Basis Report/Resources/Investor Foundations
Investor Foundations6 sections18 entries

DCF Sensitivity Analysis: How Assumptions Change Your Valuation

A DCF model is only as good as its assumptions — and small changes in growth, discount rate, or terminal value can move fair value by 30% or more. Sensitivity analysis is how serious investors turn a point estimate into a decision range.

✓After running a DCF, immediately test what happens when growth drops by 2 percentage points and discount rate rises by 1 point.
✓Calculate what percentage of your total DCF value comes from the terminal period — if it's above 65%, your model is a bet on perpetuity assumptions.
✓Build a 5x5 sensitivity table with growth rate on one axis and discount rate on the other — the range of outcomes is your real fair value zone.
✓Define three scenarios with genuinely different assumption sets, not just base ± 10%.
Go to frameworkSee a sample report

Why Every DCF Needs a Sensitivity Analysis

A discounted cash flow model produces a single number — an intrinsic value estimate. That number feels precise. It sits in a cell with two decimal places. It looks like an answer. But it isn't. It's the output of a chain of assumptions, and if any of those assumptions shift by even a small amount, the answer changes dramatically. Sensitivity analysis is how you expose that fragility before the market does it for you.

Here's the core problem: a 1 percentage point change in your revenue growth assumption can move fair value by 15–20%. A 1 percentage point change in your discount rate can move it by 20–30%. And the terminal growth rate — the assumption buried at the end of the model that most investors barely think about — often drives 60–70% of the total value. These aren't edge cases. They're the normal range of reasonable disagreement among informed analysts looking at the same company.

Sensitivity analysis doesn't make the uncertainty go away. It makes it visible. Instead of defending a single fair value estimate, you map the range of outcomes across different assumption sets. You build a sensitivity table that shows exactly how growth and discount rate interact. You define bull, base, and bear scenarios with explicit assumptions you can monitor quarter by quarter. And you probability-weight those scenarios to arrive at a fair value range that reflects what you actually know — and what you don't.

This guide walks through the complete sensitivity analysis process: which inputs matter most, how to build a sensitivity table, how to define scenarios that aren't just your base case ± 10%, and how to convert the output into a position-sizing decision. If you've ever built a DCF and wondered whether the answer was real or just the product of optimistic inputs, this framework will give you a better way to think about it.

When to use this

After building any DCF model, before sizing a position, and whenever a stock re-rates sharply and you need to know whether the new price is justified under reasonable alternative assumptions.

Why it matters now

In a market where narrative can justify almost any growth assumption, the investors who stress-test their models have a structural advantage over those who anchor on a single optimistic scenario.

Where theses break

The process breaks when investors run sensitivity tables but still anchor on the center cell, when scenarios are just the base case with cosmetic adjustments, or when the exercise becomes a ritual rather than a genuine test of conviction.

Full framework

6 sections · 18 entries — work through each before you size a position.

The illusion of precision is the most expensive mistake in DCF modeling. Sensitivity analysis replaces false confidence with calibrated ranges that improve position sizing and risk management.

18 entries in view

Why sensitivity analysis matters

A DCF model without sensitivity analysis is a guess disguised as precision. Small input changes create enormous valuation swings, and the illusion of a single 'right' answer leads to overconfident positioning.

Small input changes create huge valuation swings

A 1 percentage point increase in your revenue growth assumption can move fair value by 15–20%. A 1 point change in discount rate can shift it by 20–30%. These aren't extreme scenarios — they represent the normal range of disagreement among competent analysts. The math of discounting amplifies uncertainty at every step, which means a DCF model is inherently more sensitive to assumptions than most investors realize.

Why it matters

If you don't quantify how sensitive your valuation is to each input, you don't actually know what you're betting on. You might think you're making a revenue growth bet when you're actually making a discount rate bet.

When it matters

Every time you finish a DCF model and before you use the output for any position-sizing decision.

Investor take

Before trusting any fair value number, ask: how much does this change if my two most important assumptions are each wrong by 1 percentage point?

