How to Read an Analyst Report
A price target is arithmetic, not a forecast
① A price target is arithmetic, not prophecy. Estimate × multiple — and the analyst picks both, so moving the target means moving one of them.
② Across 525 stocks, not one carries a Sell consensus (0.00%). Yet open the individual reports and Sell ratings run 3.0% — the average erases the dissent.
③ Which means the target and the rating are the least informative parts of a report. The signal is in how the estimates changed, which assumptions were used, and what the minority view is looking at.
📐 A price target is arithmetic, not a forecast
Read a price target as "this stock will reach this price" and you will misread the whole report. In practice the number comes out of a very simple multiplication.
Price target = forecast earnings × applied multiple
Estimate next year's earnings per share at $10, apply a P/E of 30, and the target is $300. If metrics like EPS and P/E are unfamiliar, start with core concepts explained and the 50 essential US stock terms.
An important consequence follows. There are only two ways to move a price target — change the estimate, or change the multiple. And the analyst chooses both.
Changing an estimate requires an explanation — new orders, improving margins, something concrete. But a change in multiple often passes in a single line: "in line with sector average," "applying the three-year mean." When the earnings outlook is unchanged and only the target moved, this is usually what happened. Separating estimate moves from multiple moves makes what a report is actually claiming far clearer.
Take a live example. As of August 25, 2026, Micron (MU) carries a consensus target of $1,515 against a $933 share price — implied upside of +62%. Do not read that as "it will rise 62%." It means "multiply the estimates and multiple these analysts selected and you get 1,515." If the estimates are wrong, or the market assigns a different multiple, the number simply changes.
📊 What the data shows — where are the Sell ratings?
Instead of generalities, the actual numbers. Below is the analyst consensus for the 525 US stocks marketbrief tracks, as of August 25, 2026.
| Consensus rating | Stocks | Share |
|---|---|---|
| Buy | 337 | 64.2% |
| Hold | 90 | 17.1% |
| Strong Buy | 64 | 12.2% |
| No rating | 34 | 6.5% |
| Sell family | 0 | 0.00% |
Buy-family ratings account for 76.4%, Hold for 17.1%, and not a single one of the 525 carries a Sell. Implied upside points the same direction.
| Price vs target (521 stocks) | Value |
|---|---|
| Average implied upside | +15.0% |
| Median | +12.2% |
| 25th / 75th percentile | +6% / +22% |
| Trading above target | 46 stocks (8.8%) |
In short, nine of every ten stocks sit below their target price. That isn't an anomaly of one moment — it's closer to a structural property of the metric. Targets are typically set on a twelve-month view, so sitting above the current price is natural, and when a stock reaches its target, the target is frequently raised.
If the market-wide average is +15%, then a stock with 15% upside is, in effect, average. The number only carries meaning as a position within this distribution: +6% is bottom quartile, +22% is top quartile.
🔍 Consensus erases the dissent
From the table above it's tempting to conclude analysts simply never write Sell. Open the individual reports and the story changes.
Counting 2,618 individual firm reports across those same 525 stocks by rating:
| Report-level rating | Count | Share |
|---|---|---|
| Buy | 669 | 25.6% |
| Overweight | 462 | 17.6% |
| Neutral | 422 | 16.1% |
| Outperform | 392 | 15.0% |
| Equal-Weight / Hold family | 356 | 13.6% |
| Underweight · Underperform · Sell | 78 | 3.0% |
At the report level, Sell ratings exist at 3.0%. Not many — but not zero. And yet the consensus ratings for these same stocks contained no Sells at all.
The reason is simple: consensus is an average. If twenty analysts cover a stock and one says Sell while nineteen say Buy, the average is Buy. Minority views vanish the moment they are converted into a number.
"Consensus Buy" does not mean nobody said sell. It means the average landed on the buy side. Whether there was a dissenter, and what that person was worried about, leaves no trace in the average. That is why the consensus figure is not a place to stop reading.
A live case makes this concrete. Tesla's consensus target is $390 — as a single number, merely "a bit above the $350 share price." Open the individual reports:
| Date | Firm | Rating | Target |
|---|---|---|---|
| 2026-08-18 | GLJ Research | Sell | $24.86 |
| 2026-08-03 | Stifel | Buy | $491 |
| 2026-07-24 | Piper Sandler | Overweight | $450 |
| 2026-07-23 | JP Morgan | Neutral | $445 |
| 2026-07-23 | UBS | Neutral | $385 |
Same month, same company, and the targets are $24.86 and $491 — a twentyfold gap. Both are current opinions, not stale leftovers. The $390 average covers that violent disagreement without a trace.
Apple looks similar. On August 10 Jefferies published Underperform with a $263.66 target; on July 31 TD Cowen published Buy at $400. The current $309.90 price sits precisely between the two.
