Business 200Lesson 11 of 1514 min

Sensitivity and Scenario Analysis — From Point Estimate to Valuation Range

A single DCF output is not a valuation — it is an artifact of specific assumptions. McKinsey insists that every valuation must include sensitivity analysis, scenario analysis, and ideally Monte Carlo simulation to understand how the value conclusion responds to key assumption changes. The goal is not to produce a number but to understand the shape of value: what drives it, what kills it, and where the critical uncertainty lies.

What you'll learn
  • Build one-way and two-way sensitivity tables on key DCF assumptions and interpret the output correctly
  • Distinguish scenario analysis from sensitivity analysis — and when each approach is appropriate
  • Design a scenario framework with internally consistent assumption sets rather than independent variable changes
  • Apply a Monte Carlo framework conceptually — understand what probability distribution assumptions produce and how to interpret a valuation range
  • Read a tornado chart and identify which assumptions contribute most to valuation uncertainty

Sensitivity Analysis — One-Way and Two-Way Tables

Tornado Chart — EV Range by Assumption (Base = $1,695M)
Each bar shows EV range from assumption low to high. Widest bars = highest valuation sensitivity.
Terminal Growth Rate
$1155M$2.5B
WACC
$1155M$2.5B
Terminal NOPAT Margin
$910M$2.3B
Revenue CAGR (Yrs 1–5)
$1340M$2.0B
Capex / Revenue
$1560M$1.8B
Tax Rate
$1620M$1.8B
NWC / Revenue
$1640M$1.8B
EV range
Base EV ($1.7B)
Two-Way Sensitivity — EV ($B) vs. WACC × Terminal Growth
WACC \ g1.5%2%2.5%3%3.5%
9.0%$1.84B$1.97B$2.13B$2.33B$2.59B
9.5%$1.63B$1.74B$1.86B$2.01B$2.21B
10.0%$1.46B$1.55B$1.7B$1.76B$1.91B
10.5%$1.32B$1.39B$1.47B$1.56B$1.68B
11.0%$1.19B$1.26B$1.33B$1.4B$1.49B
Scenario Framework (Internally Consistent)
Bear Case$780M
25% probability — Competition, margin erosion, stalled growth
Base Case$1,695M
50% probability — Moderate expansion, stable unit economics
Bull Case$3,850M
25% probability — Market leadership, platform emergence
Probability-weighted EV: $2,004M0.25×$780 + 0.50×$1,695 + 0.25×$3,850

Sensitivity analysis asks: 'If I change one assumption while holding all others constant, how does enterprise value change?' It isolates the marginal impact of each driver. One-way sensitivity changes a single variable across a range; two-way sensitivity varies two variables simultaneously, producing a matrix of EV outcomes. The critical inputs for a DCF sensitivity: (1) terminal growth rate; (2) WACC; (3) the NOPAT margin in year 5 (or terminal year); (4) revenue growth in the explicit period. These four variables capture virtually all material valuation uncertainty.

WACC \ Growth1.5%2.0%2.5%3.0%3.5%4.0%
8.0%$2,420$2,640$2,910$3,250$3,720$4,410
8.5%$2,100$2,270$2,480$2,740$3,090$3,600
9.0%$1,840$1,970$2,130$2,330$2,590$2,960
9.5%$1,630$1,740$1,860$2,010$2,210$2,490
10.0%$1,460$1,550$1,645$1,760$1,905$2,110
10.5%$1,315$1,390$1,470$1,560$1,675$1,830
11.0%$1,190$1,255$1,325$1,400$1,490$1,615

The table above shows EV varying from $1,190M to $4,410M — a 3.7× range from the most bearish to the most bullish WACC/growth assumptions. This is not model noise; it is genuine uncertainty about long-run assumptions. The analyst's job is to identify which WACC and growth rates are defensible given the company's competitive position and industry dynamics — not to simply report the range. A common mistake: presenting the full table and implying all outcomes are equally probable. The table should be accompanied by a clear view of where the base case sits and why the analyst believes that specific WACC and growth rate are most likely.

Terminal NOPAT Margin8%9%10%11%12%13%14%15%
Enterprise Value ($M)$1,050$1,185$1,320$1,510$1,695$1,880$2,065$2,250
% Change from Base−38%−30%−22%−11%+11%+22%+33%

Scenario Analysis — Internally Consistent Assumption Sets

Sensitivity analysis changes one variable at a time, which can produce internally inconsistent combinations — e.g., simultaneously assuming high revenue growth (bullish) and a high WACC (bearish market). Scenario analysis avoids this by defining complete, coherent narratives where all assumptions are consistent with each other. A bear case is not just low growth — it includes lower margins (scale doesn't materialize), higher capex intensity (competitive investment required), and potentially higher WACC (if the bear case includes sector stress).

AssumptionBear CaseBase CaseBull Case
Revenue CAGR (5-year)3%8%14%
Terminal NOPAT Margin8%12%16%
Capex / Revenue (mature)9%7%5%
NWC / Revenue18%14%11%
Terminal Growth Rate1.5%2.5%3.5%
WACC11.5%10.0%9.0%
Implied ROIC (terminal)9%14%22%
Enterprise Value$780M$1,695M$3,850M
Equity Value per Share (73M diluted)$4.80$17.31$47.80
NarrativeCompetition intensifies, pricing erodes, growth stallsBase assumptions hold, moderate expansion, stable marginsMarket share gains, scale benefits materialize, platform business emerges

Some analysts assign probabilities to scenarios (e.g., bear 25%, base 50%, bull 25%) and compute a probability-weighted intrinsic value: 0.25×$4.80 + 0.50×$17.31 + 0.25×$47.80 = $21.81/share. This is a useful framework for communicating the expected value, but the probabilities themselves are subjective — the precision of the arithmetic should not obscure the uncertainty in the scenario definitions. McKinsey cautions against treating the probability-weighted value as a single true estimate; it is a decision-support tool, not a precise prediction.