Anchoring bias makes single-point DCFs dangerous

Once you calculate a fair value of, say, $142, your brain anchors on that number. Every subsequent analysis gets filtered through that anchor — you'll unconsciously seek confirming evidence and dismiss contradicting data. Sensitivity analysis breaks the anchor by forcing you to see $142 as one point in a range that might span $105 to $195 depending on assumptions.

Why it matters

Anchoring is the most common cognitive bias in valuation work, and it's the hardest to see in yourself. The only reliable defense is a process that makes the range of outcomes visible before you commit to a view.

When it matters

When you find yourself defending a specific price target rather than discussing a valuation range. When the market disagrees with your estimate and you feel certain rather than curious.

Investor take

Never present a DCF result as a single number. Always present the range from your sensitivity table and state which assumptions drive the width of that range.

The illusion of precision in financial modeling

A DCF model can have 200 rows, quarterly forecasts out to year 10, and margin assumptions to the basis point — and still be worthless if the three assumptions that matter most are wrong. Model complexity often creates a false sense of confidence: the more detailed the spreadsheet, the more 'real' the output feels. But precision in the model doesn't equal accuracy in the forecast.

Why it matters

The market doesn't reward model complexity. It rewards assumption quality. A simple 3-variable sensitivity analysis on a clean DCF framework often produces better investment decisions than a 50-tab model that was never stress-tested.

When it matters

When you're tempted to add more detail to your model instead of testing whether the core assumptions survive scrutiny.

Investor take

Spend 80% of your DCF time on the 3 assumptions that drive 80% of the value, not on perfecting the 47 line items that move the output by less than 5% combined.

The three inputs that move the needle

Not all DCF inputs are created equal. Revenue growth rate, discount rate (WACC), and terminal growth rate are the three assumptions that drive the vast majority of valuation outcomes.

Revenue growth rate: the assumption everyone has an opinion on

Revenue growth compounds through the entire forecast period and into the terminal value. A company growing at 12% versus 10% annually doesn't just produce 2% more revenue each year — over a 10-year forecast, the cumulative difference is roughly 25% more total revenue, which flows through margins and into free cash flow at every step. This is why a seemingly small growth disagreement can move fair value by 15–20%.

Why it matters

Growth is the most debated assumption in any DCF because it's the most visible and the most subject to narrative. But it's also the assumption where overconfidence costs the most, because the compounding effect amplifies every error forward through time.

When it matters

When building your initial DCF and when updating after each quarterly earnings release. Always compare your growth assumption to the implied growth rate embedded in the current stock price.

Investor take

Test at least five growth rate scenarios spanning the range of reasonable outcomes — don't just use consensus, consensus +2%, and consensus -2%. Include a stagnation case.

Discount rate (WACC): the assumption that moves value most per basis point

The discount rate determines how much each future dollar of cash flow is worth today. At a 9% WACC, $1 of free cash flow received in year 10 is worth $0.42 today. At 11% WACC, it's worth $0.35 — a 17% decline in present value from just a 2 percentage point rate change. This sensitivity increases with duration: long-duration growth stocks are dramatically more rate-sensitive than near-term cash flow stories. Use our WACC calculator at /tools/wacc-calculator to estimate the rate for any specific company.

Why it matters

Discount rate is where the most valuation manipulation happens — consciously or not. An analyst who needs a higher fair value can shave 50 basis points off WACC and 'find' 10–15% more value without changing a single business assumption. Sensitivity analysis makes this trick visible.

When it matters

Whenever building or reviewing a DCF. Pay special attention when interest rates are moving, when the company's risk profile changes, or when you're comparing valuations across companies with different WACCs.

Investor take

Never use a single discount rate. Always run your DCF at WACC, WACC +1%, and WACC -1% at minimum. If the thesis only works at the low end of the rate range, the margin of safety is thinner than it appears.

