shell was rehearsing a world where Arab producers restricted supply and prices quadrupled. When OPEC did exactly that in 1973, Shell was the only major oil company that had already thought through the implications. Within two years, Shell moved from the seventh-largest oil company in the world to the second.
The mechanism is not prediction. That distinction matters more than anything else about the tool. Scenario planning works by expanding the space of futures you've considered, forcing you to confront possibilities your planning process would otherwise suppress — and then testing whether your strategy is robust across that expanded space or dangerously dependent on one particular future arriving on schedule. The cognitive shift is from "What will happen?" to "What could happen, and what would we do?" That second question is harder, slower, and far more valuable.
Most organisations carry an implicit scenario in their heads: the future looks roughly like the present, plus or minus some growth. Scenario planning makes that implicit assumption explicit, places it alongside three or four radically different assumptions, and asks which strategic moves perform well across most of them. The options that only work if one specific future materialises are fragile. The options that perform reasonably well across three or four divergent futures are robust. That distinction — fragile versus robust — is the tool's primary output. Not a prediction. A portfolio of preparations.
Section 2
How to Use It — Step by Step
Instructions on the left. Worked example — a mid-size European SaaS company selling compliance software to financial institutions, deciding whether to invest heavily in AI-native product capabilities or double down on its existing rules-based platform — on the right.
Step 1 — Identify
Surface the critical uncertainties that will shape the decision
List the external forces that could materially affect your strategic options over the relevant time horizon (typically 5–10 years, though shorter horizons work for faster-moving industries). You're looking for factors that are both highly uncertain and highly impactful. Regulatory changes, technology shifts, macroeconomic conditions, competitive dynamics, customer behaviour changes. Generate 10–15 candidates, then ruthlessly prioritise. You need two — the two uncertainties that are most uncertain and most consequential. These become the axes of your scenario matrix. If a factor is important but predictable (demographics, for instance), it's a given, not an uncertainty — it belongs in every scenario, not as a differentiator between them.
Worked example
Compliance SaaS — critical uncertainties
The team generates twelve forces. After debate, two emerge as both genuinely uncertain and strategically decisive. Axis 1: Regulatory posture toward AI in financial compliance — will regulators embrace AI-driven compliance (allowing automated audit, reducing human oversight requirements) or resist it (mandating human-in-the-loop, restricting algorithmic decision-making)? Axis 2: Speed of AI capability maturation — will large language models and domain-specific AI reach reliable, auditable performance within 3 years, or will hallucination and explainability problems persist for 7+ years? Both axes are genuinely uncertain. Both would reshape the competitive landscape.
Step 2 — Construct
Build four distinct scenario narratives from the 2×2 matrix
Cross your two uncertainties to create a 2×2 matrix with four quadrants. Each quadrant is a scenario — a plausible, internally consistent story about how the future unfolds. Give each scenario a memorable name (Shell famously used evocative titles). Then flesh out each one: What does the competitive landscape look like? How do customers behave? What new entrants appear? What business models thrive? The narratives should be vivid enough that your team can inhabit them mentally — not bullet points, but stories. Each scenario should feel like it could actually happen. If a scenario feels absurd, either your uncertainties are wrong or you haven't developed the narrative enough.
Worked example
Four futures for compliance SaaS
"Autonomous Compliance" (regulators embrace AI + AI matures fast): Regulators issue frameworks for AI-driven audit. Automated compliance becomes table stakes. Rules-based platforms are legacy. New entrants with AI-native architectures capture market share rapidly. "The Careful Transition" (regulators embrace AI + AI matures slowly): Regulators signal openness but AI isn't reliable enough yet. Hybrid solutions win — rules-based engines with AI augmentation. Incumbents with strong rule engines have time to layer AI on top. "The Human Firewall" (regulators resist AI + AI matures fast): AI is technically capable but regulators mandate human oversight for all compliance decisions. AI becomes a back-office efficiency tool, not a customer-facing product. Rules-based platforms with human workflow remain dominant. "Steady State" (regulators resist AI + AI matures slowly): Neither the technology nor the regulation shifts dramatically. The current competitive landscape persists. Incremental improvement wins. Radical bets on AI are premature and expensive.
Step 3 — Test
Stress-test each strategic option against all four scenarios
Take the strategic options you're evaluating — in this case, "invest heavily in AI-native product" versus "double down on rules-based platform" — and ask how each performs in each scenario. Not just "does it work?" but "what happens to revenue, competitive position, customer retention, and margin?" Be specific. Assign rough qualitative or quantitative assessments. A simple scoring system works: strong positive (+2), moderate positive (+1), neutral (0), moderate negative (−1), strong negative (−2). The goal isn't precision — it's pattern recognition. Which options are robust (positive or neutral across most scenarios) and which are fragile (brilliant in one scenario, catastrophic in another)?
