P/E is price ÷ earnings — and both halves move.
- 22× means paying 22 years of today’s profit
- Falling profit RAISES the P/E — “cheap” can be a trap
- P/E cannot see debt
What this is about
The price-to-earnings ratio is the most quoted number in investing and the most casually misused. It is genuinely useful — and almost every way people use it in conversation is one of the traps below.
Worked example continues with Northwind Components, fictional.
What it is
Share price £34.00 Earnings per share £1.525 (61m profit ÷ 40m shares) P/E 22.3×
Two readings, both correct:
"You are paying 22.3 times one year's earnings."
"At this profit level, it would take 22.3 years of earnings to repay the price." — the more sobering phrasing, and the more useful one.
Flip it and you get the earnings yield: 1 ÷ 22.3 = 4.5%. That number is directly comparable to a bond yield or a savings rate, which is what makes it worth calculating.
Trap 1: comparing across industries
Northwind Components (industrial) 22.3× Cascade Grocers (supermarket) 14.1× Aurora Systems (software) 46.8×
Aurora is not "expensive" and Cascade is not "cheap". Different industries carry different ratios for structural reasons: growth rates, how much capital they must reinvest, how predictable earnings are, how much debt is normal.
A P/E is only informative against a comparable. Against the same company's own history, and against direct competitors. Ranking unrelated businesses by P/E produces a list, not an insight.
Trap 2: the denominator is one year
E is a single year's earnings, and a year can be unrepresentative.
A bad year inflates it. Suppose Northwind has a poor 2026 — profit falls to 20m on a one-off write-down. Price unchanged:
P/E = £1,360m ÷ £20m = 68×
The company now looks wildly expensive on a number that says nothing about its normal earning power.
A peak year deflates it. A cyclical business at the top of its cycle shows a low P/E precisely when its earnings are least sustainable — which is the single most expensive P/E mistake there is, because the number looks most attractive exactly when it is least reliable.
This is the origin of the phrase value trap: a low P/E that reflects earnings about to fall, not a bargain.
Trap 3: which E
You will see several, and they are not interchangeable:
- Trailing — the last twelve months. Factual, and backward-looking.
- Forward — analysts' estimate for next year. Useful, and an estimate, and estimates are systematically optimistic on average.
- Adjusted — the company's own preferred figure, excluding items it considers one-off. Sometimes reasonable, sometimes a way of excluding costs that recur every year under different names.
A stated P/E without saying which E is an incomplete number. Trailing and forward on the same company routinely differ by a third.
Trap 4: it ignores the balance sheet
Two companies, identical earnings and identical P/E. One has no debt; the other has debt equal to half its market value. You are not buying the same thing, and P/E cannot see the difference. Debt belongs in the judgement separately — it is step 5 of lesson 1 for exactly this reason.
Where it is genuinely useful
Against the company's own history. Northwind at 22× when it has averaged 15× over five years is a real observation: you are paying more than buyers usually have. Whether that is justified is the question, and now it is the right question.
As an earnings yield, against alternatives. 4.5% against what a government bond pays is a comparison with actual content.
As a statement of what must happen. A high P/E is the market pricing in growth. Writing down what growth it implies turns a vague "expensive" into something checkable.
A worked example
Use the ratio the one way it is informative: against comparables. Northwind at 22.3× means little alone. Against its own history — roughly 18×, 19×, 21×, 22× over the previous four years — it says you are paying the top of the company's own range, for a business whose growth has not accelerated. Flip it: a 4.5% earnings yield against, say, a 4% risk-free rate leaves half a percentage point as the reward for all the extra risk. Neither number says sell or avoid. Both state precisely what you are being asked to pay, which is the ratio's actual job.
Now watch one trap fire in real time. The 2026 write-down cuts profit to 20m and the P/E jumps to 68× with the business unchanged. A screener set to "P/E below 25" — a perfectly common filter — drops Northwind in exactly the year the number stopped meaning anything. The same trap works in reverse on a cyclical at peak earnings: 6× on a profit about to halve is not cheap, it is a denominator about to disappear. A P/E is only ever as good as the one question people skip: is this year's E representative?
What this does not tell you
It does not tell you whether a company is good. Excellent businesses trade at high multiples and bad ones at low. P/E measures price against one year of profit, and nothing about quality.
It does not tell you a low P/E is cheap. Often it is the market's judgement that earnings will fall. Sometimes the market is wrong; assuming so without a reason is not analysis.
It does not work on companies with no earnings. Negative E makes the ratio meaningless, not infinite — which is why loss-making growth companies get judged on other measures.
It does not stay comparable through a share issue or buyback. Both move EPS without the business changing.
Where to see this in the app
Chart → Key stats carries the P/E alongside market cap, so step 6 of lesson 1 is one screen.
Chart → Financials is where you check the denominator before trusting the ratio — a glance at four years of net income tells you whether this year's E is representative or an outlier, which is trap 2 defused in about fifteen seconds.
Watchlists is the honest way to use P/E: put three companies from the same industry side by side. That is the only comparison the number supports.
Chart → Analyst rating shows published expectations, which is where a forward P/E's E comes from — worth knowing before relying on one.