Terminal growth rate: the most dangerous assumption in the model

The terminal growth rate determines the value of all cash flows beyond your explicit forecast period — typically representing 60–75% of total DCF value. A company modeled with a 3% terminal growth rate is worth dramatically more than the same company at 2% terminal growth, because the terminal value formula (FCF × (1 + g) / (WACC - g)) is hyperbolic: as the growth rate approaches the discount rate, terminal value approaches infinity.

Why it matters

Terminal growth rate is dangerous because it's the most impactful assumption and the hardest to defend. No company grows forever, and long-term growth rates above GDP growth (2–3% nominal) require an argument about durable competitive advantage that most businesses can't support. Yet many models use 3–4% terminal growth without questioning it.

When it matters

During model construction and during every sensitivity check. Always calculate how much total DCF value comes from the terminal period and treat that percentage as your model's duration risk.

Investor take

Default to 2–2.5% terminal growth for average businesses. Only use 3%+ if you can articulate a specific, durable competitive advantage that justifies above-GDP perpetuity growth. Then stress-test at 1.5% to see if the stock still works.

Building a sensitivity table

A sensitivity table is the core tool of DCF stress-testing. It maps how fair value changes across combinations of two key inputs, replacing a point estimate with a visible range of outcomes.

How to construct a two-variable sensitivity matrix

Choose two inputs that drive the most uncertainty in your model — typically revenue growth rate on one axis and discount rate (WACC) on the other. Create a grid with 5 values per axis: your base case in the center, two steps above, and two steps below. For growth, use 1–2 percentage point increments. For WACC, use 0.5–1 percentage point increments. Calculate fair value at every intersection. The result is a 5×5 grid showing 25 possible fair values.

Why it matters

A two-variable table captures the interaction between assumptions — something you miss when you adjust one variable at a time. Growth and discount rate aren't independent: a higher-risk business (higher WACC) that also grows faster doesn't simply net out — the interaction determines whether the stock is attractive or not.

When it matters

After completing any DCF model and before using the output for position sizing. Update the table after each earnings release when growth or risk assumptions change.

Investor take

Build the table in your spreadsheet so it updates automatically when you change base case assumptions. The table should be the first thing you look at, not the point estimate in the summary cell.

Reading the sensitivity table: finding your fair value range

Don't look for a single number in the sensitivity table — look for the zone where most realistic assumption combinations cluster. If 15 of 25 cells show a fair value between $120 and $160, that's your working range. Then compare that range to the current stock price. If the stock trades at $100 and almost all cells show values above $120, the margin of safety is visible across scenarios. If the stock trades at $140 and half the cells show values below that, the margin of safety is narrow.

Why it matters

The density of the table tells you how much the valuation depends on getting assumptions right. A tight cluster (all cells between $130–$150) means the stock's value is robust to input uncertainty. A wide spread ($80–$250) means you're making a high-conviction bet on specific assumptions — position size accordingly.

When it matters

Every time you update the sensitivity table. Compare the range to the current price and to your position size as a percentage of portfolio.

Investor take

Color-code your sensitivity table: green where fair value exceeds current price by your required margin of safety, yellow for borderline, red for overvalued. The pattern of colors tells you the probability-weighted attractiveness at a glance.

When to add a third variable

A two-variable table captures most of the important sensitivity. But sometimes a third variable — operating margin, capex intensity, or terminal growth rate — is genuinely uncertain and impactful enough to warrant its own analysis. In that case, build a separate two-variable table for each level of the third variable. For example, three 5×5 tables at terminal growth rates of 2%, 2.5%, and 3% give you 75 scenarios across three variables.

Why it matters

Adding a third variable is worth the effort only when that variable is truly uncertain and moves value materially. If terminal growth moves fair value by ±3% across the reasonable range, it doesn't need its own table. If it moves value by ±25%, it does.

When it matters

When you've built a standard 2-variable table and one remaining assumption still feels like a significant source of uncertainty.

Investor take

Keep the total number of scenarios manageable — 75 (three 5×5 tables) is practical. More than that usually means you haven't identified which assumptions actually matter.

Scenario analysis: bull, base, and bear

Sensitivity tables test individual variables. Scenario analysis tests complete narratives — internally consistent sets of assumptions that represent different versions of the future.

