What would change the valuation answer?
Sensitivity analysis turns a model from a number-producing machine into a tool for understanding uncertainty.
A model is a set of hypotheses
Revenue growth, margins, reinvestment, cost of capital and long-term growth rates are assumptions about an uncertain future. The resulting valuation is conditional on those assumptions; it is not an independently observed fact.
Change one variable, then several
One-variable sensitivity tables can show how a discounted cash flow model reacts to a change in WACC or terminal growth. Scenario analysis then allows related variables to move together, such as weaker demand coupled with higher financing costs.
Pay attention to terminal value
In some businesses, a high proportion of modelled enterprise value depends on the period beyond the explicit forecast. This is a signal to scrutinize terminal assumptions, not a reason to accept the output as stable.
Use the model to organize debate
The best result of a sensitivity exercise is often a sharper diligence question: which assumptions are most fragile, which data could resolve uncertainty and what would genuinely change the investment thesis?
Map the biggest drivers first
Growth, margins, reinvestment, discount rates and terminal assumptions often interact. A driver register helps focus attention on inputs that are both uncertain and financially important.
Avoid impossible combinations
A scenario should have internal economic logic. A sharp revenue decline paired with rising margins and falling working-capital needs may be possible, but it requires explicit justification.
Keep the mathematics within bounds
In a perpetuity-growth DCF, the discount rate must be greater than terminal growth for the familiar formula to be well behaved. Extreme assumptions can overwhelm an otherwise reasonable forecast.
Compare changes, not just outcomes
Present the baseline, changed input, new result and absolute difference side by side. A percentage change without its starting assumptions can make a sensitivity table difficult to interpret.
Turn exceptions into diligence questions
If a small change in one assumption causes a large swing in value, the next step is not automatically to average the outcomes. Investigate the evidence, business economics and model structure behind that sensitivity.