Valuation multiples.
You see them every day. P/E, EV/EBITDA, price to book, forward this and trailing that. They get quoted in headlines, argued over on X, and printed across every screener you’ve ever opened.
Almost nobody explains what they actually measure.
So this is the piece that does. By the end, you’ll know how each one is built, when it tells you the truth, when it quietly lies, and exactly what to do with it once you have the number in front of you.
There’s a cheat sheet at the bottom you can download. Every multiple in here, what it’s good for, and what breaks it.
Let’s start with a chart that looks impossible.
The P/E and P/S multiple
P/E: what you pay per dollar of profit
Google at 17 means you pay $17 for every $1 it earns in a year.
Good for: steady businesses where profit is real and repeats.
Weak for: anything lumpy or losing money. A tiny bottom number makes a healthy company look insane.
What breaks it: one-off events. Fines, tax bills, asset sales, paper gains.
Google shows both failures on the same chart.
In 2018 it read 60. A $9.9 billion tax bill and a $2.7 billion EU fine had gutted that year’s profit. Search was fine. The denominator wasn’t. Small profit, big P/E.
Today it reads 17. Last quarter Google booked $112 billion of profit, and $99 billion of that was stakes in other companies rising on paper. No customer paid for it. Strip it out and real earnings were $2.85 per share against a headline of $9.11. Big profit, small P/E.
Open the income statement and look at two lines, "other income" and tax. If either is unusually large this year, throw the headline number away and rebuild the multiple on the company's adjusted earnings instead. Google's $99 billion paper gain roughly doubled last quarter's profit, so the honest P/E is a long way above 17.
P/S: what you pay per dollar of sales
Google at 9.4 means you pay $9.40 for every $1 of annual sales.
Good for: young or unprofitable companies, where there’s no profit to divide by yet.
Weak for: comparing across industries. A grocer at 9x sales would be madness. Software at 9x is ordinary.
What breaks it: very little. Tax bills and paper gains never touch revenue, so the number stays honest.
Google’s sits at 9.4 today, up from roughly 6.7 in 2018. Sales grew a lot over those eight years. The price grew faster.
The trap is thinking 9.4 is high or low on its own. It isn’t either until you know what happens to those sales. Google keeps about 34 cents of operating profit from every dollar it takes in. A supermarket keeps two or three. Same 9.4, wildly different meaning.
Never read it alone. Pull up operating margin first, then ask what 9.4x sales becomes once those sales turn into profit. Google keeps around 34 cents per dollar. A retailer keeping two cents would need a P/S near zero to be worth the same.
The P/Gross Profit and Earnings Yield multiple
Earnings yield: the return the business earned on your money
5.81% means that for every $100 you put in, Google earned $5.81 last year.
Good for: comparing a stock to things that already speak in percentages. Bonds, savings accounts, rental property.
Weak for: any year where profit was unusual. The percentage looks generous for reasons that have nothing to do with the business.
What breaks it: the same one-off items that bend profit. Fines, tax charges, asset sales, paper gains.
Google’s line sat near 1.7% at the start of 2018 and reads 5.81% now, a rise of 244%. Almost none of that came from Search selling more ads. The 2018 figure was small because a tax bill and a fine had flattened that year’s profit, and today’s figure is large because $99 billion of paper gains inflated it.
Look up what the 10-year government bond pays today. Subtract it from the earnings yield. What’s left is roughly what you’re being paid for taking on business risk instead of lending to a government. If that gap is thin, you’re not being paid much to own the stock. Do the subtraction with adjusted profit, never the headline.
P/Gross profit: what you pay per dollar the business keeps after delivery costs
Gross profit is sales minus the direct cost of delivering them. It’s what’s left before salaries, marketing and research. Google at 15.43 means you pay $15.43 for each dollar of it.
Good for: companies that spend their profit down on purpose. Heavy research, heavy hiring, land-grab growth. Gross profit shows what the business could earn if it stopped.
Weak for: comparing across industries, because there’s no rule about which costs sit above the line and which sit below.
What breaks it: companies quietly shifting costs between “cost of revenue” and operating expenses. The multiple improves and nothing real has changed.
Google’s reads 15.43 today against roughly 11.3 in 2018. The line wanders less than most because gross profit is difficult to bend with one-off items.
