Direct Answer
Consensus EPS and revenue estimates are the aggregated forecast, typically the mean or median, of individual sell-side analyst estimates for a company's earnings per share or revenue in a specific upcoming fiscal period. The consensus is a single summary figure, not a prediction any one analyst necessarily agrees with, and it can be built differently (mean vs. median, GAAP vs. adjusted) by different data providers.
Key Takeaways
A consensus estimate is not a single analyst's forecast, it's a roll-up of many individual sell-side forecasts into one aggregated number for a specific metric and a specific fiscal period. How that roll-up is calculated, how many analysts feed into it, and whether it's built on GAAP or adjusted earnings all shape what the headline number actually means, and none of that context shows up in the number itself.
- Consensus estimates aggregate individual analyst forecasts for a specific metric and fiscal period, most commonly using the mean or median.
- Mean and median can diverge meaningfully when one or more analysts sit far from the rest of the group.
- Coverage size varies enormously by company size, and thin coverage gives each individual analyst outsized influence on the aggregate.
- Most published consensus EPS figures are built on adjusted, non-GAAP earnings, not strict GAAP EPS.
- A single consensus number hides the spread of opinion behind it, that spread is measured separately as estimate dispersion.
What Is a Consensus Estimate?
A consensus estimate is the aggregated forecast of the individual analysts who cover a stock, rolled up into a single number for one specific metric, most commonly earnings per share (EPS) or revenue, and one specific fiscal period, such as "next quarter" or "full fiscal year 2027." Data providers collect each contributing analyst's individual estimate, then combine them, usually as a mean (simple average) or a median (the middle value once estimates are sorted).
The consensus is not a forecast any single analyst produced. It's a statistical summary of many independent forecasts, and the market widely treats it as the reference expectation against which an actual reported result gets judged. When a company "beats consensus," it means the reported number came in above this aggregated figure, not above any specific analyst's individual number.
Why the aggregation method matters
Two different data providers covering the same stock for the same quarter can publish two slightly different "consensus" figures, and the aggregation method is one of the main reasons why. A mean is calculated by summing every contributing estimate and dividing by the number of analysts; a single analyst whose estimate sits well above or below the rest of the group pulls the mean toward their number more than an equally-weighted vote would suggest. A median instead takes the middle value of the sorted estimates, which resists that kind of pull from an outlier, but a median can also make a group of analysts look more unified than they actually are, since it reflects only the center of the distribution and says nothing about how far apart the highest and lowest estimates sit.
Providers don't always disclose which method they use by default, and some let contributing analysts drop or update estimates on different schedules, which introduces further small differences between "the consensus" reported by one source and another for the identical company and period.
Common mistake
The common mistake is treating "the consensus" as a single, universally agreed-upon figure rather than a calculated statistic that depends on a specific aggregation method and a specific set of contributing analysts as of a specific date. Two sources can both be reporting an accurate consensus and still show different numbers.
How Does Analyst Coverage Size Affect the Consensus?
Coverage size, the number of analysts contributing an estimate, varies dramatically by company size, and it directly changes how much weight rides on each individual contributor. A widely followed mega-cap stock might have 30 to 40 analysts publishing estimates for the same quarter, so no single analyst's view can move the aggregate very far on their own. A small-cap company might have only two or three analysts covering it, or none at all, in which case no consensus exists at all for that stock.
With thin coverage, one analyst updating their model, after a management call, a channel-check, or simply a routine refresh, can shift the entire published consensus by itself. That's a meaningfully different situation from a mega-cap where dozens of independent views are being averaged together, and it's worth checking the number of contributing analysts, not just the consensus figure, before treating a small-cap's "consensus beat" or "consensus miss" as a strong signal.
Common mistake
The common mistake is applying the same confidence to a consensus estimate regardless of how many analysts contributed to it. A consensus built from three analysts and a consensus built from thirty-five are not equally reliable summaries of collective expectations, even when both are labeled "consensus" the same way.
Is the Consensus EPS Figure Usually GAAP or Non-GAAP?
Most consensus EPS figures published by data providers are built on adjusted, or non-GAAP, EPS rather than a company's strict GAAP EPS. Adjusted EPS typically excludes items analysts treat as one-time or non-operational, restructuring charges, certain acquisition-related costs, and in many cases the accounting impact of stock-based compensation, on the theory that these items don't reflect the ongoing, comparable earnings power of the business.
