EarnoReview
Guide

Analyst Estimates and Consensus Explained: Beats, Misses, Revisions

How analyst estimates are built, how consensus is calculated and what a beat or miss compares. Covers research independence and company disclosure rules.

Oct 8, 20268 min read

An analyst estimate is a forecast of a company’s future financial results, such as revenue or earnings per share, prepared by a research analyst at a brokerage firm or independent research provider. The consensus estimate is an average or median of those individual forecasts, compiled by a data provider. When news stories say a company “beat” or “missed” expectations, they usually mean its reported figures were above or below the consensus, which makes the consensus one of the most quoted numbers in earnings coverage and one of the least understood.

This guide explains who makes estimates, how consensus is formed, why the numbers differ between sources, what the rules say about analyst independence and company disclosure, and how to read a beat or miss with care.

In brief

  • Estimates are opinions of analysts; guidance is the company’s own outlook.
  • Consensus is a statistic computed by a data vendor, not an official figure, so the same company can have slightly different consensus numbers on different platforms.
  • A beat or miss compares reported results with consensus on a particular basis, often adjusted earnings per share.
  • Rules govern analyst independence, certifications and what companies may tell analysts privately.

Who makes analyst estimates

Two broad groups publish estimates.

Sell-side analysts work for brokerage and investment banking firms. They publish research reports with forecasts, ratings and sometimes price targets, which are distributed to the firm’s clients. They cover a defined set of companies, usually within one industry.

Independent and buy-side analysts work for independent research firms or for investment managers. Buy-side analysts prepare forecasts for internal use, and those are generally not published, so they do not enter a published consensus. Independent research providers sell or publish their work and are sometimes included in data vendors’ compilations.

Because the sell-side produces most published estimates, the word “analyst” in earnings coverage typically refers to them.

What an estimate contains

A typical model covers:

  • Revenue, often by segment or product line.
  • Margins and expenses, built from assumptions about pricing, volume, headcount and input costs.
  • Earnings per share, derived from net income and the expected share count.
  • Other measures such as cash flow, capital spending, subscribers or same-store sales, depending on the industry.

Analysts set estimates for the next quarter, the current fiscal year and often the following year or two. They build them from the company’s financial statements, management’s guidance and commentary, industry data, channel checks and their view of the economy. The assumptions behind them can differ widely, which is why estimates for the same company are rarely identical.

How consensus is calculated

Data providers collect individual estimates from contributing analysts and compute a summary statistic, most often the mean (average) and sometimes the median, along with the high, low and number of estimates. The details vary by provider:

  • Which analysts are included. Providers differ in the firms they cover and in how quickly they add new contributions.
  • How stale estimates are treated. Some providers drop estimates older than a set period; others weight recent ones more heavily.
  • Which measure is used. For earnings per share, providers may use GAAP, a company-style adjusted figure, or the figure most analysts report. The choice matters when GAAP and adjusted numbers diverge.
  • How one-time items are handled. Analysts may exclude restructuring charges, acquisition costs or other items in different ways.

For these reasons, a consensus quoted by one outlet may differ from another’s. When you see a beat or miss, look at which consensus source is cited and which measure it uses. There is no official consensus published by the SEC or any exchange.

Estimates, guidance and “whisper numbers”

Three numbers are often discussed together:

  • Guidance. The company’s own forecast. Our guide on how earnings guidance works covers it. Analysts typically adjust their estimates in response to a guidance change.
  • Consensus. The compiled average of analysts’ estimates.
  • Whisper number. An informal, unofficial expectation circulated among traders, which may be higher or lower than consensus. It has no standard definition and no published source, so treat references to it as anecdotal.

The relationship between them explains much of the market reaction to earnings. A company might meet its own guidance and still fall short of consensus if analysts had expected an upside surprise.

What “beat,” “miss” and “in line” mean

A company beats when a reported figure is above the consensus for that measure, misses when it is below and is in line when it is roughly equal. Several points make these labels less informative than they sound:

  1. They compare to expectations, not to a standard of quality. A company can grow fast and miss if analysts expected faster growth.
  2. They depend on the measure. A company can beat on revenue and miss on earnings, or the reverse.
  3. Surprise size matters. A beat of a fraction of a cent is not the same as a large difference, and rounding can decide the label.
  4. Estimates are moving targets. Analysts often adjust their forecasts in the weeks before results, and the consensus a company is compared with may be the late, lower or higher version.
  5. Quality of the beat varies. A beat caused by a lower tax rate, a one-time gain or a share buyback tells a different story from one caused by higher sales.

