Scenario Planning

When you can’t predict the future, don’t bet on one — sketch a few distinct, plausible futures and get ready for the range.

The Future of Technology · Lesson 24 · 9 min read

Base rates sharpen your odds — but some futures are so uncertain that even a good probability isn’t enough to act on. Will AI reshape your industry in 3 years or 15? Will energy get cheap or scarce? Betting everything on a single guess is how organizations get blindsided. There’s a discipline militaries, oil companies, and good strategists use precisely because they can’t predict: instead of forecasting one future, they prepare for several. Hold the question: when you truly can’t know what happens, what’s better than a single best guess?

Don’t predict one future — sketch a few

Scenario planning flips the goal from “what will happen?” to “what are the few distinct ways this could plausibly go, and am I ready for each?” You build a small set (usually 3–4) of coherent, different futures — not wild guesses, but internally-consistent stories anchored in real forces. The point isn’t to pick the winner; it’s to stop being surprised. This is “think in ranges” (lesson 5) turned into a method: instead of a single number, you hold a spread of futures and plan for the whole spread.

Build scenarios around the driving uncertainties

Good scenarios aren’t random — they’re built around the two or three uncertainties that matter most and that you genuinely can’t predict. Pick the biggest “swing factors” (e.g., does the tech get cheap fast or slowly? and do regulators clamp down or open up?), then cross them to generate distinct corners: cheap-and-open, cheap-and-restricted, expensive-and-open, and so on. Each corner is a scenario. This forces you past your favorite story into futures you’d otherwise ignore — including the ones you’re quietly hoping won’t happen.

Scenarios for “AI in my field over 5 years,” built from two swing factors (capability growth × adoption speed):
• Fast capability + fast adoption → “rapid transformation.”
• Fast capability + slow adoption → “powerful tools, slow uptake.”
• Slow capability + fast adoption → “hype outruns reality.”
• Slow both → “gradual, incremental change.”
Four coherent futures — now you can ask what you’d do in each, instead of betting on one.

Aim for robustness, not prophecy

The payoff is robustness: choices that hold up reasonably well across all your scenarios beat choices that win big in one and collapse in the others. You start asking better questions — “what’s my move if the slow future happens? what early signals would tell me which scenario we’re entering?” — and you spot leading indicators to watch (recall signal vs noise, lesson 2). Scenario planning doesn’t pretend to know the future; it makes you hard to surprise and ready to adapt, which for genuinely uncertain futures is worth far more than a confident wrong prediction.

Using scenarios to make a robust choice:
1. A company faces huge uncertainty about how fast a technology will mature.
2. It builds three scenarios: fast, medium, and slow maturity — each a coherent story.
3. Option A (bet everything on “fast”) wins hugely if fast, but bankrupts them if slow. Option B (build optionality — small bets they can scale up) does fine in all three.
4. They choose B for robustness, and list early signals (adoption numbers, cost curves) that will tell them which scenario is unfolding — so they can lean in when the evidence arrives.