Worked example
Stress-testing the two options
Option A: Heavy AI investment. Performs brilliantly in "Autonomous Compliance" (+2). Performs well in "Careful Transition" (+1) — the company is ahead of the curve. Performs poorly in "Human Firewall" (−1) — the investment is largely wasted, and competitors who invested in workflow tools pull ahead. Neutral in "Steady State" (0) — the R&D spend drags margins without payoff. Option B: Double down on rules-based. Performs poorly in "Autonomous Compliance" (−2) — the company becomes a legacy vendor within five years. Performs moderately in "Careful Transition" (+1) — the strong rule engine is a foundation for gradual AI integration. Performs well in "Human Firewall" (+2) — the product is exactly what the market needs. Performs well in "Steady State" (+1). Option A: total score +2, high variance. Option B: total score +2, lower variance but catastrophic downside in one scenario.
Step 4 — Identify
Find robust moves, hedges, and signposts
Three outputs matter. Robust moves are actions that perform well regardless of which scenario materialises — pursue these immediately. Hedges are low-cost investments that protect you in specific scenarios without committing you fully. Signposts are observable, real-world indicators that tell you which scenario is actually unfolding, so you can adjust before your competitors recognise the shift. Signposts are the most underrated output. They transform scenario planning from a one-time exercise into a continuous monitoring system.
Worked example
Robust moves, hedges, and signposts
Robust move: Build a modular architecture that can integrate AI capabilities without replacing the rules engine. This performs well in every scenario — it's the foundation for AI-native products if the market demands them, and it preserves the rules-based product if it doesn't. Hedge: Hire a small AI research team (5–8 people) and fund proof-of-concept projects. This costs 3–4% of revenue but keeps the option open for rapid scaling if "Autonomous Compliance" starts materialising. Signposts to monitor: (1) First major regulator to issue formal guidance on AI in compliance audit — and whether the guidance is permissive or restrictive. (2) Error rates in domain-specific AI models published in academic benchmarks. (3) Whether any top-20 bank publicly deploys AI-driven compliance in production. When two of three signposts point toward a specific scenario, accelerate the corresponding strategy.
Step 5 — Embed
Institutionalise scenario monitoring and strategy adaptation
Scenario planning fails when the scenarios go into a binder and the organisation reverts to single-point forecasting. Assign signpost monitoring to specific people. Review signposts quarterly. When the evidence accumulates that one scenario is becoming more likely, trigger pre-agreed strategic adjustments. The scenarios should be living documents — updated annually as uncertainties resolve and new ones emerge. The goal is not to have predicted the right future. The goal is to have prepared for enough futures that you're never caught flat-footed.
Worked example
Embedding the system
The compliance SaaS company assigns its VP of Strategy to publish a quarterly "Scenario Signal Report" — a one-page document tracking each signpost and assessing which scenario the evidence currently favours. In Q2, the EU publishes draft guidance that is cautiously permissive toward AI in compliance. One signpost triggered. The company increases AI team headcount from 6 to 12 and accelerates the modular architecture timeline. They haven't committed to a full AI pivot — but they've moved from hedge to active preparation. The scenario framework gave them the language and the triggers to make this shift without a crisis.
