Discounted Cash Flow Guide
Full DCF mechanics — projecting cash flows, terminal values, and sensitivity analysis
Valuation foundations
How to Value a Stock
Without fooling yourself — start with what the price already believes, then decide if you know something it doesn't.
Implied expectationsDCFMultiplesMargin of safetyNormalization
Stock valuation is the process of estimating what a company is actually worth — its intrinsic value — and comparing that estimate to the current market price. If your estimate of intrinsic value is materially higher than the price, you may have found an undervalued stock. If it is lower, the stock may be overpriced regardless of how good the business looks.
The three most common approaches to stock valuation are the DCF model (discounted cash flow), comparable multiples analysis (P/E, EV/EBITDA vs. peers), and earnings power value. Each method has strengths and blind spots. A DCF model forces you to project future cash flows explicitly but is sensitive to assumptions about growth and discount rates. Multiples are intuitive but inherit every mispricing in the peer group. The most reliable valuations triangulate across all three.
This guide teaches a practical, assumption-first framework for valuing any stock. You will learn how to reverse-engineer the expectations embedded in the current price, choose the right valuation method for the business type, normalize financial inputs to avoid garbage-in-garbage-out errors, and set margin of safety rules before committing capital. Along the way, you can use our free DCF calculator, P/E context tool, and intrinsic value calculator to apply each framework on live ticker data.
In this guide: Read the price · Three valuation lenses · Implied expectations workout · Normalize inputs · Decision rules · Common mistakes
The core insight
The stock already has a valuation. Your job isn't to create one from scratch — it's to find the assumption the market is getting wrong.
Most valuation guides start with a blank spreadsheet and ask you to build up a fair value from scratch. That is the wrong starting point. Every traded stock already has a price, and that price encodes a set of assumptions about growth, margins, reinvestment, and risk. Before you build a single model, your first job is to decode those assumptions. Then — and only then — you look for the one or two assumptions where you have a genuine, evidence-based disagreement with the market.
This guide teaches valuation as a process of disagree or pass, not “build a model and hope the number is bigger than the price.”
Step 1
The price is a set of assumptions. Read them first.
A stock trading at $85 per share is not an arbitrary number. It is the market's consensus on what the company's future cash flows are worth today, discounted for risk and time. Embedded in that $85 is an assumed revenue growth rate, an assumed operating margin trajectory, an assumed reinvestment rate, and an assumed cost of capital. Change any one of those assumptions and the “right” price changes with it.
This is why starting with a blank DCF is backwards. You are trying to solve for a number that already exists. The smarter move: reverse-engineer the price. Take the current stock price, the company's trailing financials, and a reasonable discount rate. Then ask: what revenue growth rate and what margin level does this price require to be justified? If the answer is 25% annual growth for a company that has never grown faster than 12%, the price is pricing in a miracle. If the answer is 6% growth for a company that has compounded at 14% for a decade, the price might be pricing in a disaster that isn't coming.
Key reframe
Valuation is not “what is this stock worth?” It is “what does the current price assume — and do I have evidence that those assumptions are wrong?”
Example: Suppose MedDevice Corp (illustrative) trades at $120 per share with 200M diluted shares, $8B in trailing revenue, 18% operating margins, and $1B in net debt. At a 10% discount rate, reverse-engineering the price reveals the market is baking in ~16% annual revenue growth at 22% steady-state margins. The company has grown revenue 9% CAGR over the past 5 years and never posted margins above 19%. The price is not “expensive” or “cheap” in the abstract — it is specifically assuming a growth acceleration and margin expansion that the historical record does not support. That is a concrete, testable disagreement, and it is where your edge starts.
Step 2
Three valuation lenses — and when each one lies to you
There is no single “correct” valuation method. There are three major lenses, each useful in specific contexts and misleading in others. The skill is knowing which lens to pick for the stock in front of you — and how to triangulate when they disagree.
