What is the difference between forecasting and scenario planning?

Forecasting estimates what is most likely to happen. Scenario planning tests a decision against several plausible futures so the underlying assumptions are forced into view.

Is scenario planning better than forecasting?

Neither is better on its own. A forecast suits situations where the variables can be estimated with tolerable confidence, while scenario planning is more useful when the decision has to survive materially different futures.

When should you use forecasting instead of scenario planning?

Use forecasting when the decision depends on conditions that are stable enough for a projected range to mean something. Use scenario planning when uncertainty is broader and the decision would be exposed if only one future were considered.

How do assumptions fit into forecasting and scenario planning?

Assumptions are the claims that both methods rest on, whether or not a team writes them down. In forecasting, they sit behind the projected range. In scenario planning, they shape which futures are treated as plausible and which signals should trigger a rethink.

Forecasting and Scenario Planning Both Dodge the Decision

A chief executive I worked with once had both forecasting and scenario planning on the table for a plant expansion: a five-year demand curve to one decimal place, four named futures on coloured paper, and three months before the lease option expired. He still could not tell me what had to be true for the plant to earn its keep.

I want the live assumption in the decision pack, because that is usually where the failure is hiding. Neither the forecast nor the scenario set is the output that matters. The assumption is.

Forecasting and scenario planning are assumption-testing tools whose value depends on whether they surface enough doubt to justify committing resources.

Forecasting and scenario planning ask different questions, then stop too early

Even the Bank of England had to admit one official future was not enough. After an inflation forecasting embarrassment, it commissioned Ben Bernanke's 2024 review, which surveyed six peer central banks. The verdict was plain: under high uncertainty, a baseline forecast should be supplemented with alternative scenarios. That is the right instinct, but Bernanke still did not ask which assumption was carrying the policy decision.

Question	Forecasting	Scenario planning	What I use instead
Starting move	Build a baseline view of the future	Describe several plausible futures	State the live decision and its Purpose
Main question	What is most likely?	What could happen?	What would have to hold for this option to work?
Respectable output	A number, range, or central case	A deck, matrix, or set of narratives	Named assumptions ranked by influence and confidence
Usual hiding place	Precision turns into authority	Variety turns into theatre	The weak assumption is exposed early
What survives after approval	A revised forecast cycle	Another workshop next year	A monitoring trigger tied to the decision

The forecasting and scenario planning camps disagree about tools, but both are often trying to spare the Decider from admitting the wager. Forecasting tries to settle the future. Scenario work tries to loosen it. Neither is sufficient if the decision itself is still resting on unnamed assumptions, which is where the science of decision making starts.

Why the forecasting and scenario planning argument is the wrong one

The UK Department for Transport comes closest to saying this in public. Its TAG Unit M4 guidance says a scenario is just a forecast produced under one set of assumptions, and asks what needs to be true for a scheme to justify itself. That is the right question, or very close to it.

Old question struck through: Which tool predicts the future better? Replaced by: What assumptions must hold for the decision to work? — Both methods dodge the question. The assumption carries the decision.

The trouble is that most organisations stop just before the useful embarrassment. They have the forecast, they have the scenarios, and still no agreed view on which assumption is carrying the decision. That stopping point suits sponsors, planners, and anyone who wants the workshop without reopening the recommendation. It produces the appearance of rigour without the inconvenience of doubt.

A handsome matrix can still leave the real assumption unnamed. I write about that in critical uncertainties scenario planning. Roger Estall and I wrote in Deciding that today's facts decay into tomorrow's assumptions. If you do not say them aloud, they still run the decision. They just do it anonymously.

Forecasts fail numerically, scenarios fail politely

The GAO's review of FAA aviation forecasting is a lovely example of respectable error. Its 2016 report found the FAA had overestimated aviation activity every year since 2004, with mean errors reaching 54.7 percent at ten years for total operations. Nobody had set error thresholds, and nobody had insisted the assumptions be made visible. The forecast looked official, so nobody asked what was holding it up.

Silicon Valley Bank managed the same trick with fresher language. The Federal Reserve's 2023 review says the bank's primary liquidity stress scenario relied on inappropriate assumptions about deposit outflows and contingency funding. Management also altered model assumptions so its interest-rate risk looked lower than it was. You can call that forecasting failure or scenario failure. I call it the old problem in a new suit: the assumption was weak, and nobody with authority forced it into the light.

That is why I do not give scenario work moral credit merely because it has several futures in it. A bad assumption survives a forecast spreadsheet and a scenario workshop with equal comfort. In one it hides behind precision. In the other it hides behind variety. The rare corporate case where scenario work earned its keep is Shell scenario planning, because it forced managers to doubt the house forecast before events did it for them.

What I use instead of either

The best evidence I have seen for scenarios changing judgement comes from Phadnis and colleagues at MIT. When long-range investment experts saw multiple scenarios, support for the least flexible option fell from 229 of 351 judgements to 176, while support for a more flexible option rose from 43 to 81. The room finally had to treat flexibility as a decision issue, not an intellectual hobby.

So I start somewhere less glamorous. I use the Universal Decision-Making Method to pin down the Purpose of the decision. Then I get the real options on the page and ask the only question I trust: what are we assuming here? Once the assumptions are visible, I judge them by influence on the outcome and confidence they will hold. That is where sufficient certainty becomes a real standard instead of another respectable phrase.

The plant executive got his answer when we wrote down the three assumptions his forecast was hiding. Two held. The third did not, and that saved him the lease.

If the critical assumption does not hold up, I either go back for information or change the decision. Sometimes the honest move is to stop there. Monitoring belongs inside the decision, not in an appendix. I keep pointing people to cases where the scenarios actually changed the decision, not just the slide deck. The useful output is the named assumption and the point at which the decision gets reopened, with an owner attached.

Use forecasts and scenarios if you must. Neither belongs in the board pack pretending to be judgement. I say more about why in the scenario planning guide.

You could approve the lease with a forecast yet never test what must be true.

Work through your decision

No sign-up. Just pick your decision and start.

Grant Purdy is the co-author, with Roger Estall, of Deciding (2020), and the architect of the Universal Decision-Making Method.