📏 Same stock, different targets — the spread is the signal
To check whether Tesla is a freak case, I recalculated the spread between firms across 495 stocks using only targets published within the last 90 days (stale targets inflate apparent disagreement).
| High-to-low target spread | Value |
|---|---|
| Median | 20.5% |
| Average | 29.3% |
| 25th / 75th percentile | 13% / 32% |
| Top decile | 48% or wider |
For more than half of all stocks, firm targets differ by over 20%. That is why treating a price target as a precise figure makes little sense. And the spread itself is worth reading.
| Stock | Target range (5 firms) | Spread | How to read it |
|---|---|---|---|
| NVDA | $325 – $352 | 8% | Narrative agreed; the debate is over pace |
| AMD | $550 – $700 | 27% | Same direction, differing magnitude (all five bullish) |
| AAPL | $264 – $400 | 52% | The growth story itself is contested |
| MU | $1,150 – $2,000 | 74% | Disagreement over where the cycle peaks |
| TSLA | $25 – $491 | 1,875% | They are analyzing different businesses |
A narrow spread like Nvidia's means the market has broadly agreed on how to value the company. A wide one like Tesla's means analysts are closer to analyzing different companies — one sees a carmaker, another an autonomy and robotics platform. The wider the spread, the less the average target means and the more it matters which narrative each writer assumed. Those competing narratives in semis are covered in the AI chip three-way and the semiconductor sector outlook.
Super Micro's (SMCI) target list still shows $1,000. The stock trades at $38. That target was published in September 2024, and all four other entries beside it are from 2024 as well. Without checking when a number was published, you mistake a two-year-old view for today's outlook.
In fairness, this is rare. Of the 525 stocks, only 12 (2.3%) have no report newer than 90 days, and the median age of the latest report is 14 days. Most are fresh — which is exactly why filtering out the stale minority matters.
💡 Where the real information sits
Invert everything above and the conclusion appears. The price target and the rating are the least informative parts of a report. Sells are 3%, implied upside clusters near +12%, and targets differ by more than 20% between firms. Three numbers that decide very little.
The information is elsewhere.
| Component | Information density | Why |
|---|---|---|
| Estimate revisions | Highest | Shows what the analyst actually learned. Both direction and size are signal. |
| Assumptions | High | Revenue growth, margins, applied multiple. Agreeing with these is agreeing with the target. |
| The minority view | High | Contradicting consensus demands clearer reasoning. More to read. |
| Risk section | Medium | Often boilerplate — but newly added items are worth the time. |
| Price target | Low | Arithmetic on estimate × multiple. 20% spread. |
| Rating | Lowest | Sells at 3%. Effectively close to binary. |
Estimate revisions deserve special attention. Price targets tend to trail the share price and move late, whereas a revision reflects what the analyst newly verified through management meetings, channel checks, or an earnings print. "Target raised" says far less than "FY revenue estimate raised 8%." How to read the prints themselves is covered in the big-tech earnings guide.
⏱️ In practice — reading a report in five minutes
There is no need to read forty pages front to back. With a fixed order, most reports yield their core in about five minutes.
| Step | What to check | Test |
|---|---|---|
| 1 | Date and firm | Older than 90 days, treat as reference only. Note initiation vs update. |
| 2 | Whether the rating changed | Reiteration isn't news. Only upgrades, downgrades, and initiations are. |
| 3 | Size of estimate revisions | By what % did revenue and EPS estimates move? This beats the target change. |
| 4 | How the target was derived | Estimate up, or multiple up? The latter often rests on thinner ground. |
| 5 | Position vs consensus | Far above or below the mean? The reason for that gap is the report's real content. |
| 6 | Whether dissent exists | If any Sell or Neutral exists, read that one first. |
Having AI summarize a long PDF is useful, with two cautions. First, conditionals disappear in summarization — "may happen under certain conditions" compresses easily into "will happen." Second, verify numbers in the source, not the summary. Prompts that separate fact from interpretation are in writing with AI and making AI read your documents; why models are shaky with figures is in why AI gets numbers wrong.
✅ Checklist, and the limits of this analysis
Six questions to ask whenever a report lands in front of you.
- When was it published? — start with the date.
- What changed? — rating held and target held means almost no new information.
- Estimate or multiple? — the real driver behind any target move.
- Where does it sit versus consensus? — near the mean is not a distinctive claim.
- What does the other side say? — go find the 3% the consensus erased.
- Does my decision rest on this report? — if so, it rests on a single source.
This data describes a structure; it does not assess anyone's integrity. There are well-known structural reasons Sell ratings are rare — the practice of quietly dropping coverage rather than publishing a negative view, access to company IR, relationships with the investment-banking side, and the plain fact that equity markets have risen over long horizons. The point is not "they can't be trusted" but "read them knowing this distribution." Once you know +15% is the average, that number stops being impressive.
A report is raw material, not a conclusion. A price target is someone else's assumptions multiplied together; a consensus is an average with the dissent removed. Both are starting points, not destinations. To check how far your own portfolio leans on a single narrative, use the portfolio check; for sector-level concentration, see the sector rotation piece.
※ Written August 27, 2026. Statistics are aggregated from analyst consensus data as of 2026-08-25 for the 525 US stocks marketbrief tracks. The 2,618 individual reports represent up to the five most recent per stock — a sample, not the full universe of published research, so the true share of Sell ratings may differ. Rating labels (Buy, Overweight, Outperform, and so on) follow different scales at different firms, which limits direct comparison. Stocks named here illustrate the data and are not buy or sell recommendations. Investment decisions and their consequences rest with the investor.
※ This report is provided for informational and educational purposes only and does not constitute a recommendation to buy or sell any security.
New research, when it lands
Subscribe and the next deep dive comes to you, along with the daily market brief — free, unsubscribe anytime.