| DCF | Comparable multiples | Earnings power | |
|---|---|---|---|
| Best for | Stable cash-flow businesses where you can project 5+ years with some confidence — industrials, consumer staples, mature SaaS | Any stock with a meaningful peer set — most useful in sectors where the market has priced similar companies and you need to understand relative positioning | Mature, slow-growth businesses where the current earnings run rate is roughly what the company will earn indefinitely — utilities, REITs, regulated industries |
| How it works | Project free cash flows for 5–10 years, then add a terminal value. Discount everything back at a rate that reflects the risk of those cash flows. | Divide price by a fundamental (earnings, EBITDA, revenue, book value). Compare the ratio against peers and the stock's own history. | Take current normalized earnings, divide by your required return. The result is the capitalized value of the earnings stream as a perpetuity. |
| Where it lies | Terminal value dominates — often 60–80% of the total. Small changes in terminal growth or exit multiple swing the output by 30%+. Gives false precision on genuinely unknowable inputs. | Inherits every mispricing in the peer group. If the entire sector is overvalued, comps will tell you the stock is 'cheap relative to peers' while it's expensive in absolute terms. | Assumes current earnings persist forever. Misleads badly for any company with a growth or decline trajectory. Ignores reinvestment needs. |
| Failure mode | Analyst builds a model that works at exactly one discount rate, declares the stock 40% undervalued, and never runs a sensitivity table. | Analyst sees 8x P/E, calls it cheap, doesn't notice earnings are at a cyclical peak and the company is a terminal declining business. | Analyst capitalizes peak earnings from a commodity upcycle and declares a mining stock worth 3x its current price. |
Practical rule
Use DCF for your independent estimate. Use comps as a sanity check. Use earnings power as the floor. If all three disagree violently, you probably don't understand the business well enough to invest.
The choice of method should follow from the type of business, not your personal preference. High-growth companies with negative current cash flow are almost impossible to value on earnings power — you need a DCF that captures the trajectory. Stable utilities with predictable rate bases are natural earnings-power candidates. Consumer discretionary companies with 10 publicly traded peers are tailor-made for relative valuation. Using the wrong lens for the business type is how smart people build elegant models that produce garbage outputs.
Run valuation models yourself
Apply DCF and multiples analysis on any ticker — no spreadsheet required.
Step 3
The implied expectations workout
Theory is worth nothing until you run the numbers. Here is a concrete worked example, followed by an interactive calculator so you can run the same analysis on any stock.
Worked example: IndustrialCo (illustrative)
Given: Stock price $85 · 450M shares diluted · $12B trailing revenue · 15% operating margin · $2B net debt · 10% discount rate.
Step 1 — Enterprise value: Market cap = $85 × 450M = $38.25B. Add net debt: EV = $40.25B.
Step 2 — Reverse-engineer: Using a simplified DCF (22% tax rate, 5% capex/revenue, 3% terminal growth), we solve for the 5-year revenue CAGR that produces an enterprise value of $40.25B. The answer: approximately 14–15% annual growth at the current 15% margin, or ~11% growth if margins expand to 18%.
Step 3 — Compare to base rates: IndustrialCo's sector (diversified industrials) has grown revenue at 5–7% CAGR over the last decade. The company itself has done 9%. The price implies growth 50–60% faster than the company's own track record and 2–3× the sector average.
Verdict: Unless you have a specific, defensible reason to believe growth will accelerate dramatically — a new product cycle, a regulatory tailwind, a market share shift you can quantify — the price is embedding an expectation the base rates don't support. You either need to find the catalyst, or pass.
Implied expectations calculator
The price implies 34.7% annual revenue growth for 5 years at 15.0% steady-state margins. If you believe growth is closer to 34.7% at 15.0% margins, the stock is roughly fairly priced by approximately 0.0%. Your assumptions imply a fair value of $85.00 vs. the current price of $85.00.
Simplified DCF model. Assumes 22% tax rate, 5% capex-to-revenue, 3% terminal growth. All figures are illustrative — not investment advice. Adjust the discount rate input above to test sensitivity.
The calculator above does the same math from the worked example, but lets you plug in any stock. Enter the basics, see what the market is pricing in, then adjust the sliders to reflect your own view. The gap between “what the price implies” and “what you believe” is your thesis. If the gap is small, there is no trade. If the gap is large and you have evidence, you might have an edge.
Step 4
Normalize before you model
Garbage in, garbage out. Every valuation model — DCF, comps, earnings power — is only as good as the inputs you feed it. And the inputs that most people feed it are wrong, because they use raw trailing numbers without adjusting for distortions. Before you project anything forward, normalize.
The naive inputs produce a stock that looks like it's growing 18% with 22% margins and $3.60 in FCF per share. The normalized inputs reveal an 11% grower with 19% margins and $2.77 in FCF per share. Same company. Same year. Wildly different valuation outputs.