How to define scenarios with different assumption sets

Each scenario should tell a coherent story, not just vary one number. A bull case isn't just 'higher growth' — it's 'higher growth because market share gains accelerate, pricing power holds, and reinvestment needs decline as the platform matures.' Each assumption should connect to a specific business outcome. Define 3–5 key variables for each scenario and explain why they move together in that direction.

Why it matters

Internally consistent scenarios force you to think about the business rather than the model. When you have to explain why margins expand while growth accelerates, you're testing whether the story makes economic sense — not just whether the math produces a higher number.

When it matters

After building sensitivity tables and before making a position-sizing decision. Each scenario should map to a distinct thesis about the company's competitive trajectory.

Investor take

Write one paragraph for each scenario explaining the business narrative before filling in the numbers. If you can't articulate why the assumptions move together, the scenario is just arithmetic, not analysis.

Probability-weighting outcomes

Assign a probability to each scenario based on your assessment of likelihood — for example, 25% bull, 50% base, 25% bear. Multiply each scenario's fair value by its probability weight and sum the results to get a probability-weighted expected value. This single number incorporates both your fundamental analysis and your confidence in each outcome.

Why it matters

Probability weighting forces honest calibration. If your bull case produces $200 fair value and you weight it at 50%, you're implicitly saying there's a coin-flip chance of the best outcome. Most analysts who are forced to assign explicit probabilities realize their base case should carry more weight and their extreme cases less than they initially assumed.

When it matters

After defining all scenarios and before comparing fair value to market price. Revisit weights after major catalysts — a strong earnings print might shift weight from bear to base, not just change the base case numbers.

Investor take

Track your scenario weights over time. If you find yourself consistently increasing the bull case weight to justify the position, you may be rationalizing rather than analyzing.

When to use scenarios versus sensitivity tables

Sensitivity tables are better for testing how robust a thesis is to input uncertainty — they answer 'does this stock work across a range of assumptions?' Scenario analysis is better for testing narrative uncertainty — it answers 'what are the 2–3 distinct futures this business could face, and what is each one worth?' Use both. The sensitivity table gives you the mechanical range. The scenarios give you the strategic range.

Why it matters

The two tools complement each other because they test different types of uncertainty. A sensitivity table might show the stock works at every combination of growth and WACC, but a scenario analysis might reveal that the bear case — a new competitor disrupts the business model — produces a fundamentally different cost structure that the sensitivity table doesn't capture.

When it matters

Always use a sensitivity table. Add scenario analysis when the company faces discrete strategic uncertainties — new product launches, regulatory decisions, competitive threats — that can't be captured by sliding a continuous variable up and down.

Investor take

Present your investment conclusion using both frameworks: the sensitivity table shows the valuation is robust to input error, and the scenario analysis shows you've considered the strategic range of outcomes.

Common mistakes in DCF sensitivity analysis

Even investors who run sensitivity tables often make systematic errors that undermine the value of the exercise. These are the most common mistakes and how to avoid them.

Using management guidance as the base case

Management guidance is a negotiation with Wall Street, not a forecast. Companies set guidance to be beatable — conservative enough to beat but ambitious enough to maintain credibility. When you use the midpoint of guidance as your DCF base case, you're anchoring on management's marketing, not on your own analysis. Your base case should reflect your independent view of normalized growth and margins.

Why it matters

Analysts who anchor on guidance tend to cluster around the same fair value estimates, creating crowded positioning. Independent base cases — built from your own unit economics analysis — are where variant perception comes from.

When it matters

Every time you build or update a DCF. Compare your base case to guidance and to consensus, but make sure your assumptions were built before you checked those references, not after.

Investor take

Build your base case assumption by assumption before looking at guidance or consensus. Then compare. If your number is far from consensus, ask whether you have a genuine insight or a genuine mistake.

Ignoring mean reversion in growth and margins

High-growth companies tend to slow down. High-margin companies tend to face competitive pressure. These aren't bearish assumptions — they're base rates. Yet many DCF models project current growth and margins forward with minimal fade, producing terminal values that assume a company will be one of the rare businesses that sustains exceptional economics indefinitely.

