Sensitivity Analysis in Business Valuation: Testing Key Assumptions

Every valuation is built on assumptions that could reasonably have been set somewhat differently. A slightly higher growth rate, a slightly lower discount rate, a different comparable set. Sensitivity analysis is the discipline of testing exactly how much the final valuation actually moves when those assumptions shift within a genuinely plausible range, rather than presenting a single point estimate as if it were the only defensible answer.

Why a Single Number Is Never the Whole Story

A DCF valuation that ends with a number can make people feel too certain even though the basic assumptions. Growth rate, margin path, discount rate, terminal value. Are all truly uncertain. Sensitivity analysis makes that uncertainty visible rather than hiding it behind a confident-looking single output.

This isn't a sign of a weaker valuation, it's the opposite. A valuation showing how the result changes across a reasonable range of assumptions, and explaining why the base case sits where it does, is far more defensible than one simply stating a number without acknowledging the judgment behind it.

Which Variables Actually Deserve This Treatment

Not every assumption, in a valuation model needs the level of sensitivity testing. The effort should focus on the inputs that truly shift the outcome the most.

The discount rate is almost always near the top of this list, given how directly it compounds across every year of projected cash flow and disproportionately affects terminal value. A half-point shift in WACC can move an entire valuation by a meaningful margin.

The terminal growth rate deserves equally close attention, since terminal value typically represents the majority of a DCF's total figure, and the underlying formula is mathematically most sensitive right where growth approaches the discount rate.

Near-term revenue growth and margin assumptions matter considerably for earlier-stage or fast-changing businesses, where a modest change compounds meaningfully across a multi-year projection.

Volatility, in any option-pricing or Monte Carlo-based valuation, deserves the same scrutiny, given how much a private company's estimated volatility depends on the comparable peer set chosen.

Structuring the Analysis

One-variable sensitivity tables show how the valuation changes as a single input moves across a defined range, holding everything else constant — the simplest way to isolate one assumption's specific effect.

Two-variable data tables, commonly built around discount rate and terminal growth rate together, show how the valuation shifts across combinations of the two most influential inputs, since these two interact rather than moving independently.

Tornado charts rank every tested variable by the magnitude of its effect, making it immediately visible which assumptions matter most and which are comparatively inconsequential.

Why This Isn't Just Good Practice — It's Often Required

Under both ASC 820 and Ind AS 113, Level 3 fair value measurements specifically require a quantitative sensitivity analysis showing the effect of reasonably possible changes in significant unobservable inputs. This is not an extra, for a valuation report. For these measurements this is a disclosure requirement. A missing disclosure requirement is an audit finding.

Common Mistakes in Applying Sensitivity Analysis

Testing only favorable directions. A table showing only upside scenarios, without an equally rigorous downside range, misrepresents the genuine uncertainty rather than illuminating it.

Ignoring interaction effects between variables. Discount rate and growth rate, or revenue growth and margin, often move together in a coherent scenario rather than independently testing them in isolation can understate the true range of plausible outcomes.

Treating the base case as self-evidently correct. Sensitivity analysis works best paired with a clear explanation of why the base case was chosen from within the tested range, not just a table of alternatives with no judgment attached.

Sensitivity analysis does not hurt the trustworthiness of a valuation. A good one makes it stronger. It shows clearly how much the outcome relies on known assumptions. It does not give one number and say that is the final answer. Focusing this analysis on the variables that genuinely move the outcome, testing both directions honestly, and explaining the reasoning behind the base case is what separates a valuation built to withstand real scrutiny from one that simply hopes nobody asks.