Section 3
When It Works Best
✓
Ideal Conditions for Scenario Planning
| Dimension | Best fit |
|---|---|
| Decision type | Large, irreversible, or slow-to-reverse commitments — market entry, major capital allocation, platform architecture, M&A strategy. The tool's overhead (weeks of work for a full exercise) is only justified when the stakes are high enough that being wrong about the future is genuinely costly. |
| Uncertainty profile | Environments with deep structural uncertainty — where the key variables could plausibly go in multiple directions and no amount of data will resolve the ambiguity in advance. Regulatory shifts, technology discontinuities, geopolitical realignments. If the future is merely risky (quantifiable probabilities), a decision tree or Monte Carlo simulation is more precise. |
| Time horizon | 3–10 years. Shorter than 3 years and the future is usually constrained enough that forecasting works. Longer than 10 years and the combinatorial explosion of uncertainties makes any scenario set feel arbitrary. The sweet spot is the range where current trends could plausibly continue or break. |
| Organisational readiness | Requires senior leadership willing to entertain uncomfortable possibilities. Scenario planning surfaces futures where the current strategy fails — and that's the point. If the CEO treats the exercise as a validation of the existing plan, the scenarios will be sanitised into irrelevance. |
| Team composition | Cross-functional and, ideally, including external perspectives. The scenarios are only as good as the range of mental models in the room. A team of strategists will produce strategist scenarios. Add an engineer, a regulator, a customer, and a contrarian, and the scenario set becomes genuinely challenging. |
| Information environment | Works best when there's enough information to construct plausible narratives but not enough to assign reliable probabilities. The tool fills the gap between "we know nothing" (where it's useless) and "we can model this quantitatively" (where it's unnecessary). |
Section 4
When It Breaks Down
⚠
Failure Modes
| Failure pattern | What goes wrong | What to use instead |
|---|---|---|
| Scenario as forecast | The team unconsciously assigns probabilities to the scenarios and gravitates toward the "most likely" one. The other three become decorative. The entire exercise collapses back into single-point planning with extra steps. This is the most common failure and the hardest to prevent because it exploits the brain's deep preference for prediction over preparation. | Pre-Mortem on the "most likely" scenario to force engagement with alternatives |
| Axes too correlated | The two uncertainties chosen for the matrix are not independent — they tend to move together. This produces two plausible scenarios (the diagonal) and two implausible ones (the off-diagonal). A 2×2 matrix with only two real quadrants is a waste of the framework. | Spend more time on Step 1; test independence by asking "Could Axis 1 go high while Axis 2 goes low?" If not, choose different axes. |
| Sanitised scenarios | Political dynamics within the organisation prevent genuinely challenging scenarios from being developed. Nobody constructs the scenario where the CEO's pet initiative fails, or where the company's core market disappears. The scenarios become variations on the status quo — mildly optimistic, mildly pessimistic, and two flavours of "roughly the same." | Use an external facilitator with no organisational loyalty; include at least one "nightmare" scenario by design |
| Too many scenarios | Teams generate six, eight, ten scenarios to "be thorough." Cognitive overload sets in. Nobody can hold ten futures in their head simultaneously, so the stress-testing becomes superficial. The 2×2 matrix exists for a reason: four scenarios is the maximum that most teams can meaningfully engage with. | Stick to the 2×2 structure; if a third uncertainty is critical, run a separate exercise rather than expanding to a 2×2×2 cube |
| No signposts defined | The scenarios are developed, the strategies are stress-tested, and then the organisation returns to business as usual with no mechanism for detecting which scenario is actually unfolding. By the time the answer is obvious, it's too late to adjust. Scenario planning without signpost monitoring is a strategy offsite, not a decision system. | Require each scenario to have 2–3 observable, time-bound signposts; assign monitoring responsibility by name |
| Rapidly shifting environment | In domains where the relevant uncertainties change faster than the scenario cycle (quarterly or faster), the scenarios become stale before they're useful. Startups in hypergrowth, markets in acute crisis, or domains with weekly regulatory shifts outpace the tool's natural tempo. | OODA Loop for faster-cycle adaptation; Cynefin Framework to classify whether the domain is complex (probe-sense-respond) rather than complicated (analyse-plan-act) |
The most dangerous failure mode is scenario as forecast — and it's nearly invisible while it's happening. The tell: watch how the team talks about the scenarios after the exercise. If they say "Scenario B is the most likely," the framework has already collapsed. The whole architecture of scenario planning depends on treating the scenarios as equally plausible thought experiments, not as probability-weighted predictions. The moment one scenario becomes "the base case," the others become afterthoughts, and you're back to the single-future planning that the tool was designed to replace. The defence is cultural, not procedural: the facilitator must repeatedly, almost annoyingly, insist that the question is never "Which scenario will happen?" but always "What would we do if this one happened?"
A subtler version of the same failure: teams that build robust strategies but never define signposts. They've done the intellectual work of preparing for multiple futures, then thrown away the monitoring system that would tell them which future is arriving. It's like buying insurance and then never filing a claim because you forgot you had the policy.
Section 5
Visual Explanation
Scenario matrix for the compliance SaaS worked example. Two critical uncertainties create four plausible futures. Strategic options are scored in each quadrant. Signposts (bottom) indicate which scenario is materialising.
Section 6
Pairs With
Scenario planning is the strategic equivalent of a stress test. It tells you which options are fragile and which are robust. But it doesn't generate the options, define the problem, or execute the response. The tools around it determine whether the scenarios produce action or just interesting conversation.
Use before
First Principles Thinking
Before constructing scenarios, decompose the industry's fundamental economics and physics. What must be true for any future to hold? First principles separate the structural constants (which belong in every scenario) from the genuine uncertainties (which differentiate the scenarios). Without this step, teams waste scenario axes on factors that aren't actually uncertain.
