Company Analysis 101: Understanding Earnings
“NVIDIA just posted record earnings!” “That company had an earnings miss.” You’ve probably heard lines like these on the news. And yet — you might have assumed “good earnings mean the stock goes up,” only to watch a stock fall right after a great report and wonder what on earth just happened.
If the financial statements we covered earlier are like a company’s “health check-up,” then earnings (the quarterly report) are more like a “report card” that arrives every three months. They’re your freshest read on whether a company is doing well right now. Let’s walk through how to read that report card — including the classic beginner puzzle: “great earnings, so why is the stock dropping?”
What earnings are, and when they arrive
Earnings are the numbers that show how much a company sold (revenue) and kept (operating income, net income) over a set period — usually a quarter (three months). Just as a student gets a report card each term, public companies have to give the market a report card every quarter.
The stretch when lots of companies report at once is called earnings season. Because reports cluster starting about a month after a quarter ends, that’s when the news fills up with phrases like “earnings surprise” and “earnings miss.”
The structure of the numbers — from revenue to EPS
One thing to keep in mind: earnings flow down through four layers.
For a beginner, two of these are the must-watch numbers: operating income and EPS.
2) Consensus is usually set in EPS, so “EPS beat/missed expectations” becomes the news headline.
3) It ties straight to the P/E ratio. P/E = Price ÷ EPS. When EPS rises, the same stock price looks cheaper.
4) Buybacks move it. When a company buys back and retires its own shares, the profit is split among fewer shares, so EPS goes up. NVIDIA’s $80B buyback authorization is, in the end, a move to boost each shareholder’s slice (EPS).
The three things to always check
A single earnings number tells you nothing on its own. Split it into three angles and it snaps into focus. Let’s follow along with a made-up company, Acme Electronics. Say this is their quarter (figures in $ millions):
1How big — size
The most basic question is “how large is it?” Acme did $1.2 billion in revenue and $240 million in operating income. But size alone can’t tell you whether that’s good — every company is a different size. That’s why you need the next two.
2How much it grew — growth rate (vs. a year ago vs. last quarter)
For growth, what you compare against is everything. There are two baselines.
Using Acme’s operating income: YoY = (240 − 150) ÷ 150 × 100 = +60% (up 60% from a year ago). QoQ = (240 − 220) ÷ 220 × 100 = about +9% (up 9% from last quarter).
Same result, yet YoY looks dazzling and QoQ looks modest. That’s why YoY is the default — because of seasonality. An ice-cream maker naturally sells more in summer than in winter; comparing to last quarter (QoQ) creates an illusion. Comparing to the same season last year (YoY) shows the true growth.
A real example — In its FY2027 Q1 (reported in 2026), NVIDIA posted $81.6B in revenue on booming AI-chip demand. Versus the same quarter a year earlier, that was +85% YoY — dazzling. Versus the quarter right before, it was +20% QoQ. Same result, very different impression depending on the baseline.
3How much it kept — margin (profitability)
Big sales with nothing left over is a losing game. So you look at “how much it keeps” — the margin. Watch two of them together.
Acme’s operating margin is 240 ÷ 1,200 × 100 = 20% — for every $100 sold, $20 is left as operating profit. But what really matters with margins is the trend. Rising means pricing power or good cost control; falling means competition is heating up or costs are climbing.
Did the profit turn into real cash?
So far we’ve looked at “how much it kept (profit).” But there’s one more thing to check: did that profit actually land in the bank?
Ever heard “profitable on paper, but still went under”? Book profit and the cash actually in the bank can differ. That’s why serious investors watch cash flow just as closely as profit — and earnings releases include these figures too.
Operating cash flow — did profit become real money?
This is the cash the core business actually brought into the bank. If a sale is made but not yet paid for (sold on credit), the income statement books it as revenue and profit — but the cash isn’t there yet. So if net income is positive while operating cash flow stays negative quarter after quarter, that’s a warning sign.
Free cash flow (FCF) — the real spare cash
Free cash flow = operating cash flow − capital expenditures. It’s what’s left after investing in factories and equipment — the cash a company is free to actually use.
2) It funds dividends and buybacks. NVIDIA’s $80B buyback and 25× dividend hike are only possible with plenty of spare cash behind them.
3) It shows whether growth is sustainable. For capital-heavy companies (like AI), the key question is “does any cash survive all that spending?”
Where the money goes — and can you trust it?
Once the profit has landed as cash, two questions follow: where does the company spend that money, and how much can you trust the reported numbers?
Where the money goes — capital allocation
How a company uses the cash it earns is worth watching just as much as the earnings themselves. It broadly splits four ways.
NVIDIA raising its dividend 25× and announcing an $80B buyback this quarter is one example of capital allocation — returning earned cash to shareholders.
How much to trust it — quality of earnings
Even good numbers deserve a second look.
Related to those “one-off” items, US earnings come in two flavors: GAAP and adjusted (Non-GAAP).
For context: most markets outside the US use a different rulebook called IFRS, but the goal is the same — a shared yardstick so companies can’t just count things however they like.
In NVIDIA’s FY2027 Q1, GAAP EPS was $2.39 — but that included a one-time gain from the rising value of stock holdings. Strip it out and adjusted EPS was $1.87, which is the number the market treated as the company’s real strength.