Take two companies in the same industry, one profitable and one spending everything to grow. Compare them on this, then check whether the cheaper one’s gross margin is holding steady. A falling gross margin with a falling multiple is a business getting worse, and it is not a bargain.
P/OCF: what you pay per dollar of cash the business collects
Operating cash flow is the cash that actually came in from running the business, before any of it gets spent on buildings and equipment. Google at 22.57 means you pay $22.57 for each dollar of it.
Good for: checking whether reported profit is real. Cash is harder to invent than profit.
Weak for: capital-heavy businesses. It counts the money coming in and ignores the money that has to go straight back out to keep the lights on.
What breaks it: paying suppliers late, or collecting from customers early. Both flatter the number for a quarter and reverse the next one. Shares handed to staff get added back too, so companies that pay in stock look better than they are.
Google’s sits at 22.57 against roughly 20 in 2018. A calm line for eight years, because the cash keeps arriving.
Open the cash flow statement and find two things, the working capital lines and stock-based compensation. Add them up. If they explain most of the year’s operating cash flow, the multiple is flattering the company and you should discount it.
P/FCF: what you pay per dollar of cash left over
Free cash flow is what remains after the company pays for the equipment it needs. Google at 78.66 means you pay $78.66 for each dollar of leftover cash.
Good for: mature businesses that fund themselves. This is the cash that can actually pay a dividend, buy back stock or clear debt.
Weak for: companies in a heavy building phase. Money spent today is meant to earn money later, and this multiple records it as though it vanished.
What breaks it: one big investment year. Leases and acquisitions also sit outside this calculation, so a company can look cash-rich while still burning cash.
Google’s jumped from about 31 to 78.66. Nothing went wrong with ads. Capital spending went from $52.5 billion in 2024 to $91.4 billion in 2025, and the company has guided to between $195 and $205 billion for 2026. That comes straight out of the leftover cash, so the multiple tripled while the business kept growing.
Compare capital spending to depreciation. Roughly equal means the company is just maintaining what it owns, and the free cash flow is genuine. Far above depreciation means it’s building something, and your job becomes working out what those new assets will earn. Google’s capital spending is now several times its depreciation, which tells you this multiple is measuring an investment decision rather than a valuation.
FCF yield: the leftover cash as a percentage of what you pay
1.27% means that for every $100 you put in, Google had $1.27 of cash left over last year after paying for its equipment.
Good for: a quick sanity check against anything else you could buy. A savings account, a bond, a flat you’d rent out.
Weak for: companies in the middle of a build. The number describes what they chose to spend, not what they’re capable of earning.
What breaks it: one heavy investment year. Acquisitions and leases also sit outside the calculation, so real cash can leave without ever touching this number.
Google’s was around 3.2% in 2018 and reads 1.27% today, a fall of 61%.
Ask one question. Could the company stop spending tomorrow? Google could slow its data centre build and this number would leap. A steel mill cannot stop replacing its furnaces. If the spending is optional, a low yield is a choice. If it’s mandatory, a low yield is the business.
P/B: what you pay per dollar of what the company owns on paper
Book value is everything the company owns minus everything it owes, as recorded by the accountants. Google at 6.54 means you pay $6.54 for each dollar of it.
Good for: banks, insurers and property companies. Their assets are the business, and those assets are carried at something close to what they’d fetch.
Weak for: nearly everything else. Google’s brand, its people, its search index and decades of software cost billions to build and appear nowhere on that balance sheet. Book value simply doesn’t see them.
What breaks it: anything that moves the balance sheet without touching the business. New shares, buybacks, write-offs, paper gains.
Google’s read above 10 earlier this year and sits at 6.54 now. Shareholders’ equity reached $640 billion by the end of June, lifted by $99 billion of paper gains landing in retained earnings and $49.6 billion raised in June from selling new shares and preferred stock. The multiple got a third cheaper. The company did not.
Only reach for this when the balance sheet actually holds the business, and then pair it with return on equity. A bank earning 15% on its book value has earned a higher multiple than one earning 6%. For a company like Google, put this one down and use something else.
First, what enterprise value means
Imagine buying a house for $500,000. It comes with a $200,000 mortgage you have to take over, and the seller leaves $50,000 in cash in a safe in the kitchen.
The sticker says $500,000. What you’re really taking on is $650,000.