The four normalizations that matter most:
- Cyclicality: Use mid-cycle revenue and margins, not peak or trough. For commodity producers, model at a 10-year average commodity price, not today's spot.
- One-time items: Strip asset sales, restructuring charges, legal settlements, and gain/loss on investments. If a company has a “one-time charge” every quarter, it's not one-time — it's the business.
- Stock-based compensation: SBC is a real cost. Subtract it from free cash flow. Use fully diluted shares in every per-share calculation. Companies that report “adjusted earnings” excluding SBC are flattering their numbers by ignoring the cost of their own compensation program.
- Capital allocation quality: A company reinvesting at 25% ROIC deserves a higher multiple than one reinvesting at 8% ROIC, even if current earnings are identical. Factor in the quality of incremental returns when setting your growth assumptions.
Step 5
Set your decision rules before you see the answer
The most dangerous moment in valuation is when you have a number and a stock you already like. Confirmation bias will find a way to make the model say “buy.” The countermeasure: set your decision rules before you run the final model. Write down — in advance — what margin of safety you require, what position size you'll take at different price levels, and what specific facts would make you sell.
Margin of safety is not optional. Every intrinsic value estimate is wrong. The discount rate is a guess. The terminal growth assumption is a guess. The margin trajectory is a guess. The margin of safety is the cushion that absorbs all of those errors simultaneously. For a high-quality business with predictable cash flows (Visa, Costco, Microsoft), a 20–25% margin of safety may be sufficient — the range of outcomes is narrow. For a cyclical, capital-intensive, or competitively threatened business, demand 35–50%. If the margin of safety isn't there, the answer is not to lower your standards. It's to wait.
Position sizing follows from conviction, not excitement. A high-conviction idea with a wide margin of safety and a verifiable catalyst warrants 5–8% of a portfolio. An interesting idea with a narrower margin of safety or less clear catalyst is a 2–3% position. An “it might work out” idea is a watchlist entry, not a position. Sizing discipline is what separates investors who survive drawdowns from investors who are forced to sell at the bottom.
Anti-anchoring discipline
If you wouldn't buy the stock at today's price without knowing what you paid last time, you shouldn't hold it either.
Your cost basis is irrelevant to the stock's current value. The question is always: “Would I buy this stock today at this price with this information?” If the answer is no, you should be selling, regardless of whether you're up 40% or down 40%. Pre-commit to this rule before you enter any position.
Pre-commitment checklist
Check every box before committing capital. If you can't, you either need more work or the trade isn't ready.
- I have a written estimate of intrinsic value with explicit assumptions
- I can name the specific assumption the market is getting wrong
- I have defined my margin of safety and it exceeds 20% for quality businesses, 35%+ for cyclicals
- I have a stop-loss trigger — the fact or event that would invalidate my thesis entirely
- I have written down the price at which I would add to the position
- I have written down the price at which I would trim or sell completely
- My position size is proportional to my conviction — and survivable if I'm 100% wrong
- I have checked that this position doesn't create unintended sector concentration in my portfolio
Failure modes
The five fastest ways to get valuation wrong
These are not theoretical risks. They are the specific errors that destroy returns for investors who otherwise do solid fundamental work. Every one of them feels reasonable in the moment. That's what makes them dangerous.
1. Anchoring on a low P/E
A stock trading at 8x earnings is not automatically cheap. It might be 8x peak earnings in a cyclical business, 8x earnings inflated by a one-time asset sale, or 8x earnings that are about to fall 40%. The P/E tells you the price relative to last year's accounting profits. It tells you nothing about whether those profits are sustainable, growing, or about to collapse. Every value trap in history looked cheap on trailing P/E.
2. Using management guidance as your base case
Management teams are paid to be optimistic. They set guidance to a number they're reasonably confident they can beat, then beat it by 2% so the stock pops on the "earnings surprise." If you model management's long-range revenue targets as your base case, you are building a bull case and calling it neutral. Use management guidance as the upper bound of your scenario range, not the midpoint. Your base case should come from industry base rates and your own analysis of the competitive landscape.