Use before
Cynefin Framework
Classify the decision domain before choosing your tool. Scenario planning belongs in the "complex" or "complicated" domains of Cynefin — where cause and effect exist but aren't fully predictable. In the "chaotic" domain, you need to act first and sense later; scenarios are too slow. In the "obvious" domain, you don't need them at all.
Use after
Pre-Mortem
Run a pre-mortem on your chosen strategy within each scenario. "It's 2028, Scenario B materialised, and our strategy failed. Why?" This catches implementation risks that the scenario matrix — which tests strategic direction, not execution — will miss.
Use after
Reversible vs. Irreversible Decisions
Once scenarios reveal which moves are robust and which are fragile, classify each move by reversibility. Robust moves can be committed to immediately. Fragile moves that are also irreversible should be deferred until signposts clarify the landscape. Fragile moves that are reversible can be treated as experiments.
Use after
Decision Matrix
When scenario planning surfaces a robust move that still involves choosing among several implementation paths, a weighted decision matrix can evaluate those paths on criteria that the scenario exercise identified as critical — adaptability, cost to reverse, speed to deploy.
Mental model
Second-Order Thinking
Scenario narratives improve dramatically when the team applies second-order thinking within each scenario. "If regulators embrace AI, then what? Incumbents scramble to acquire AI startups. Then what? Acquisition prices spike. Then what? Only well-capitalised players can afford the transition." This depth is what separates a scenario from a headline.
Section 7
Real-World Application
Royal Dutch Shell — navigating the 1973 oil crisis before it happened
The scenario
In 1971, Royal Dutch Shell was the seventh-largest oil company in the world, operating in an industry that had enjoyed two decades of cheap, abundant crude. Every major oil company's planning function assumed this would continue. The question wasn't whether oil would stay cheap — that was treated as a given — but how fast demand would grow and where to build the next refinery. Pierre Wack, a planner in Shell's London headquarters, thought the assumption itself was the vulnerability.
How the tool appliedWack and his colleague Ted Newland identified two critical uncertainties that the industry was ignoring. First: the political stability of the OPEC cartel and the willingness of Arab oil-producing states to use supply as a geopolitical weapon. Second: the response of Western governments to a supply shock — would they coordinate or fragment? Wack constructed scenarios that included a world where OPEC restricted supply and oil prices rose dramatically. He didn't predict the 1973 embargo specifically. He constructed a plausible future in which something like it happened and asked: "If this world materialises, what should Shell have done differently starting now?"
The scenarios were presented to Shell's managing directors — the Committee of Managing Directors, Shell's top decision-making body — in a series of sessions through 1972 and early 1973. Wack's genius was not the scenarios themselves but how he presented them. He didn't ask the directors to believe the oil shock would happen. He asked them to mentally inhabit a world where it had happened and notice what they wished they'd done. This reframing — from prediction to preparation — was what made the exercise actionable.
What it surfacedThe scenario exercise revealed that Shell's entire downstream strategy (refining, distribution, retail) was optimised for a world of cheap crude and growing demand. In a supply-constrained world, the strategic priorities inverted: refining flexibility mattered more than refining capacity; diversification of crude sources mattered more than volume contracts with the cheapest supplier; and the ability to rapidly adjust product mix (heating oil vs. gasoline vs. petrochemicals) became a decisive competitive advantage. Shell began making operational changes — diversifying supply sources, investing in refining flexibility, building crude oil inventory — months before the October 1973 embargo.
The non-obvious factorShell's advantage wasn't that it predicted the oil crisis. Several analysts and commentators had warned about OPEC's growing power. The advantage was that Shell had already rehearsed its response. When the embargo hit, Shell's competitors entered a period of strategic paralysis — their planning assumptions had been invalidated and they had no alternative framework to fall back on. Shell's managers, by contrast, recognised the unfolding situation as one of their scenarios and activated pre-considered responses. The cognitive preparation — having already thought through the implications — was worth more than any specific operational hedge. Wack later described this as the real product of scenario planning: not better predictions, but "the microcosm of the decision-maker's mind" being prepared for a wider range of futures. By 1975, Shell had risen from seventh to second among the world's oil majors, a position it maintained for decades.
Section 8
Analyst's Take
Faster Than Normal — Editorial View
Scenario planning endures because it solves a problem that no other strategy tool addresses: the human inability to take seriously a future that differs from the present. Forecasting, by design, extrapolates from what's happening now. Decision trees require probability estimates that, in conditions of genuine uncertainty, are fabricated precision. Scenario planning is the only widely used tool that forces a leadership team to inhabit — emotionally, not just intellectually — a future where their current assumptions are wrong. That emotional rehearsal is the mechanism. Wack understood this. Most practitioners don't, which is why most scenario exercises produce interesting documents and zero behavioural change.