Why it matters

Mean reversion is the strongest base rate in corporate finance. Only about 10% of companies sustain above-average profitability for more than a decade. Models that don't incorporate fade are implicitly betting that this company belongs in that 10% — and that bet should be explicit, not hidden.

When it matters

When building long-term forecast assumptions. Apply fade rates to growth and margins unless you have specific evidence of a durable competitive advantage (switching costs, network effects, regulatory moats) that justifies sustained outperformance.

Investor take

Default to building in margin and growth fade over your forecast period. If someone asks why, the answer is base rates. If you want to model sustained outperformance, make that the bull case, not the base case.

Letting terminal value dominate without testing it

In most DCF models, terminal value represents 60–75% of the total equity value. This means the majority of your valuation depends on assumptions about what happens after your explicit forecast period ends — the period where you have the least visibility. Many investors check their 5-year revenue and margin forecasts carefully but never stress-test the terminal growth rate that drives most of the answer.

Why it matters

A DCF where terminal value is 75% of total value is really a bet on perpetuity assumptions, not a bet on the next 5 years of cash flows. Knowing this doesn't make the model wrong — it makes the position-sizing conversation honest.

When it matters

Every time you complete a DCF. Calculate terminal value as a percentage of total equity value and include it in your investment memo.

Investor take

If terminal value exceeds 70% of total DCF value, run the model with terminal growth rates from 1% to 3.5% and see how wide the fair value range becomes. That range is the real uncertainty in your valuation.

Putting it into practice

Sensitivity analysis is a process, not a one-time exercise. This section walks through the step-by-step workflow for turning a DCF point estimate into an actionable fair value range.

Step 1: Run your base DCF and identify key assumptions

Build or update your DCF model with base case assumptions. Then identify the 2–3 inputs that drive the most value — typically revenue growth, WACC, and terminal growth rate. You can find these by changing each input by ±1 percentage point and noting which changes move fair value the most. Rank them by impact. These are the variables that will go into your sensitivity table.

Why it matters

This step is about triage. A DCF model might have 30 inputs, but usually 2–3 drive 80% of the valuation outcome. If you stress-test the wrong variables, the sensitivity analysis looks thorough but misses the actual risk.

When it matters

At the start of every new DCF and after any major model revision. Also after earnings releases that change the growth or margin trajectory.

Investor take

Write down the three assumptions that move value most and the dollar impact of a 1-point change in each. Keep this on the first page of your model — it's the most important output you'll produce.

Step 2: Build sensitivity tables and identify the fair value range

Construct a 5×5 sensitivity table using the two highest-impact variables. Calculate fair value at each intersection. Identify the cluster of scenarios you consider most likely and define that as your working fair value range. A stock is attractive when the current price sits below the low end of the likely range, and expensive when it sits above the high end.

Why it matters

The sensitivity table transforms your DCF from a point estimate into a range with visible assumptions. This changes the investment conversation from 'the stock is worth $142' to 'the stock is worth $118–$165 depending on growth and discount rate, and the current price implies...' — which is a much better starting point for decision-making.

When it matters

After identifying key assumptions and before defining scenarios. Update the table whenever base case inputs change.

Investor take

Share the sensitivity table, not the point estimate, when discussing the investment with others. It creates a more productive conversation because people can debate assumptions rather than conclusions.

Step 3: Define scenarios, assign probabilities, and size the position

Create bull, base, and bear scenarios with internally consistent assumption sets. Assign probability weights. Calculate the probability-weighted expected fair value. Compare to the current stock price and calculate the expected margin of safety. Use this margin of safety — not the base case upside — to inform position sizing. A wide margin of safety with high-probability scenarios supports a larger position; a narrow margin with a wide scenario range calls for smaller sizing.

Why it matters

This is where sensitivity analysis becomes a portfolio management tool, not just a valuation exercise. Position sizing should reflect your confidence-adjusted view of value, not your most optimistic scenario.