The practical consequence is that a company's own GAAP EPS, as reported in its official earnings release, and the "consensus EPS" quoted in a headline are frequently not measuring the same thing. Comparing the two directly, without checking whether the consensus figure was built on an adjusted basis, is a common analytical error, it can make a company look like it beat or missed by a margin that doesn't actually reflect a genuine gap against what analysts were modeling.
Fiscal year timing also matters
"This quarter's consensus estimate" doesn't mean the same calendar period for every company. Fiscal year-ends vary, many companies use a calendar year, but a meaningful number use fiscal years ending in months like June or September. A company's "Q3" can therefore refer to different calendar months depending on when its fiscal year starts, so the actual fiscal period being described, not just the quarter label, needs to be checked before comparing estimates across companies or against a prior period.
Common mistake
The common mistake is assuming a "Q3" or "fiscal 2027" label means the same calendar window across every company being compared. Two companies with different fiscal year-ends can have "Q3" estimates covering entirely different calendar quarters.
Worked Example: What "$2.40 Consensus" Actually Hides
Illustrative numbers, not live market data.
Assume a hypothetical company is covered by 12 sell-side analysts, each of whom has published an individual EPS estimate for the upcoming quarter. Across those 12 analysts, the individual estimates range from roughly $2.10 at the low end to $2.75 at the high end, and the mean of all 12 works out to approximately $2.40.
A headline or data feed reporting "consensus EPS: $2.40" is accurate, but it collapses a $0.65 range of professional opinion into a single number. That headline alone doesn't tell a reader whether all 12 analysts are clustered tightly around $2.40 with only minor variation, or whether the group is genuinely split, some confident in a strong quarter near $2.75, others expecting real weakness near $2.10, with the $2.40 mean simply sitting in the middle of a wide disagreement.
Those are two very different situations that can produce the identical consensus figure. A tightly clustered set of estimates suggests broad analyst agreement about the company's near-term trajectory. A widely spread set suggests real uncertainty or genuine disagreement about the outlook, even though the single published number looks the same either way. Measuring that spread directly, rather than inferring it from the consensus alone, is the specific gap that Estimate Dispersion covers, and it's worth reading alongside this page for anyone who wants more than the headline number.
Misconceptions Versus Reality
| Misconception | Reality |
|---|---|
| "The consensus" is a single, objective number every data source agrees on | Different providers can use different aggregation methods (mean vs. median) and different contributor sets, producing slightly different consensus figures for the same stock and period |
| A consensus estimate reflects broad agreement among all covering analysts | A single mean or median can result from either tight agreement or wide, offsetting disagreement, the consensus figure alone doesn't distinguish between the two |
| Consensus EPS is the same as the company's reported GAAP EPS | Most published consensus EPS is built on adjusted, non-GAAP earnings; comparing it directly to GAAP EPS without adjusting is a common analytical error |
| A small-cap's consensus estimate is just as reliable as a mega-cap's | Coverage can be as thin as two or three analysts for a small-cap versus 30-40 for a mega-cap, so each contributor's view carries far more weight on a thinly covered stock |
Risks, Limitations, and Exceptions
- No consensus exists at all for stocks with zero analyst coverage, which is common among micro-caps and thinly traded names.
- Consensus figures can lag real-world developments if some contributing analysts haven't refreshed their estimate recently, even though the aggregate is reported as current.
- Mixing GAAP and non-GAAP figures when comparing consensus to an actual reported result can produce a misleading "beat" or "miss" that doesn't reflect the real gap analysts were modeling.
- A consensus estimate describes an aggregated expectation, not a prediction guaranteed to be accurate, and should not be treated as investment advice or a standalone trading signal.
- Fiscal year-end differences between companies mean the same quarter label can cover different calendar periods, which can distort side-by-side comparisons if not checked.
One Number Standing In for a Crowd
A consensus figure is a summary statistic, and the useful move is to look behind it before relying on it. How many analysts contributed, how widely they disagree, how recently each estimate was updated, and whether the basis is reported or adjusted earnings all change what the single number means. Two identical consensus figures can rest on very different amounts of information.
That matters most where coverage is thin. A consensus assembled from a handful of contributors is highly sensitive to any one of them, and a single revision can move the figure in a way that looks like a shift in view across the whole market.
The comparability problem is what catches people out. Providers apply different inclusion rules, different staleness cut-offs and different treatment of adjusted measures, so a figure from one source and a reported result compiled on another basis can disagree without either being wrong.