Estimate revisions

After each report and each guidance change, analysts revise their numbers. Terms used in coverage include:

  • Upward or downward revisions. Changes in the consensus for a future period over time. A pattern of revisions in one direction shows a shift in expectations.
  • Estimate dispersion. The spread between high and low estimates. A wide spread suggests more disagreement or uncertainty.
  • Number of analysts. A consensus based on two analysts is less representative than one based on twenty.

Revisions reflect new information, but they also reflect analysts’ tendency to follow one another, so a consensus can move together without independent confirmation.

Rules that shape analyst research and company disclosure

Independence and conflicts. FINRA Rule 2241 governs research analysts and research reports of member firms. It requires written policies and procedures reasonably designed to identify and manage conflicts of interest related to the preparation, content and distribution of research reports, public appearances by analysts and the interaction between analysts and others in the firm, including investment banking personnel. The rule also covers disclosures in research reports. Check the rule text on FINRA.org for the specific provisions.

Analyst certifications. The SEC’s Regulation AC, at 17 CFR 242.501, requires a clear and prominent certification in a research report that the views expressed accurately reflect the analyst’s personal views, along with a statement about whether the analyst’s compensation was related to the specific recommendations or views in the report, and a description of any such compensation if it was.

Selective disclosure. Regulation FD (17 CFR 243.100) requires a company that discloses material nonpublic information to covered recipients, which include broker-dealers and investment advisers and their associated persons, to make that information public, simultaneously if the disclosure was intentional and promptly if it was not. The rule is the reason companies hold open conference calls and generally avoid giving individual analysts private guidance. The rule does not stop companies from discussing information that is already public.

Non-GAAP reconciliation. When a company publicly discloses a non-GAAP measure, Regulation G (17 CFR 244.100) requires the most directly comparable GAAP measure and a reconciliation. This helps readers match the measure used in estimates with the measure in the earnings release.

A worked example, with made-up numbers

Suppose five analysts forecast a fictional company’s quarterly adjusted earnings per share at 0.98, 1.00, 1.02, 1.03 and 1.07. The mean is 1.02 and the median is 1.02. The company reports 1.04.

  • Against the mean, that is a beat of 0.02.
  • The result is above four of the five estimates and below one, so the beat is not large relative to the spread.
  • If the same company’s GAAP earnings per share came in below the estimates because of a restructuring charge, a headline that says “beats” would be describing the adjusted measure only.
  • If the company also cut its full-year outlook, analysts would likely lower next quarter’s and next year’s estimates, and the consensus would drift down even after a beat.

A careful reader would check the measure used, the source of the consensus, the cause of the difference and the outlook.

Limits of consensus

  • It is backward-looking at the moment of the report. It reflects what analysts knew before the release.
  • It can reflect shared assumptions. Analysts draw on the same public information, and conflicts of interest are the reason for rules such as FINRA Rule 2241. Do not assume any single estimate is neutral.
  • It does not capture risk. An average does not describe the range of outcomes.
  • It is not advice. Neither the consensus nor a beat or miss is a recommendation to buy or sell.

Reading estimates responsibly

  1. Identify the source and date of the consensus.
  2. Confirm the measure (GAAP or adjusted, revenue or earnings).
  3. Look at the number of analysts and the high-low range.
  4. Compare the result with the company’s own guidance as well as with consensus.
  5. Read the cause of any gap in the release and the filing.
  6. Watch the direction of revisions after the report.

FAQ

Is the consensus estimate an official number?

No. It is a statistic compiled by data vendors from analysts’ published or submitted forecasts. Different vendors can report slightly different consensus figures for the same company and period.

What is the difference between guidance and an analyst estimate?

Guidance comes from the company and reflects management’s outlook. An estimate comes from an analyst and reflects the analyst’s own model. Analysts often use guidance as an input but may set estimates above or below it.

Why do companies sometimes “beat” and still see a drop in their share price?

The beat compares results with consensus, while the market also weighs guidance, margins, the quality of the beat and prior expectations such as unofficial whisper numbers. The price reaction reflects all of those, and this guide does not predict any of them.

Can a company influence analyst estimates?

Companies can shape estimates through public guidance and communications. Regulation FD prohibits selectively disclosing material nonpublic information to covered analysts and investors without making it public. Firm-level rules also aim to prevent research from being improperly influenced.

Are analyst ratings and price targets part of the estimate?

They are separate from the earnings forecast. A rating or target is an analyst’s opinion on a security. This publication does not provide ratings, targets or recommendations.

Where can I check an analyst’s compliance with the rules?

Research reports of broker-dealers carry disclosures and certifications required by FINRA rules and Regulation AC. The rule text is available on FINRA.org and the Regulation AC text on the eCFR.

Prepared with AI assistance from public sources and reviewed under our editorial policy. Not investment advice.

Related