The whisper number: the bar earnings must clear that headlines don't show
84% of S&P 500 companies beat earnings estimates in the third quarter — yet earnings came in only 2.7% above them. The published consensus is not the bar that moves stocks; the whisper number is.
Two bars, one report
Every earnings report is graded against two bars, but only one of them is printed in the headlines. The first is the consensus estimate: the average of the earnings-per-share forecasts published by analysts who cover the company. The second is the whisper number: the unofficial, unpublished forecast that circulates among professionals on Wall Street — the number analysts actually expect, which is often different from the number they published.
Investopedia describes the whisper as the newest, although unwritten, consensus expectation. The logic is simple: an analyst's published estimate can sit unchanged for weeks while their real view moves with each new piece of information, and the updated view is what gets passed around trading desks. Investopedia gives a clean example of how the two bars collide: suppose the whisper for a company is $0.85 per share, and the company reports $0.80. The result beats the published consensus, so the headlines call it a beat — but it falls short of what investors were actually expecting, and the stock can fall anyway. The beat was real. The bar was wrong.
Why beats are so common — and so thin
Third-quarter 2026 shows how the game works at index level. FactSet's November 6 scorecard, reported by MarketWatch, found that 88% of S&P 500 companies had reported results, and 84% of them had beaten earnings-per-share estimates — above the five-year average of 78% and the ten-year average of 77%. On revenue, 74% beat estimates, above the five-year average of 70% and the ten-year average of 68%.
But the size of the beats tells a different story. In aggregate, companies reported earnings only 2.7% above estimates, below the five-year average of 7.5% and the ten-year average of 8.0%. If that figure holds, it will be the smallest aggregate earnings surprise since the fourth quarter of 2022, when the surprise was 1.3%. Revenue told the same story in miniature: aggregate revenue was 2.1% above estimates, slightly above the five-year average of 1.8%. The blended earnings growth rate for the quarter was 31.4%.
The market's reaction matched the thin margins. FactSet found that companies beating estimates rose 0.8% on average from two days before to two days after their reports, slightly below the five-year average of 0.9%, while companies that missed fell 1.1%, against a five-year average decline of 3.1%. Beating a bar that has been set low earns a small reward; missing it still costs, just less than usual. The frequency of beats says the bar was low. The size of the beats says how low.
- Companies that had reported88%
- Beat EPS estimates84% (five-year average: 78%)
- Aggregate earnings above estimates2.7% (five-year average: 7.5%)
- Beat revenue estimates74%
- Blended earnings growth31.4%
Nvidia is the live test
Nvidia reports its fiscal third-quarter results on Tuesday, November 17, after the market closes — the quarter ended October 25, 2026 — with the results at about 1:20 p.m. Pacific time and the conference call at 2 p.m. Pacific, the company said. Wall Street expects revenue of $109.4 billion and net income of about $60.9 billion, according to Visible Alpha's analyst survey cited by Investopedia, while the company said in August that it expected $108 billion in revenue.
The published bar is visible. The real bar is not. Nvidia's shares were down 12% in the week to Friday's close and more than 30% below October's record high, and options pricing points to a move of about 8% in either direction by the end of next week, Investopedia calculated. The report is also the company's first under acting chief executive Colette Kress, the chief financial officer since 2013, after co-founder Jensen Huang stepped down this week. A result above $109.4 billion would beat the published consensus — and could still disappoint if the whisper sits higher. That is the test on Tuesday: not whether Nvidia beats the number in the headlines, but whether it clears the number nobody printed.
How to read the next report
Start by writing down both bars before the report lands: the published consensus and the company's own guidance for the quarter. Then, when the numbers arrive, compare the result with both — and look at the size of the beat, not just whether there was one. An aggregate surprise of 2.7% across the S&P 500 is a reminder that most beats this season were thin.
Then read the guidance for the next quarter, which often moves a stock more than the quarter just reported, and check the cash behind the profit, as the earlier pieces in this series explain. And remember what the whisper is: an unpublished expectation, not a number you can look up. If a stock falls on a headline beat, the likely explanation is not that the market misread the report — it is that the market was grading against a bar the headlines never showed. Educational content, not investment advice.