Enterprise value does that to a company. Take the share price times the shares, add the debt, subtract the cash. It’s the cost of owning the whole thing, debts and all.
EV/Sales: what the whole company costs per dollar of sales
Google at 9.15 means the entire business, debts included, costs $9.15 for each dollar of annual sales.
Good for: comparing two companies where one borrowed heavily and the other didn’t. A share price hides debt. This doesn’t.
Weak for: comparing across industries, because a dollar of sales is worth wildly different amounts depending on what it costs to produce.
What breaks it: a very large cash pile or a very large debt pile. Either one moves the number without the business changing at all.
Google’s reads 9.15 against roughly 5.8 in 2018. It held $242 billion of cash and marketable securities against $98 billion of long-term debt at the end of June, so subtracting that net cash makes the whole company slightly cheaper to buy than the share price alone suggests.
Reach for this whenever a company carries real debt. Then check what the cash is for. Cash already promised to an acquisition or a dividend isn’t yours to subtract, and treating it as though it is makes a company look cheaper than it is.
EV/Gross profit: what the whole company costs per dollar kept after delivery costs
Google at 15.02 means the entire business costs $15.02 for each dollar of gross profit.
Good for: ranking direct rivals. It handles different debt loads and different spending habits at the same time.
Weak for: crossing industries, because there’s no rule about which costs sit above the gross line and which sit below.
What breaks it: a company shifting costs from “cost of revenue” into operating expenses. Gross profit rises, the multiple falls, and nothing real happened.
Google’s sits at 15.02 today, up from around 10 in 2018.
Line up four or five direct competitors and rank them on this number. Then rank them again on how their gross margin has moved over three years. A company that’s cheapest on the multiple with a steady or rising margin is worth your afternoon. Cheapest with a falling margin usually means everyone else already spotted the problem.
EV/EBIT: what the whole company costs per dollar of operating profit
Operating profit is what’s left after every cost of running the business, before interest and tax. Google at 27.63 means the entire business, debts included, costs $27.63 for each dollar of it.
Good for: comparing companies in different countries or with different debt loads. Tax rates and interest bills sit below this line, so they can’t distort it.
Weak for: businesses where interest is a genuine cost of doing business, like banks and leasing companies. Skipping past it hides the point.
What breaks it: acquisitions. Buying a company creates a yearly write-down of the price paid, and that charge lands inside operating profit. The multiple gets worse while the business gets bigger.
Google’s reads 27.63 against roughly 24.5 in 2018. The $99 billion of paper gains never touched this number, because gains on investments sit below operating profit. Neither did the 2017 tax bill. What did land inside it is Google’s June purchases of Wiz for $29.5 billion and Intersect for $5.9 billion.
Find the amortization of acquired intangibles in the notes. If it’s large and the company is a serial acquirer, add it back and recalculate. Then ask whether those acquisitions actually produced the revenue growth. If they didn’t, keep the charge in.
EV/EBITDA: the same thing with wear and tear added back
Take operating profit and add back the yearly charge for buildings and equipment wearing out. Google at 23.6 means the whole business costs $23.60 for each dollar of that figure.
Good for: comparing companies with different asset ages. A firm with a new factory and a firm with an old one look more alike on this.
Weak for: anything that has to keep replacing its equipment. Adding back the wear and tear pretends the replacement never has to be bought.
What breaks it: a heavy building phase. The wear-and-tear charge grows every year afterwards, and adding it back flatters the company most at the exact moment it’s spending most.
Google’s sits at 23.6 today, up from about 19.4 in 2018. Capital spending is running toward $195 to $205 billion this year, and the charge for wearing all of that out grows behind it. Every dollar of that charge gets added straight back into this multiple.
Treat the added-back figure as a bill that arrives later. Subtract the company’s actual capital spending from it. If capital spending is bigger, the multiple is describing a business that keeps needing money, and you should reach for a measure that counts the spending instead.
Quick note on where every chart in this piece came from.
All of them are from Fiscal. It’s what I use daily for analyzing stocks, listening to earnings calls, and pulling up metrics, multiples, and fundamentals without hunting through filings for twenty minutes first.
Full transparency: I’m not sponsored, but I am affiliated. You get 20% off your subscription and I get a small commission. Plenty of you already know this is the platform I actually use, and I’d recommend it either way.
