P/E, PEG, P/S, EV/EBITDA: How to Read Valuation Multiples

Whether a stock is “expensive” or “cheap” is not decided by its price — a €500 stock can be a bargain, a €5 stock wildly overpriced. What matters is how much earnings, revenue, or cash flow you get for your money. That is exactly what valuation metrics measure. This guide explains the four most important ones — P/E, PEG, P/S, and EV/EBITDA — with the formula, a worked example, and above all: the situations in which each one will lead you astray.

The Principle: Price Divided by an Earnings Measure

All four metrics follow the same pattern: you divide the price (share price or enterprise value) by an earnings measure (profit, revenue, EBITDA). The result is a multiple — “you’re paying X times Y.” The multiple by itself tells you nothing; it only becomes meaningful in comparison: with the industry, with the company’s own history, and with expected growth. Skip this triple comparison and you’re reading numbers without meaning.

P/E Ratio: The Classic — and Its Cyclical Trap

Formula in words: Share price divided by earnings per share. Alternatively: market cap divided by annual net income — same result.

Worked example: A stock trades at €60 and the company earns €3 per share. P/E = 60 ÷ 3 = 20. You’re paying 20 times annual earnings; if earnings stayed constant and flowed entirely to you, you would arithmetically recoup your investment after 20 years. Read the other way around: 1 ÷ 20 = 5% “earnings yield” per year.

When it works: For profitable companies with reasonably stable earnings — consumer staples, software, insurers. Here the P/E ratio is the fastest valuation anchor and comparable across decades.

When it misleads: The classic trap is cyclicals — carmakers, chemicals, steel, semiconductors. At the peak of the cycle, earnings explode and the P/E drops to a seemingly dirt-cheap 4 to 6 — and that is often exactly when the stock is most dangerous, because the next downturn will halve or wipe out those earnings. In the downturn, the opposite happens: earnings collapse, the P/E shoots up to 40 or becomes incalculable for lack of profit — precisely when the stock may be historically cheap. Rule of thumb: for cyclicals, an optically low P/E is a warning sign, not a buy signal. Second trap: one-off items like the sale of a stake inflate a single year’s earnings and artificially depress the P/E — so always check whether the earnings are “clean.”

PEG: The P/E Ratio Relative to Growth

Formula in words: P/E ratio divided by the expected annual earnings growth in percentage points.

Worked example: A software company has a P/E of 30 and is expected to grow 25% per year. PEG = 30 ÷ 25 = 1.2. A consumer goods conglomerate with a P/E of 22 and 5% growth comes out at a PEG of 4.4. As a rough rule of thumb, a PEG around 1 is considered fair and well above that demanding — which explains why the optically higher P/E of 30 can be the more relaxed valuation here.

When it works: For exactly this case — making growth companies comparable when they always look “too expensive” on a raw P/E. The PEG answers the question: how much valuation am I paying per percentage point of growth?

When it misleads: The PEG stands or falls with the growth estimate — and that is a forecast, not a fact. Assume 25% growth for five years where only two years are realistic, and you can talk yourself into any stock. The metric also breaks down for slow growers (at 2% growth the values become absurd) and completely ignores debt and the quality of the growth: 20% growth with shrinking margins and share dilution is worth less than 12% funded from a company’s own resources.

P/S Ratio: When There Is No Profit (Yet)

Formula in words: Market cap divided by annual revenue.

Worked example: A company is worth €5 billion on the stock market and generates €2 billion in revenue. P/S = 5 ÷ 2 = 2.5. You’re paying 2.5 times annual revenue.

When it works: For unprofitable growth companies, where a P/E ratio simply doesn’t exist for lack of earnings — young tech and biotech companies, turnaround situations. Revenue is also harder to dress up than profit, which makes the P/S ratio robust against accounting cosmetics.

When it misleads: The P/S ratio ignores margins — and with them, half the business model. A software vendor with an 80% gross margin may deserve a P/S of 8, while a grocery retailer with a 2% net margin is already richly valued at a P/S of 1: one euro of revenue turns into perhaps 25 cents of profit for the former, 2 cents for the latter. P/S comparisons across industry lines are therefore worthless. And for chronically unprofitable companies: a low P/S ratio doesn’t turn a business with no path to profitability into a bargain — revenue that never becomes profit is worth little to shareholders.

EV/EBITDA: The Metric That Makes Debt Visible

Formula in words: Enterprise value (market cap plus net debt) divided by EBITDA — earnings before interest, taxes, depreciation, and amortization.

Worked example: Market cap €10 billion, net debt €4 billion → enterprise value €14 billion. With €2 billion of EBITDA, EV/EBITDA = 14 ÷ 2 = 7. The key insight: a debt-free company with the same market cap and the same EBITDA would come out at 10 ÷ 2 = 5 — and despite an identical “price tag” on the stock market, it would be significantly cheaper. The P/E ratio would largely swallow this difference.

When it works: Whenever debt plays a role — telecoms, utilities, real estate, private equity targets — and when comparing companies with different capital structures, tax rates, or depreciation policies. That is why it is the standard metric in acquisitions: the buyer takes on the debt, too.

When it misleads: EBITDA pretends that depreciation isn’t a real expense. In asset-heavy businesses — machinery, airlines, networks — it very much is: the machines eventually have to be replaced. A “cheap” EV/EBITDA of 5 can then correspond to a very expensive ratio to free cash flow. That’s why Charlie Munger polemically called EBITDA “bullshit earnings.” Rule of thumb: never read EV/EBITDA without a look at capital expenditure needs and free cash flow.

Industry Comparison and Own History: the Double Yardstick

Each of these metrics needs two axes of comparison. First, the industry: software is structurally valued higher than steel, because margins, capital intensity, and growth differ — a P/E of 25 can be normal for the software vendor and absurd for the steel producer. So always compare against direct competitors. Second, the company’s own history: if a stock trades at 30 times earnings when it has averaged 20 times over the past ten years, you’re paying a 50% premium — which is only justified if the business has measurably improved (more growth, better margins, a stronger moat, the hallmark of quality stocks). If it hasn’t, you’re only buying a raised expectation.

Why No Single Metric Is Enough

Every multiple illuminates one facet and hides others: the P/E ratio ignores debt and growth, the PEG hangs on a forecast, the P/S ratio ignores margins, EV/EBITDA ignores capital expenditure needs. Only as a bundle do they form a picture — and only in the context of business model, balance sheet, and risk does the picture become a verdict. How that assessment works step by step is shown in our guide on how to analyze stocks; if you want to move on from multiples to intrinsic value, our article on fair value via DCF is the sequel. Further metrics, from payout ratio to interest coverage, are explained in the glossary.

In practice, this means: check several metrics in parallel, mirror each one against industry and history, and keep each metric’s weaknesses in mind. That is exactly the principle behind the AI analysis from aktienanalyse.ai, by the way — it condenses more than 35 fundamental metrics into a score and a fair value range instead of relying on a single multiple. Whether you build the bundle by hand or get it as a structured summary: what matters is that you never believe a single number.

Bottom Line: Multiples Are Thermometers, Not a Diagnosis

P/E, PEG, P/S, and EV/EBITDA take the temperature of a valuation — fast, useful, comparable. But a thermometer doesn’t make a diagnosis: why a stock trades at 8 times or at 40 times earnings can only be answered by looking at the business model, the quality of growth, and the risks. Use the metrics as a filter and an alarm system — and distrust anyone who tries to sell you a stock with a single number.