When it matters

Before initiating or changing a position. Revisit after each earnings cycle or major catalyst.

Investor take

If the probability-weighted expected value is less than 15% above the current price, the margin of safety may not justify a full position — even if your bull case is compelling. Let the probabilities do the work, not the best-case scenario.

Related resources

Continue your research

DCF Calculator

Run your own DCF model free

WACC Calculator

Estimate your discount rate for any stock

DCF Analysis Guide

Step-by-step guide to building a DCF model

How to Value a Stock

Complete stock valuation framework

Common questions

What investors ask about investor foundations for investor foundations stocks.

What is sensitivity analysis in DCF?
Sensitivity analysis in a DCF model tests how much the estimated intrinsic value changes when you adjust key assumptions — typically revenue growth rate, discount rate (WACC), operating margin, and terminal growth rate. The goal is to map the range of reasonable fair values rather than relying on a single point estimate. A standard approach is building a two-variable sensitivity table that shows fair value at every combination of, say, five growth rates and five discount rates. This reveals which assumptions matter most and how much uncertainty exists in the valuation.
How do you do a sensitivity analysis for a discounted cash flow model?
Start by identifying the 2–3 assumptions that move your valuation most — usually revenue growth, WACC, and terminal growth rate. Then build a sensitivity table: put one variable on each axis and calculate fair value at every intersection. A typical table uses 5 values per variable (your base case plus two above and two below in realistic increments). Next, define 3 complete scenarios (bull, base, bear) with internally consistent assumption sets. Finally, assign probability weights to each scenario and calculate a probability-weighted fair value. Compare that weighted value to the current stock price to gauge margin of safety.
What discount rate should I use in a DCF?
Most equity analysts use the weighted average cost of capital (WACC), which typically falls between 8% and 12% for established U.S. companies. The exact rate depends on the company's cost of equity (driven by beta, equity risk premium, and risk-free rate), cost of debt, and capital structure. Small-cap stocks, emerging market companies, and businesses with volatile cash flows warrant higher rates — often 12–15%. The critical insight is that a 1–2 percentage point change in discount rate can move your fair value by 20–30%, so you should always run a sensitivity table rather than treating any single rate as correct. Use our WACC calculator to estimate it for specific tickers.
Why does terminal value dominate DCF results?
Terminal value captures all cash flows beyond your explicit forecast period (typically years 6–10 and beyond), discounted back to today. For most companies, terminal value represents 60–75% of total DCF output because the explicit forecast period is short relative to the company's remaining life. This dominance means your terminal growth rate and exit multiple assumptions — which are the hardest to forecast accurately — have an outsized impact on the final number. The best defense is to stress-test terminal assumptions aggressively: vary the terminal growth rate by ±1% and see how much fair value moves. If your investment thesis depends on a specific terminal assumption, acknowledge that and size the position accordingly.
How many scenarios should a DCF model have?
Three scenarios (bull, base, bear) is the standard framework and sufficient for most investment decisions. The key is making each scenario internally consistent — a bull case should have higher growth, higher margins, and potentially a lower discount rate, all supported by a specific narrative about what goes right. Some analysts add a fourth 'catastrophe' scenario for tail risk. More than four scenarios usually adds complexity without improving decision quality. The value comes not from the number of scenarios but from the discipline of defining explicit assumptions for each one and assigning honest probability weights.

Continue research

Keep building the foundation

DCF Analysis Guide for Investors — How to Build a DCF ModelInvestor FoundationsHow to Value a StockInvestor FoundationsHow stock multiples actually workInvestor FoundationsEBITDA for investors: useful bridge, dangerous destinationInvestor FoundationsReturn on invested capital: how to use ROIC without kidding yourselfInvestor FoundationsTech multiples: what actually deserves a premiumInvestor Foundations

Put it on a live stock

MSFT stock overviewJump to the live ticker page and report flowAAPL stock overviewJump to the live ticker page and report flowGOOGL stock overviewJump to the live ticker page and report flowAMZN stock overviewJump to the live ticker page and report flow