Why Analyst Price Targets Disagree—and What Revisions Tell You
Distinguish changed opinions from changed coverage when reading analyst estimates.
General learningIntermediate4 min
Estimates, not promises
Plain English
An average hides the views inside it
An analyst price target is an estimate built from a method and assumptions, not a promise about a future price. Analysts may disagree about growth, profitability, discount rates or valuation multiples. They may also use different horizons or information dates. An average becomes interpretable only after checking whether the observations being combined are sufficiently comparable.
Our introductory target-price article explains the basic reading rules. Here the focus is what disagreement and revisions can reveal about the collection of estimates. Dispersion describes differences between opinions; it is not automatically a statistical measure of future price risk. A tight group of targets can reflect shared assumptions, including assumptions that later prove wrong.
Worked example
Hold the panel and horizon constant
Use three fictional analyst notes, A, B and C, with targets of £80, £100 and £120. They are issued on 10 August 2026 for the same twelve-month horizon ending 10 August 2027, using the same currency and share basis. Their arithmetic mean is £100 and the high-minus-low range is £40. The reference share price is held at £90 solely to isolate the target calculation.
Later on the same hypothetical issue date, analyst A corrects its model and revises £80 to £95, explicitly retaining that horizon. B and C are unchanged. The new mean is (95 + 100 + 120) / 3 = £105, and the range is 120 − 95 = £25. All three contributors remain in the panel. This is a deliberately controlled revision example; ordinary updates weeks later can roll the horizon forward.
The mean-to-price gap changes from 100 / 90 − 1, approximately 11.11%, to 105 / 90 − 1, approximately 16.67%. Those are arithmetic comparisons, rounded to two decimal places using unrounded inputs. They are not expected returns, probabilities, total-return estimates or recommendations. The mean contains no stated probability weights or distribution assumptions.
Calculation
(95 + 100 + 120) / 3 = 105
Targets and their mean are in pounds per share, with the same currency, share basis and fixed horizon.
Original mean target
£100
Revised mean target
£105
Original target range
£40
Revised target range
£25
Original mean-to-price gap
11.11%
Revised mean-to-price gap
16.67%
Reading the Revision
Ask what caused the aggregate to change
The worked result tells us that one lower estimate moved closer to the others. It does not establish that all analysts became more optimistic, or that uncertainty about the business fell. Read A’s explanation: changed earnings inputs, a corrected model and a new valuation multiple would be different reasons. A number alone cannot identify which reason applies.
Now consider a separate coverage example using the original targets. If the £80 analyst simply disappears, the remaining £100 and £120 average £110 without either analyst revising anything. Calling that an upward revision would confuse membership change with changed opinion. Preserve contributor identities and distinguish unavailable data from an analyst explicitly ending coverage.
Real comparisons must also check stale estimates, corporate actions, currency and horizon. A provider’s current average may combine notes written at different times. Record what is known rather than assuming every target refers to the same future date. The reference price can move too; a changing percentage gap might reflect that denominator rather than any change in targets.
Methods and Conflicts
Read method and conflicts alongside the numbers
For covered US research reports, FINRA Rule 2241 addresses the basis for price targets, valuation methods, risks and conflict disclosures. These are US member-firm requirements, not a guarantee of accuracy or a rule governing every estimate worldwide. Their educational relevance is that a target needs explanatory context, including interests that may affect how research is produced.
A consensus collection is not necessarily representative of all informed views. Coverage can be sparse, correlated or selected by a provider. Neither an expanding nor a narrowing range supplies a trading rule. The useful conclusion names the change that can be established and the information still needed to interpret it.
Common mistake
Do not read an average as a vote on future returns
A common mistake is treating every higher mean as broad agreement that shares will rise. Check the matched panel first, then the reasons, dates and assumptions. In this exercise one revision increases the mean; in the alternative example one removal does. Identical-looking aggregate movements can therefore describe very different research events.
Self-check
Check your understanding
Did all three analysts raise their targets?
No. Only A changed, from £80 to £95. B and C remained at £100 and £120.
What happens if the original £80 contributor disappears?
The remaining mean becomes £110 without a revision by either remaining analyst. It is a coverage-composition change.
Does a 16.67% gap imply that expected return?
No. It compares a mean target with a fixed price and provides neither outcome probabilities nor a total-return model.
Optional workflow context
A connection to Strata Value
Strata Value brings analyst expectations, fundamental quality and momentum into one view. These perspectives can disagree; the review method above also works with your own notes.
This article is for information and financial education only. Examples are hypothetical and are not personalised investment advice or recommendations to buy, sell or hold a security.