A consensus carries no predictive authority. It records what a set of forecasters currently expects, and expectations get revised precisely because they turn out to be wrong.
Frequently Asked Questions
What are consensus EPS and revenue estimates?
A consensus estimate is the aggregated forecast, typically the mean or median, of the individual estimates published by sell-side analysts who cover a stock, for a specific metric (EPS or revenue) and a specific fiscal period. Instead of relying on any one analyst's number, the consensus rolls up every contributing estimate into a single reference figure that the market treats as the benchmark expectation for that period.
Why do mean and median consensus figures sometimes differ?
The mean (simple average) is sensitive to outliers: a single analyst with an estimate well above or below the rest of the group pulls the average toward their number. The median (the middle value when estimates are sorted) is more resistant to that kind of distortion, but it can also understate genuine, widespread disagreement among analysts since it only reflects the middle of the distribution, not its spread. Different data providers sometimes default to different aggregation methods, which is one reason the reported "consensus" for the same company can vary slightly from one source to another.
Is the consensus EPS figure usually GAAP or non-GAAP?
Most consensus EPS figures published by data providers are built on adjusted, or non-GAAP, EPS, earnings with items like one-time charges, certain stock-based compensation effects, and other exclusions analysts treat as non-recurring stripped out, rather than a company's strict GAAP EPS. Comparing a company's own GAAP EPS headline against a non-GAAP consensus figure without adjusting for the difference is a common analytical error, since the two numbers are not measuring the same thing.
How does a stock split affect a historical consensus EPS series?
A split changes the share count, so every per-share figure before the split date is stated on a different basis than every figure after it. Data vendors normally restate historical estimates to be split-adjusted, but the restatement is applied retroactively, which means a series that looks continuous today did not look that way at the time. For research that depends on what the consensus was on a past date, the adjustment convention needs to be confirmed rather than assumed.
What happens to consensus estimates when a company changes its fiscal year end?
The period labels stop lining up. Analysts have to rebuild their models around the new calendar, often producing a short transition period that belongs to neither the old nor the new annual cycle. During the changeover, some analysts publish on the new basis while others have not updated, so the aggregate can briefly mix incompatible periods. Any comparison of "next fiscal year" estimates across a fiscal-year change needs the period definitions checked explicitly.
Why do two data providers show different consensus numbers for the same quarter?
Providers differ in which analysts they collect from, how long they keep an estimate before treating it as stale, whether they use the mean or median, and how they treat estimates published on different accounting bases. A firm whose analyst has not updated in four months might be included by one provider and excluded by another. These are methodology differences rather than errors, which is why a consensus figure is more useful when quoted alongside the provider and the as-of date.
Are revenue estimates generally easier to forecast than EPS estimates?
Revenue sits at the top of the income statement and is affected by fewer discretionary accounting choices, while EPS sits at the bottom after cost of sales, operating expenses, interest, tax and share count have all been applied. Small proportional errors in several of those lines can compound into a large proportional error in earnings per share. That structural difference is why revenue and EPS surprises for the same quarter can point in opposite directions.
How are consensus estimates handled when a company reports in a foreign currency?
Vendors typically publish the consensus in the company's reporting currency and may also provide a converted version. Conversion introduces an exchange-rate assumption that belongs to the vendor rather than to any analyst, and a consensus that moves purely because of currency translation can look like a revision when no analyst changed their view. When comparing a reported result to a consensus for a cross-border listing, confirm both figures are stated in the same currency on the same basis.
Is a whisper number part of the consensus?
No. A whisper number is an informal expectation circulating among traders and commentators that sits above or below the published consensus, and it is not collected, verified or aggregated by data vendors. It has no defined methodology and no audit trail, so two people quoting a whisper number for the same company may be quoting different figures. It can help explain why a stock reacted unexpectedly to a reported beat, but it is anecdote rather than data.
References
This guide describes long-standing, widely documented sell-side research conventions around estimate aggregation and disclosure. Key reference sources include:
- U.S. Securities and Exchange Commission, Regulation FD: sec.gov: the fair-disclosure framework governing how companies communicate financial information to analysts and the market.
- CFA Institute, Equity Research and Valuation: cfainstitute.org: professional standards and methodology references for equity research practice.
The worked example in this guide uses clearly labeled illustrative numbers, not live consensus data. This content was reviewed by the Swoopr Editorial Team in August 2026.