Tornado Chart — Ranking Assumptions by Valuation Impact

The tornado chart is a structured visualization of one-way sensitivity analysis, ranking assumptions from most to least impactful on enterprise value. By showing the EV range from each assumption's low and high bound side by side, it reveals which variables drive the most valuation uncertainty. The widest bars (most impactful assumptions) deserve the most analytical attention; the narrowest bars can be treated as given and require no further investigation.

AssumptionLow BoundEV at LowEV at HighHigh BoundTotal EV Range
Terminal Growth Rate1.0%$1,155M$2,465M4.0%$1,310M
WACC8.0%$2,465M$1,155M12.0%$1,310M
Terminal NOPAT Margin7%$910M$2,250M16%$1,340M
Revenue CAGR (Yrs 1-5)2%$1,340M$2,050M15%$710M
Capex / Revenue5%$1,820M$1,560M10%$260M
Tax Rate20%$1,765M$1,620M30%$145M
NWC / Revenue10%$1,750M$1,640M18%$110M

Instead of point estimates at the low and high bounds, Monte Carlo simulation treats each assumption as a probability distribution (e.g., terminal growth = Normal(2.5%, 0.8%); WACC = Uniform(8.5%, 11.5%)). Running 10,000 simulations produces a distribution of enterprise values, from which analysts compute: expected EV, percentile ranges (10th–90th), probability of EV exceeding or falling below key thresholds (e.g., P(EV > $2B) = 34%). In practice, Monte Carlo is most valuable when assumptions are genuinely uncertain and correlated — for example, in resources and mining, oil price uncertainty correlates with both revenue and WACC. For most corporate DCFs, the three-scenario framework captures the essential uncertainty at lower analytical cost.

Reverse DCF — What Is the Market Implying?

The reverse DCF inverts the valuation question: instead of asking 'what is this company worth given my assumptions?', it asks 'what assumptions must be true for the current market price to be justified?' This is one of the most powerful tools in the applied valuation toolkit because it tells the analyst exactly what the market is pricing in — and whether those implied assumptions are realistic.

  1. Start with the current market capitalization and add net debt to compute implied enterprise value. This is the market's 'answer' to the valuation question.
  2. Set WACC at your best estimate — this is not what you are solving for. WACC is an external input based on capital market data, not a function of growth expectations.
  3. Choose a terminal growth rate assumption consistent with the industry's long-run potential (typically GDP growth ± 1%).
  4. Solve for the revenue growth rate and/or terminal margin that, when plugged into your model structure (capex, NWC, and ROIC assumptions held constant), exactly produces the implied enterprise value.
  5. Compare the implied growth/margin to: (a) historical growth rates; (b) analyst consensus; (c) industry averages; (d) physical capacity constraints. If implied growth = 20% per year for 5 years and the industry has never grown faster than 8%, the market is pricing in an assumption that requires extraordinary evidence to support.

For growth stocks trading at high multiples, the reverse DCF is far more informative than asking 'is the P/E too high?' The reverse DCF tells you exactly what growth rate and margin profile are required — and you can assess whether those are achievable. Damodaran's approach: compute the implied revenue (or FCFF) growth rate at which the DCF equals the market price; then apply a probability to that growth rate being achieved; if (probability × EV at implied growth) + (1−probability) × EV at realistic growth < market price, the stock is overvalued. This is how systematic investors quantify the risk priced into a stock, not just whether it is 'cheap' or 'expensive' on a multiple.

Key Takeaways

  • Two-way sensitivity tables (WACC × terminal growth) reveal the range of DCF outcomes — but the analyst must anchor on defensible assumptions, not simply present the table as all-equally-possible outcomes
  • Scenario analysis produces internally consistent assumption bundles where all inputs (growth, margins, capex, WACC) reflect a coherent narrative — unlike sensitivity analysis which changes one variable at a time
  • Tornado charts rank assumptions by their EV impact range, identifying which variables deserve the most analytical attention; in most DCFs, terminal growth rate and WACC produce the widest ranges
  • Monte Carlo simulation treats assumptions as distributions and produces a probability distribution of enterprise values; it is most valuable for correlated, highly uncertain assumptions (commodities, early-stage companies)
  • The reverse DCF inverts the question: what assumptions justify the current market price? It is the most direct test of whether market pricing is grounded in realistic expectations

Quiz — 3 Questions

Answer one at a time
Question 1 of 30 answered

A two-way sensitivity table shows EV ranging from $800M to $3,200M when WACC varies from 8%–12% and terminal growth varies from 1%–4%. Your base case is WACC=10%, g=2.5%, giving EV=$1,695M. The company is trading at EV=$2,100M. What can you conclude?

AThe stock is overvalued — EV > base case DCF
BThe market is pricing assumptions above your base case. The $2,100M market EV is achievable in the sensitivity table (e.g., WACC=9.5%, g=3.0% or WACC=9.0%, g=2.5%). The question is whether those market-implied assumptions are more or less realistic than your base case. If your WACC and growth assumptions are well-founded, then the market is pricing the company at a premium to your intrinsic value. But if your WACC is too high (e.g., you used book value weights) or your growth is too conservative, the market price could be correct. The correct conclusion requires: (1) verifying your WACC methodology; (2) running a reverse DCF to identify exactly what WACC/growth combination produces $2,100M; (3) assessing whether those implied assumptions are defensible.
CDCF valuation is unreliable — the range is too wide
DThe market is wrong because EV > DCF base case