Skip to content
Log in
← Glossary

scenario analysis

Also called: scenarios, scenario planning

Building several consistent versions of the future (for example base, weak and strong), each with its own set of assumptions, and comparing the results.

Scenario analysis builds a few consistent versions of the future, each a story in which several drivers move together, and runs each through the same model. A weak year might combine fewer customers, more discounting and a weaker currency; a strong year more customers and a stronger currency. The result is a range with a reason behind each end, not one number.

It complements sensitivity analysis. Sensitivity moves one driver at a time to find which ones matter; scenarios combine the drivers that tend to move together in real life. Three scenarios (base, weak, strong) are usually enough; the base case is the most likely one, not necessarily the middle of the range.

The real value comes after the numbers: agree in advance on a trigger that says the weak story is starting (a driver crossing a level by a date) and on what the team will do when it sees it. A response decided in calm weather is cheaper and calmer than one invented in a crisis.

Example

Klasna's base forecast for next year: 2,251,968 UAH of operating profit. Two more stories run through the same monthly sheet:

Scenario
Drivers
Operating profit, UAH
Weak
learners −10%, average price 630, USD at 45
−740,802
Base
as planned
2,251,968
Strong
learners +10%, USD at 41
3,742,508

The plan for the weak case: if October's learners come in 10% under plan, halve marketing for June to August (saves 295,000) and postpone one hire (45,000 a month, 540,000 over the year). Then the weak year ends at −740,802 + 295,000 + 540,000 = +94,198 UAH: not great, but not a loss.

Common mistakes

  • Making scenarios ±10% on everything. A scenario is a story in which drivers move together for a reason.
  • Too many scenarios. Three well-built ones beat ten that nobody reads.
  • Presenting the base case as the average. It is the most likely case; the range around it is rarely symmetric.
  • Stopping at the range. Without a trigger and agreed actions, the weak case is just a scary number.
  • Building the weak case to look survivable. It should be weak enough that the plan for it is tested.

Learn it in the course