How to value a stock: a practical, evidence-first process

Stock valuation is not about finding one perfect number. It is a disciplined process for translating business performance and explicit assumptions into a range of values, then testing how fragile those values are.

1. Start with the business, not the formula

A valuation is a set of assumptions about a business expressed in numbers. Before forecasting anything, understand what the company sells, how it earns money, which segments drive revenue and profit, how cyclical demand is, what capital the business must reinvest, how strong its competitive position is, and which risks could permanently change its economics. A precise spreadsheet cannot rescue a weak understanding of the business.

For a mature subscription software company, recurring revenue, retention, margins, and reinvestment may dominate the analysis. For a semiconductor manufacturer, cycle conditions, utilization, capital expenditures, technology leadership, and customer concentration may matter more. Banks, insurers, REITs, commodity producers, and early-stage businesses often require different frameworks from a standard industrial FCFF model.

2. Build a normalized financial base

Start with reported revenue, operating income, taxes, depreciation and amortization, capital expenditures, working-capital changes, cash, debt, diluted shares, and free cash flow. Then decide which figures are representative of ongoing operations. One-time restructuring charges, unusually high or low margins, acquisitions, asset sales, temporary tax items, or a cyclical peak can distort a single period.

3. Choose a valuation method that matches the business

No single valuation method is universally correct. Discounted cash flow is useful when operating cash generation can be modeled with reasonable assumptions. Earnings multiples are useful when earnings are meaningful and comparable companies exist. Free-cash-flow yield can provide a fast cash-based cross-check. Asset-based methods can be more relevant for businesses whose balance-sheet assets are central to value. A sum-of-the-parts approach may be useful when a company contains economically different businesses.

See the deeper stock valuation methods comparison for method selection.

4. Forecast cash flow explicitly

In an enterprise FCFF framework, a common starting relationship is:

Forecast revenue, operating margins, taxes, depreciation, reinvestment, and working capital in a way that is internally consistent. Growth usually requires investment. A model that assumes rapid revenue growth while capital spending and working-capital needs disappear may be mathematically valid but economically implausible.

Use an explicit forecast period long enough for unusual growth or margins to move toward a more sustainable state. Five years is common, but the correct horizon depends on the business and the purpose of the model.

5. Estimate a defensible discount rate

FCFF is commonly discounted using weighted average cost of capital (WACC). Conceptually, WACC blends the required returns of equity and debt according to the company's financing structure. The discount rate should reflect the risk of the cash flows being valued; it is not a knob to turn until the model produces a preferred answer.

A higher discount rate lowers present value. A lower rate raises it. Because the effect compounds over time, small changes can move a DCF materially, especially when much of the value lies in distant cash flows.

For a step-by-step explanation of the mechanics, read Discounted Cash Flow (DCF) Valuation.

6. Handle terminal value carefully

Terminal value estimates the value of cash flows beyond the explicit forecast. In a perpetual-growth approach, the terminal growth rate should be consistent with a mature business operating in a real economy. If a model assumes a company grows faster than the economy indefinitely, the mathematics can produce an impressive number that is not economically credible.

Watch the percentage of total enterprise value coming from terminal value. A very high terminal-value share does not automatically invalidate a model, but it tells you that the result depends heavily on distant assumptions. That deserves lower confidence and stronger sensitivity analysis.

7. Cross-check with earnings and comparable companies

A DCF should not live in isolation. Compare implied valuation with normalized earnings, free cash flow, historical valuation ranges, and genuinely comparable businesses. Peer selection should be based on economics—not simply sector labels. Growth, margins, capital intensity, balance-sheet risk, geography, product mix, and business quality can justify different multiples.

When DCF and earnings-based estimates disagree sharply, do not average them mechanically. Investigate why they disagree. One method may be reacting to heavy reinvestment, unusual free cash flow, an aggressive terminal assumption, an unsustainable earnings period, or a peer multiple that is itself expensive.

8. Use reverse valuation to test the market price

Instead of asking only “what is the stock worth?”, reverse the problem: “what future performance would justify today's price?” Solve for the revenue growth, operating margin, free-cash-flow conversion, or terminal assumptions implied by the market price. This reframes valuation as an expectations test.

If the current price requires assumptions you believe are unrealistic, the valuation concern becomes clearer. If the required assumptions look modest relative to the company's history and competitive position, the market price may be easier to defend.

9. Stress-test the assumptions

Do not rely on one base case. Test combinations of revenue growth, margins, WACC, terminal growth, capital spending, share dilution, and cyclicality. Sensitivity tables are especially useful because they show how quickly the conclusion changes when two major inputs move together.

10. Document what would change the conclusion

A useful valuation is falsifiable. Write down the facts that would strengthen or weaken the thesis: revenue growth above or below a threshold, a durable margin change, a major change in capital intensity, new debt, dilution, customer loss, regulatory change, competitive displacement, or a different interest-rate environment. When new information arrives, update the model instead of defending the old answer.

How to tell whether the result is actually useful

Ask four questions: Are the source data current and traceable? Are the assumptions economically plausible? Do different valuation methods tell a coherent story or materially disagree? How sensitive is the result to a small change in major inputs? Confidence should fall when source quality is weak, assumptions are fragile, or methods disagree substantially.

For a practical screening process after you calculate value, see How to Tell if a Stock Is Undervalued.

How Simple AI Stock Valuation applies this framework

The app's sample reports place DCF and earnings/peer valuation beside financial analysis, technical context, risks, catalysts, source references, confidence labels, and validation results. The point is not to produce a single unquestionable number; it is to make the inputs, disagreements, and limitations visible enough for the user to inspect.

See the report methodology and historical sample reports for concrete examples of how the framework is presented.

Apply the framework to a stock or ETF

Simple AI Stock Valuation can generate a point-in-time research PDF that puts valuation beside fundamentals, technical context, risks, sources, assumptions, confidence, validation, and disclosures.

Educational only: This material is general information, not personalized financial, investment, tax, or legal advice. Valuation estimates are uncertain and can be wrong.