Guide

Discounted cash flow (DCF) valuation explained

A discounted cash flow (DCF) model estimates what a business is worth today by projecting the cash it will generate in the future and translating those future dollars into present value. The core idea is time value of money: $100 received five years from now is worth less than $100 today because you could invest today's dollars and earn a return. DCF is the formal version of that intuition — sum the present value of expected free cash flows (FCF), add a terminal value for cash flows beyond your forecast horizon, subtract net debt, and divide by shares outstanding to get intrinsic value per share. Unlike a P/E ratio, which compares price to last year's accounting earnings, DCF asks what the business can earn in cash over its life. This guide walks through building a DCF from scratch, choosing discount rates (WACC), handling terminal value responsibly, running sensitivity analysis, pairing DCF with fundamental analysis, and the mistakes that turn spreadsheets into false precision.

Why DCF exists and what it answers

Stock prices reflect what buyers will pay right now; intrinsic value reflects what a business is worth to an owner who cares about cash, not quotes. DCF bridges that gap by answering: "If I owned 100% of this company and collected its surplus cash for decades, what would those flows be worth discounted at my required return?" The output is a dollar estimate — often a range, not a single point — that you compare to the current market price. A stock trading at $80 with an intrinsic value estimate of $110 suggests margin of safety; $80 against $65 suggests overpayment relative to your assumptions.

DCF shines for businesses with predictable cash generation: mature software with high recurring revenue, industrial companies with stable capex, consumer brands with pricing power. It struggles with pre-profit startups, commodity cyclicals at peak earnings, and banks where "cash flow" is not defined the same way. For those cases, value investors lean on asset-based metrics, normalized earnings multiples, or sum-of-the-parts models instead of a single FCF stream.

Free cash flow: the input that matters

DCF models discount free cash flow — cash the business generates after operating expenses and capital investments needed to maintain or grow the asset base. The most common definition for equity holders is unlevered free cash flow (UFCF), also called free cash flow to the firm (FCFF):

UFCF = EBIT × (1 - tax rate) + D&A - Capex - Change in NWC

Start with operating profit (EBIT), tax it at the marginal rate to approximate cash taxes, add back non-cash depreciation and amortization (D&A), subtract capital expenditures (capex) required to sustain operations, and subtract increases in net working capital (NWC) — the cash tied up in inventory and receivables minus payables. Read financial statements carefully: capex in the cash flow statement is factual; maintenance vs growth capex requires judgment. Aggressive DCF models understate capex and overstate FCF.

An alternative path starts from net income and walks through the cash flow statement to FCF to equity, then discounts at the cost of equity instead of WACC. Both approaches work if you stay consistent — never mix levered cash flows with an unlevered discount rate.

Projecting cash flows: the forecast period

Most models use an explicit forecast of five to ten years. Year one often anchors on the latest trailing FCF or management guidance, then applies assumptions for revenue growth, operating margin, capex as a percent of sales, and working capital intensity. Growth should decay toward mature-economy rates (2-4% nominal) unless you have a durable competitive advantage that justifies sustained outperformance — and even then, competition and law of large numbers cap how long hypergrowth lasts.

Document every assumption. "Revenue grows 12% for five years because TAM" is not analysis; tie growth to unit economics, market share limits, and historical reinvestment returns. A DCF with ten adjustable knobs and no sensitivity table is a storytelling device, not a valuation.

Discount rates: WACC and the cost of equity

Future cash flows are discounted back using a rate that reflects the risk of receiving them. For unlevered FCF, use the weighted average cost of capital (WACC):

WACC = (E/V) × Re + (D/V) × Rd × (1 - tax rate)

E is market value of equity, D is market value of debt, V = E + D, Re is cost of equity, Rd is cost of debt. Cost of equity is often estimated with the Capital Asset Pricing Model (CAPM): Re = risk-free rate + beta × equity risk premium. Higher beta (more volatile, cyclical businesses) demands a higher discount rate and compresses present value.

WACC links directly to interest rates: when the risk-free rate (typically 10-year Treasury yields) rises, WACC rises and intrinsic values fall even if cash flow forecasts are unchanged. That is why growth stocks with distant cash flows reprice sharply in rate-hiking cycles — most of their value sits in terminal value years away.

Typical WACC ranges: 7-9% for stable large caps, 9-12% for mid-cap operating companies, 12%+ for high-risk or emerging-market assets. Using 8% because "it feels right" without tying to current yields and beta is a common error. When in doubt, run the model at WACC minus 1%, base, and plus 1% and report the range.

Terminal value: where most DCF value hides

Businesses do not stop after year ten. Terminal value (TV) captures all cash flows from year eleven onward. Two standard methods:

Gordon growth (perpetuity) model: TV = FCF_n × (1 + g) / (WACC - g), where g is the perpetual growth rate. g must be less than WACC and should not exceed long-run GDP growth (roughly 2-3% real) unless you enjoy proving you were wrong in court. A 3% perpetual growth rate on a company already dominating a mature market is usually optimistic.
Exit multiple method: apply an EV/EBITDA or EV/FCF multiple to year-ten metrics based on comparable mature peers. This cross-checks the perpetuity assumption against what the market pays for similar businesses today.

