First principles: a stock is a share of future cash
A share is a claim on all the cash a business will generate for its owners from now until forever, discounted back to today. Every valuation method is just a different way of estimating that number. None is "correct" — they are structured guesses, and the goal is a range, not a decimal.
Method 1 — Discounted Cash Flow (DCF)
Project the company's free cash flow for 5–10 years, pick a discount rate r (your required return, typically 8–12%), add a terminal value for everything beyond the projection, and discount it all to today. Strength: forces you to think about the actual drivers — growth, margins, reinvestment. Trap: tiny changes in growth or discount rate swing the answer wildly. A DCF is a thermometer, not a scalpel: use conservative inputs and treat the output as a ceiling-and-floor, not a target.
Method 2 — Fair multiples
Ask: what multiple of earnings (P/E, EV/EBIT, P/FCF) has this business — and comparable businesses — deserved through a full cycle? Apply that to normalised earnings (not a one-off great year). Strength: fast, anchored in market reality. Traps: comparing against a bubble period's multiples; using a calendar year's price against a different fiscal year's earnings (align the periods!); and paying an "average" multiple for a below-average business.
Method 3 — Owner earnings
Buffett's yardstick: the cash an owner could pull out each year without hurting the business. Divide by market cap and you get an owner-earnings yield you can compare directly against bonds or other stocks. Strength: cuts through accounting noise. Trap: separating maintenance capex from growth capex takes judgement.
Why you should average methods (and then demand a discount)
Each method fails differently: DCF is assumption-sensitive, multiples import market mood, owner earnings needs judgement. Averaging independent estimates cancels part of each method's bias. Then apply the most important idea in investing — the margin of safety: only act when price sits well below your value estimate (classically 25–40% below), so being somewhat wrong still doesn't hurt you.
The five mistakes that ruin valuations
- Extrapolating a peak year's earnings as "normal".
- Mismatching fiscal periods and prices.
- Ignoring share dilution (value per share is what you own).
- Forgetting net debt — equity value ≠ enterprise value.
- Trusting one method's single number instead of a range.
