Scenario and Sensitivity Analysis
Varying the inputs to see how much the answer moves. It converts a point estimate into a range and identifies which assumption is actually carrying the conclusion.
MadStockAlerts Research · Updated August 28, 2026
What to take away
- Sensitivity varies one input at a time; scenarios vary several coherently.
- The output is a range, which is a more honest representation.
- The most useful finding is which input the answer depends on.
- A reverse calculation solves for the assumption the price implies.
- None of it removes uncertainty; it makes the uncertainty visible.
MAD Academy Training Video · 0:45
One Number Is Never the Answer
Sensitivity analysis shows which assumption actually drives the valuation, which is more useful than the valuation itself.
This lesson is part of a Stock Alerts + Tools plan.
The two techniques
| Sensitivity | Scenario | |
|---|---|---|
| What varies | One input at a time | Several together, coherently |
| Output | A grid or a table of values | A small number of complete cases |
| Strength | Isolates which input matters | Reflects that inputs are related |
| Weakness | Treats inputs as independent, which they are not | Requires judgement about what goes with what |
The fourth row is the substantive limitation of a sensitivity grid. Growth and margin are not independent, so a cell combining high growth with unchanged margins may describe a case that cannot occur.
The finding that matters most
The primary output is not the range of values. It is the identification of which input the answer actually turns on, because that determines where research effort belongs.
Scroll the chart sideways to see all of it.
When the two tallest bars are the two least knowable inputs, that is itself the conclusion: the model's answer is a restatement of assumptions rather than a measurement, which is the criticism the DCF article makes.
Building coherent scenarios
- 1Define the cases by a storyNot by moving every input up or down, but by a coherent account of what happens.
- 2Make the inputs consistent with itA demand shock affects volume, price and margin together.
- 3Keep the number smallThree or four. More cases produce a range so wide it says nothing.
- 4Attach rough probabilitiesEven approximate weights force a view about which case is central.
The reverse construction
Rather than forecasting inputs to produce a value, take the market price and solve for the assumption that justifies it. The output is a testable statement about what is currently priced.
That is a more answerable question than what the company is worth. Whether a company can grow at the implied rate for the implied period is something evidence can bear on; what it is worth is not.
Presenting the output honestly
The analysis produces a range, and how the range is presented determines whether it informs a decision or is quietly discarded in favour of a point estimate.
- Report the range rather than a midpoint, since the midpoint restores the false precision the analysis removed.
- State which assumption drives the width, so a reader knows what the answer depends on.
- State the assumptions of each scenario explicitly rather than labelling them optimistic and pessimistic.
- Include the price at which the analysis would be wrong, which is what makes it actionable.
The first item is the discipline that is hardest to keep. A range from forty to a hundred and ten is uncomfortable, and the temptation to report seventy-five is exactly the temptation the analysis was performed to resist.
The last item connects the exercise back to the risk pillar. A valuation that produces a range is far more useful for sizing a position than one that produces a number, because the width of the range is information about how much conviction the analysis supports.