Terminal value often represents 60-80% of total enterprise value in a ten-year model. That concentration means small changes in g or WACC swing intrinsic value wildly — which is why professional analysts publish sensitivity matrices, not single numbers.

Discount terminal value back to present: PV(TV) = TV / (1 + WACC)ⁿ. Sum PV of explicit FCFs plus PV(TV) to get enterprise value (EV). Subtract net debt (debt minus cash), add non-operating assets, divide by diluted shares for equity value per share.

From enterprise value to buy, hold, or pass

Compare intrinsic value per share to the current stock price. Value investors often demand a margin of safety — buying only at a 20-30% discount to intrinsic value to absorb modeling error and unforeseen competition. If your base case says fair value is $100 and the stock trades at $98, the margin is thin; one wrong terminal growth assumption flips the verdict.

Triangulate DCF with relative multiples. If your DCF implies 35x forward earnings but peers trade at 18x with similar growth, ask which assumption diverges — your growth forecast, your margin path, or your discount rate. Disagreement with the market is fine; unexplained disagreement is not.

Sensitivity analysis table

Build a two-way table: rows = WACC (e.g., 8.0% to 10.0%), columns = terminal growth (e.g., 2.0% to 3.5%). Each cell is intrinsic value per share. Investors should see a range — "$85 to $125, base $102" — rather than a false-precision "$101.47." If the stock trades at $70 and only the most pessimistic cells exceed $70, the risk-reward may be attractive; if only optimistic cells justify the price, caution is warranted.

DCF vs shortcut valuation methods

Method	Strengths	Weaknesses	Best for
DCF	Fundamental, captures growth and capex, works across sectors with cash flow	Assumption-heavy, terminal value dominates, false precision risk	Mature companies with visible FCF, acquisition pricing, sum-of-parts
P/E, EV/EBITDA multiples	Fast, market-anchored, good for relative comparison	Ignores balance sheet, cyclical distortion, growth not explicit	Quick screens, peer sets, cyclicals on normalized earnings
Dividend discount model	Simple when payouts are stable	Useless for non-dividend growth stocks and buyback-heavy firms	Utilities, REITs, mature dividend aristocrats
Asset-based (book, liquidation)	Floors value for asset-heavy businesses	Ignores earning power and intangibles	Banks, holding companies, distressed situations

Use DCF when you need an absolute anchor; use multiples for relative context. The best workflows run both and reconcile differences before committing capital.

Common DCF mistakes

Overstating near-term growth — extrapolating pandemic-era or acquisition-boosted growth rates without mean reversion.
Understating capex — treating all capex as "growth" and ignoring maintenance spend required to keep revenue flat.
Aggressive terminal growth — 4%+ perpetual growth on a large-cap implies the company eventually exceeds global GDP; math fights you.
Discount rate too low — anchoring WACC to historical averages when risk-free rates have structurally risen.
Ignoring dilution — using basic share count when stock-based compensation and convertibles expand the denominator.
Double-counting synergies — in M&A models, baking best-case cost saves into base case without probability weighting.
Single-point outputs — publishing one intrinsic value without sensitivity to WACC, growth, and margin assumptions.
DCF on cyclical peak earnings — discounting record FCF at the top of a commodity cycle produces "cheap" stocks that are expensive on normalized cash flow.

Worked intuition (simplified)

Imagine a business generating $50M UFCF this year. You forecast 8% growth for five years, then terminal growth of 2.5% with WACC of 9%. Explicit-period FCF grows to roughly $73M by year five. Discount each year at 9%, sum to about $280M present value. Terminal value at year five: $73M × 1.025 / (0.09 - 0.025) ≈ $1.15B; discount back five years at 9% ≈ $750M. Enterprise value ≈ $1.03B. Subtract $100M net debt → $930M equity. At 50M shares, intrinsic value ≈ $18.60/share. Change WACC to 10% or terminal growth to 2.0% and the answer moves several dollars — that swing is the model telling you uncertainty dominates precision.

Production checklist

FCF definition documented — UFCF vs FCFE matches discount rate.
Forecast assumptions tied to revenue drivers, not arbitrary growth cells.
Capex and NWC modeled as ratios of sales, stress-tested in downturn.
WACC built from current risk-free rate, beta, and market risk premium.
Terminal growth ≤ long-run GDP; cross-checked with exit multiples.
Sensitivity table on WACC and terminal growth (and margin if thin).
Net debt and dilution reflected in per-share equity value.
Triangulation with P/E, EV/EBITDA, and peer multiples before decision.
Margin of safety threshold defined before comparing to market price.

Key takeaways

DCF values a business by discounting projected free cash flows and terminal value to present dollars.
Free cash flow — not accounting earnings — is the economic input; capex and working capital matter as much as revenue growth.
WACC translates risk into a discount rate; rising interest rates lower intrinsic values for long-duration cash flows.
Terminal value dominates most models — treat perpetual growth and exit multiples conservatively.
Publish ranges via sensitivity analysis, not false-precision single prices; demand margin of safety before buying.