You can start a trial without a card and decide for yourself.
EV/OCF: what the whole company costs per dollar of cash collected
Google at 21.97 means the entire business, debts included, costs $21.97 for each dollar of cash it collects from running.
Good for: comparing a company that borrowed to grow against one that didn’t. Interest payments come out of this cash, and the debt itself sits in the price you’re measuring against.
Weak for: capital-heavy businesses. It counts what came in and ignores what has to go straight back out.
What breaks it: paying suppliers late or collecting early. Both lift the cash for a quarter and reverse the next. Shares paid to staff get added back too.
Google’s reads 21.97 against roughly 18.4 in 2018. A steady line for eight years, even while the company issued $51.8 billion of new notes and raised $49.6 billion in June from selling stock. All of that borrowing lands in the top of the fraction, so this number would have punished the company if the cash coming in hadn’t grown alongside it.
Use this one when a company has been buying rivals or borrowing to expand. Then check the interest cover in the notes. Cash flow that comfortably covers interest several times over means the debt in the numerator is safe. Cash flow that barely covers it means the multiple is the least of your worries.
EV/FCF: what the whole company costs per dollar of cash left over
Google at 76.57 means the entire business costs $76.57 for each dollar of cash remaining after it pays for its equipment.
Good for: the closest thing to a real price on a settled business. It’s the full cost of ownership set against the cash you could actually take out.
Weak for: anything mid-build. The spending is meant to create future cash and this treats it as gone.
What breaks it: one heavy investment year, and the gap between what counts as capital spending and what doesn’t. Leases and acquisitions drain cash without appearing here.
Google’s went from around 27 to 76.57. Capital spending guided at $195 to $205 billion for 2026 is doing nearly all of that, and the $29.5 billion Wiz deal and $5.9 billion Intersect deal never touched the calculation at all.
Turn it upside down before you judge it. Divide 1 by 76.57 and you get 1.3%, which is the cash return on your money at today’s price. Compare that to a government bond. Then decide whether the spending that crushed it is buying something worth waiting for. If you can’t name what that something is, the multiple is telling you to walk.
The forward versions
Every multiple in this article can be built two ways. Backwards, using what the company already reported. Forwards, using what analysts expect over the next twelve months.
The rule for reading a forward number is the same no matter which multiple it’s attached to.
Good for: businesses where last year was strange. Google’s trailing profit is carrying $99 billion of paper gains that nobody forecasts and nobody expects again, so a forward P/E quietly drops all of it. The distortion fixes itself.
Weak for: anything unpredictable, or anything thinly covered. Two analysts on a small company is a rumour with a decimal point.
What breaks it: optimism. Estimates start high and get cut as the year goes on, so a forward multiple usually looks cheaper in January than the same year deserves. Analysts also mix GAAP and adjusted figures, and different data providers pick differently, which is why two websites can show the same stock at 19 and at 24 on the same afternoon.
Check three things before you trust one. How many analysts contribute. Which way the estimates have moved over the last ninety days. Whether the figure is GAAP or adjusted. Rising estimates from a dozen analysts on an adjusted basis is a usable number. A falling estimate from three is a warning that the multiple hasn’t caught up yet.
And a habit worth building. Calculate the multiple both ways and look at the gap. A forward number far below the trailing one means the market is counting on a big jump in profit. Your job is to find out what that jump depends on, then decide whether you believe it.
Getting your cheat sheet
I promised you the cheat sheet, and it’s ready.
Here’s how to get it. Drop a comment below with the words cheat sheet in it. That’s the whole trick. I’ll send it over.
If you have a thought about the piece, I’d love to read it. Which multiple you’ve been misusing, which one you’d argue with me about, what you’d add. Comments like that are the best part of running this thing. A plain “cheat sheet” works perfectly well too.
One catch. Make sure you’re subscribed, because that’s how I get your email address. If you’re not, fix that now, and you’ll never miss another piece.
Why the comment instead of an attachment? Attaching files to a newsletter this size is a fast way to get the whole thing dumped into spam by Gmail and Outlook. So we do it this way instead, and everyone’s inbox stays happy.
—Yorrin
Sehr gut verständliche Übersicht. Freue mich auf den Spickzettel !
Cheat sheet