The failure mode I see most often is what I'd call "scenario tourism." The team visits four futures, finds them intellectually stimulating, and returns home to the same strategy they had before. The scenarios were vivid. The discussion was rich. Nothing changed. The root cause is almost always the same: the team never completed Step 4. They built scenarios and stress-tested options but didn't define signposts or robust moves. Without those outputs, the exercise has no mechanism for connecting to actual decisions. It's strategic entertainment. The fix is to make the final deliverable not the scenarios themselves but a one-page document with three columns: "Do now" (robust moves), "Prepare to do" (hedges), and "Watch for" (signposts). If that page doesn't exist, the exercise didn't finish.
The highest-leverage modification is to run the scenario exercise with the people who will execute the strategy, not just the people who set it. Shell's breakthrough wasn't that Pierre Wack built brilliant scenarios in London. It was that he presented them to the managing directors and, crucially, to the country managers who ran Shell's operations in dozens of markets. Those operators then carried the scenarios into their own planning. When the crisis hit, the response was distributed — hundreds of managers recognising the situation and adjusting locally, not waiting for headquarters to issue new instructions. Scenario planning's value scales with the number of decision-makers who have mentally rehearsed the scenarios. Confine it to the C-suite and you get a smarter strategy document. Push it to the operating level and you get an organisation that adapts faster than its competitors in real time.
Section 9
Top Resources
01
The Art of the Long View — Peter Schwartz (1991)
BookThe definitive practitioner's guide. Schwartz led scenario planning at Shell after Wack and later founded Global Business Network. This book translates Shell's methodology into a step-by-step process accessible to any organisation. Chapters 3–6 cover the mechanics; Chapter 7 on "rehearsing the future" captures the psychological dimension that most how-to guides miss entirely. If you read one book on scenario planning, this is it.
02
Book
Not a scenario planning book, but essential for understanding why the tool works. Kahneman's research on the planning fallacy, overconfidence, and the availability heuristic explains precisely why organisations default to single-point forecasts and why structured consideration of alternative futures is a genuine cognitive intervention. Read Part III on overconfidence to understand the disease that scenario planning treats.
03
Book
Grove's account of Intel's strategic inflection points is scenario planning in practice, even though he doesn't use the term. His framework for detecting "10X changes" — forces that alter the competitive landscape by an order of magnitude — is the best guide to identifying the critical uncertainties that belong on your scenario axes. Chapter 5 on signal versus noise is particularly relevant for defining signposts.
04
Scenarios: The Art of Strategic Conversation — Kees van der Heijden (1996)
BookVan der Heijden was Wack's successor at Shell and this is the most rigorous academic treatment of the Shell methodology. Denser than Schwartz, more theoretical, but invaluable for understanding why certain scenario structures work and others don't. His distinction between "predetermined elements" and "critical uncertainties" is the conceptual foundation that makes Step 1 of the process work. Best for practitioners who've already run a few exercises and want to deepen their craft.
05
Book
Porter's industry analysis framework — the five forces — provides the structural vocabulary for building scenario narratives. Each scenario should describe how the five forces shift: Do barriers to entry rise or fall? Does buyer power increase? Do substitutes emerge? Without this structural discipline, scenarios become vague stories rather than analytically grounded futures. Chapter 13 on industry scenarios explicitly connects Porter's framework to scenario methodology.
Decision Tools Library — Browse by phase
Why this matters next
mental modelsSecond-Order Thinking
First Principles Thinking applied the Second-Order Thinking mental model
mental modelsFirst Principles Thinking
First Principles Thinking applied the First Principles Thinking mental model
mental modelsLeverage
First Principles Thinking applied the Leverage mental model
mental modelsOODA Loop
First Principles Thinking applied the OODA Loop mental model
mental modelsNarrative
First Principles Thinking applied the Narrative mental model
mental modelsScale
First Principles Thinking applied the Scale mental model
Continue exploring
DMSHCADT
Decision tool
Decision Matrix
Score options across weighted factors to identify a clear winner
Decision tool
Six Thinking Hats
Examine a decision from six cognitive perspectives — facts, emotions, risks, ben
Decision tool
Cost-Benefit Analysis
Quantify the economic trade-offs of each option to determine net value
Decision tool
Decision Tree
Map sequential outcomes with probabilities and expected values to find the highe
More like this, in your inbox
I send a newsletter every week — free, no spam, unsubscribe anytime.