Sensitivity Tables and Tornado Charts

Any valuation model rests on assumptions — a growth rate, a discount rate, a margin forecast — that are estimates, not certainties. Presenting a single output number without showing how sensitive it is to those assumptions can be misleading, since a small change in an uncertain input might swing the valuation dramatically. A sensitivity table addresses this directly by recalculating the output across a grid of different input values, typically two assumptions at once (say, revenue growth down one axis and discount rate across the other), so a reader can see the full range of plausible answers rather than one point estimate.

A tornado chart answers a related but different question: of all the uncertain inputs in a model, which ones actually move the output the most? Each input is varied one at a time across its plausible range while holding everything else fixed, and the resulting swing in output is plotted as a horizontal bar. Sorting these bars from largest swing to smallest produces a shape that tapers like a tornado, immediately showing which two or three assumptions deserve the most scrutiny and which barely matter.

Worked example. A discounted-cash-flow model's base case values a company at $50 per share. A tornado chart that varies terminal growth rate, discount rate, and near-term margin one at a time shows the discount rate swings the valuation from $38 to $68 (a $30 range), terminal growth swings it $20, and near-term margin swings it only $4. This tells the analyst to spend their diligence time stress-testing the discount rate assumption, since it dominates the uncertainty in the final number far more than the margin assumption does.