Four times a year, thousands of public companies open their books, and for a few hours the numbers seem to speak for themselves. They rarely do. The same report that sends one company’s stock up 10 percent can send an outwardly similar company’s stock down just as far — not because the arithmetic is different, but because the market was pricing in something the report didn’t deliver.

Understanding earnings season means understanding two separate scorecards: what a company actually did, and what Wall Street had already bet it would do.

Two numbers, and what each one measures

Every earnings report centers on two headline figures. Revenue is the total amount of money a company took in from customers — the top line, before any costs are subtracted. Earnings per share, or EPS, is what’s left of profit after every expense, tax, and interest payment, divided by the number of shares outstanding. A company can grow revenue while EPS falls, if costs are rising faster than sales; it can also grow EPS while revenue is flat, by cutting costs, buying back shares, or catching a favorable tax rate. Reading only one number tells half the story.

Companies also report both GAAP figures, calculated under standardized accounting rules, and “adjusted” or non-GAAP figures that strip out items management considers one-time or non-operational, like restructuring charges or stock-based compensation. The two can diverge significantly, which is why analysts typically track both and are wary of a widening gap between them.

The number that actually moves the stock: consensus

Long before a company reports, Wall Street analysts who cover the stock publish their own estimates for what revenue and EPS will be. Data providers like FactSet and LSEG average those individual forecasts into a single “consensus estimate,” and it’s the gap between the actual result and that consensus — not the result in isolation — that usually drives the immediate stock reaction. A company growing profit 20 percent year-over-year can still fall if analysts had modeled 25 percent growth; a company with flat profit can rally if analysts feared a decline.

Beating consensus is also far more common than intuition suggests, which is part of why the market doesn’t treat every beat as good news. According to FactSet’s Earnings Insight research, 86 percent of S&P 500 companies beat EPS estimates in the second quarter of 2026, well above the five-year average beat rate of 78 percent and the ten-year average of 76 percent. When the vast majority of companies beat estimates most quarters, a beat alone carries less information than the size and source of that beat.

McKinsey’s research on what it calls the “consensus-earnings trap” found the market’s reaction is more measured than the drama around earnings season implies: missing consensus by 1 percent was associated with only about a 0.2 percent drop in share price over the following five days, on average. The exception McKinsey identified was chronic underperformance — companies that missed consensus in four or more of seven consecutive years saw statistically significant, lasting stock declines, regardless of the size of any single miss.

Why guidance can matter more than the quarter just reported

Alongside historical results, most companies issue guidance: management’s own forecast for revenue or earnings in the coming quarter or year. Because a stock price reflects expectations about the future, not a scorecard of the past, guidance often moves shares more violently than the reported quarter itself.

Netflix’s second-quarter 2026 report, released July 17, 2026, is a clean real-world illustration. The company beat EPS estimates by a penny, landing at $0.80, and matched revenue expectations at $12.56 billion — by traditional measures, a solid quarter. Yet shares fell as much as 12.2 percent the next morning, settling to roughly a 9.1 percent decline, according to reporting on the results. The culprits weren’t the historical numbers: Netflix’s third-quarter guidance came in just below Wall Street’s consensus, engagement-hours growth of just 2 percent in the first half of the year looked weak relative to subscriber growth, and the company’s decision to report engagement data only annually going forward, starting in 2027, was read by some investors as an attempt to obscure a slowing trend.

What a report doesn’t tell you

An earnings report is a snapshot of accounting outcomes, not a verdict on the business. It doesn’t capture competitive dynamics shifting outside the reporting window, doesn’t reveal cash burn hidden by favorable working-capital timing, and doesn’t distinguish between profit growth from genuine operating improvement and profit growth from financial engineering, like a lower share count from buybacks or a one-time tax benefit. McKinsey’s research specifically flagged that firms sometimes manage to a beat by cutting R&D or marketing spend, or by offering steep end-of-quarter discounts to pull sales forward — moves that can flatter one quarter’s numbers while quietly damaging the next several.

The most useful way to read an earnings report, then, is less as a single verdict and more as one data point in a trend: how results compare to what was expected, how guidance is shifting, and whether the underlying growth and margin trajectory over several quarters looks more like the Netflix pattern of a “beat” masking real concerns, or like a genuine, durable improvement in the business.

Sources: FactSet Insight, “S&P 500 Earnings Season Update: July 24, 2026” (insight.factset.com); McKinsey & Company, “Avoiding the consensus-earnings trap” (mckinsey.com); Yahoo Finance, “Netflix Beat Estimates Again; Why Did the Stock Drop 12% Anyway?” July 2026 (finance.yahoo.com).