COMPARE METHODS
Sensitivity vs scenario analysis: isolate a response or compare a coherent case
Sensitivity analysis tests how an output responds to selected changes while keeping the remaining assumptions and model relationships fixed. Scenario analysis changes a coherent set of assumptions to represent a different business situation. The useful distinction is the question being answered, not which method is more sophisticated.
| Decision | Sensitivity | Scenario analysis |
|---|---|---|
| Question | What changes if this input changes? | What happens under this business situation? |
| Assumptions | Selected values or ranges; other assumptions held fixed | A coordinated set of revenue, cost, timing or financing assumptions |
| Result | Response curve, grid or tested difference | A complete case and differences from Base |
| Probability | Not assigned by the test | Not supplied by a scenario name |
On this page
The same revenue increase can produce different EBITDA
This synthetic annual example uses EUR, not EURk. Base revenue is 100 EUR, variable costs are 60% of revenue and fixed costs are 20 EUR. EBITDA = Revenue × (1 − Variable cost rate) − Fixed costs.
The sensitivity raises revenue to 110 EUR while retaining the 60% cost relationship. It does not freeze the absolute variable cost at 60 EUR: that cost rises to 66 EUR. In the scenario, higher revenue comes with a 70% variable cost rate and fixed costs of 25 EUR. Those additional assumptions change the conclusion.
| Assumption / output | Base | Revenue sensitivity | Coherent scenario |
|---|---|---|---|
| Revenue | 100 | 110 | 110 |
| Variable cost rate | 60% | 60% | 70% |
| Variable costs | 60 | 66 | 77 |
| Fixed costs | 20 | 20 | 25 |
| EBITDA | 20 | 24 | 8 |
Choose according to the decision
Use sensitivity when you need to understand a price, margin, growth or financing assumption in isolation. A two-input grid can test combinations, but the grid does not itself explain why those combinations would occur.
Use a scenario when the reasoning depends on assumptions moving together: for example, additional sales requiring a more expensive delivery mix and extra operating capacity. Document that narrative before changing the numbers. An Upside label does not make every metric favorable.
Often both are useful. First compare credible business cases; then test the assumptions that matter within each case. Keep the same output, period, currency and scenario source when comparing results.
What the comparison does not establish
Neither the 24 EUR result nor the 8 EUR result has an assigned likelihood. Neither proves a causal relationship outside the model. A finely spaced sensitivity grid can still omit a structural change, while a detailed scenario can still depend on unsupported assumptions.
If the question is a frequency under stated input distributions, read Monte Carlo vs sensitivity. It requires additional probability assumptions rather than a new label for the same cases.
Reproduce the comparison and continue with the right workflow
Download the synthetic comparison model. The saved scenario “Higher Revenue / Higher Cost” contains the coordinated changes; the Base model contains the original assumptions. The separate price sensitivity changes price from 10 EUR to 11 EUR with volume fixed at 10.
For practical case creation, continue to financial-model scenario analysis. The Sensitivity reference and scenario-modeling reference explain the respective controls. The Analyze overview helps choose a different Layer.
Sources and conventions
Synthetic examples. Model-based results were reconciled with the current BalanceCheat calculations; assumptions, units and boundaries are specified in each example.