3. Ignoring dilution
A company reports $2B in net income. You divide by 500M shares and get $4 EPS. But the company issued 25M shares in stock-based comp this year, 15M in a secondary offering, and has 40M in dilutive options outstanding. The real share count is growing 3–4% a year, meaning your per-share value is shrinking even as total earnings grow. Always use fully diluted shares. And subtract stock-based comp from free cash flow — it is a real cost paid by existing shareholders through dilution.
4. Confusing peak earnings for normal earnings
Commodity producers, homebuilders, semiconductor companies, and shipping firms all have earnings that spike during upcycles and crater during downturns. If you value a copper miner on its earnings when copper is $4.50/lb, you will massively overpay for the stock. The right approach is to normalize — average earnings across a full cycle, or model the company at a mid-cycle commodity price. Ask: what does this company earn in an average year, not what does it earn in the best year?
5. Terminal multiple you can't defend
In a DCF, the terminal value often represents 60–80% of total value. If you slap a 20x exit multiple on year-5 earnings because "the sector trades at 20x," you've just anchored 70% of your valuation to a single assumption you pulled from today's market environment. Today's multiples reflect today's interest rates, today's growth expectations, and today's risk appetite — none of which you can predict five years out. Use a perpetuity growth model as a cross-check, and run your terminal at 2–3 different multiples to see how much it matters.
Free tools
Apply these valuation frameworks on any stock.
DCF CalculatorBuild a discounted cash flow model with live ticker data. Get intrinsic value, margin of safety, and full sensitivity analysis.P/E Context ToolPut any P/E ratio in historical and sector context. See how the market has priced this stock across earnings cycles.Intrinsic Value CalculatorCompare DCF, dividend discount, and earnings power valuations side by side for any stock.Earnings Quality ScoreVerify that the earnings you're valuing are real — cash conversion, accruals, GAAP vs. adjusted spread.
Common questions
Stock valuation — answered directly.
What is intrinsic value?
Intrinsic value is an estimate of what a business is actually worth based on its future cash flows, growth rate, and risk profile — independent of the current stock price. Investors compare intrinsic value to market price to identify stocks that may be undervalued or overvalued. You can estimate it with our free intrinsic value calculator.
What is a DCF model?
A DCF (discounted cash flow) model projects a company's future free cash flows over 5–10 years, adds a terminal value, and discounts everything back to today using a rate that reflects the risk of those cash flows. It is the most widely used method for estimating the intrinsic value of a stock. Try it yourself with our DCF calculator.
How do you calculate intrinsic value?
The most common approach is to build a DCF model: forecast free cash flows, estimate a terminal value, and discount both at an appropriate rate (typically 8–12% for equities). You can cross-check with comparable multiples (P/E, EV/EBITDA) and an earnings power value to triangulate a reasonable range. Our intrinsic value calculator runs all three methods on live data.
What discount rate should I use for stock valuation?
Most analysts use 8–12% for established, large-cap companies, and higher rates (12–15%+) for small-caps or volatile businesses. The exact number matters less than running a sensitivity table — a 1-point change can swing your estimated value by 15–20%, so test your thesis across multiple rates.
DCF vs comparable analysis — which is better?
Neither is reliable alone. A DCF forces you to state assumptions explicitly but gives false precision on unknowable inputs. Comparable analysis is grounded in real market prices but inherits every mispricing in the peer group. The best practice is to use both: DCF for your independent estimate and P/E comps as a sanity check.
Go deeper
Each valuation lens has its own dedicated guide.
This page gives you the framework. The guides below go deep on specific methods — DCF mechanics, multiples interpretation, cash flow analysis, and earnings quality signals.
How Stock Multiples Work
P/E, EV/EBITDA, and why cheap-looking stocks aren't always cheap
Free Cash Flow Yield Guide
Measuring what you actually get for what you pay
Earnings Quality Checklist
Spot accounting tricks before they blow up your valuation model
How To Analyze A Stock
The full research process from filing to decision
P/E Ratio by Industry
Benchmark any stock's earnings multiple against sector medians and historical ranges
Ready to value a specific stock?
Plug any ticker into our DCF Calculator to reverse-engineer the market's implied assumptions, estimate intrinsic value, and run sensitivity analysis — no spreadsheet required.
Apply it
Value any stock in minutes.
Basis Report generates a complete fundamental analysis — financial overview, earnings quality, valuation, and risk factors — on any ticker. Start with the implied expectations, then decide if the market is wrong.