Good numbers aren’t enough — they must beat expectations
Now for the part beginners miss most. In the stock market, earnings are less about the number itself and more about how it compared to expectations. Think of a test: a score of 90 matters less than “what did people expect, and then what came in?” If everyone expected 70 and you got 90, applause. If everyone expected 100 and you got 90, disappointment.
That “score everyone expected” is the consensus.
Reported vs. consensus = a four-tier verdict
In practice, each figure gets compared to consensus and sorted into four tiers. These are common in the news, so they’re worth knowing.
In practice, you line it up in one table
Real earnings write-ups lay the key figures side by side: reported · consensus · growth (YoY) · verdict. Here’s NVIDIA’s FY2027 Q1 in that format.
Lined up this way, what jumps out isn’t “is the number big?” but “did it clear the bar (consensus)?” If you can just build this table, you’ve read half of any earnings report.
A real example — Microsoft’s FY2026 Q4 (reported July 2026) posted adjusted EPS of $4.74 vs. a $4.24 consensus — about 12% above — with revenue of $90B topping the ~$87.7B estimate. With cloud (Azure) growth also well ahead of expectations, it was a textbook surprise.
Beyond the total: which business grew?
Once you’ve judged whether results beat expectations, peel the numbers back one more layer. Total revenue rising doesn’t mean it’s all the same growth. You have to see which business drove it to get the real picture — which is why earnings releases always break out revenue by segment.
NVIDIA — in FY2027 Q1, about 92% of revenue came from the data center segment. It’s essentially a “data-center company” now. That concentration is explosive while the AI boom lasts, but if that flow reverses, the whole company shakes. Concentration is both a strength and a risk.
Microsoft — splits broadly into Cloud (Azure, etc.), Productivity (Office), and Personal Computing (Windows, Xbox). In FY2026 Q4, Azure surged +43% YoY while the PC segment actually fell about 4%. Look only at total revenue (+18%) and you’d miss that contrast.
The same goes for Apple (iPhone vs. Services) and Google/Alphabet (Search ads, YouTube, Cloud): every quarter, “which segment carried it?” is the thing to watch.
In short, read earnings total first, then segment — always go one level deeper.
As important as results: the outlook
As much as past results — maybe more — the market fixates on what’s ahead. Alongside its earnings, a company shares its own forecast: “here’s what we expect next quarter.”
Here too, the key is the comparison to expectations (consensus).
One more thing to watch: did they raise or cut it versus last time (direction)? When a company lifts its outlook, it signals both a good quarter and a brighter road ahead — markets love that. In the phrase “beat & raise,” the “raise” refers exactly to this upward revision.
NVIDIA, in its FY2027 Q1 report, guided next-quarter revenue to $91B — more than $4B above the ~$87B consensus at the time. Strong results plus a raised outlook: a textbook beat & raise.
Behind the numbers: the earnings call
If guidance is the “forecast in numbers,” there’s also a place where management explains things out loud.
As a beginner you don’t need to listen to the whole thing — just tune in for three things.
Even just gauging “does management sound optimistic or cautious about next quarter?” from the news headlines helps a lot.
Where to find all these numbers
Everything we’ve covered — revenue, profit, cash flow, guidance — can be checked at the official sources.
Why a stock can fall on great earnings
Now, the question we teased at the very start. When a stock falls despite great earnings and even an earnings surprise — the opposite of common sense — why does it happen? Three things drive the post-report move.
The biggest trap — priced in and valuation
That third one — priced in — is the biggest beginner trap. A stock trades on the future, not the past. If the expectation of good results was already baked into the price months ago, then when the good results actually land, people shrug (“as expected”) and profit-taking can kick in instead.
Add valuation (was the stock already expensive?) on top. If the P/E ratio (Price ÷ EPS) is already high, the price is carrying a lot of “and it’ll keep being great” expectation. In that case, merely good results won’t satisfy — even a small blemish can send the stock down.
NVIDIA, FY2027 Q1 — revenue, data center, and EPS all beat, and it raised guidance too (a “beat & raise”) — yet the stock was muted right after. Why? (1) the good results were already priced in, (2) the size of the beat had been shrinking each quarter, and (3) attention had shifted from this quarter to the future — China export rules, competition, and whether AI investment keeps going.
Apple, FY2026 Q3 was similar. Revenue and EPS both beat in its “best-ever June quarter,” yet the stock fell 3–6% right after — because Services revenue came in a touch light, next-quarter guidance was cautious, and the stock had already run up a lot, setting the bar high.
So when you read earnings, ask not only “are the numbers good?” but also “is all this good news already in the price?”
Summary
Key formulas at a glance
The four-tier verdict
Glossary
Earnings-reading checklist
In practice — read it in this order
Wrapping up
Reading earnings comes down to a few ideas. First, a number only means something next to a point of comparison — against a year ago (YoY) and against consensus (expectations).
Second, go beyond the total to the segments, and beyond the past to the guidance — that’s when the picture is complete. Third, a stock trades on the future, not the past — which is why even record earnings can send a stock down.
Internalize just these, and you’ll read the flood of earnings-season headlines far more calmly. Instead of celebrating or panicking over a single number, ask what that number means against expectations and the road ahead — that’s where good earnings analysis begins.
