TL;DR: A sell-side analyst rating is an opinion—typically some flavor of Buy, Hold or Sell—paired with a 12-month price target derived from a valuation model. Rating scales are not standardized across Wall Street firms; a “Neutral” at one bank is a “Market Perform” at another and an “Equal-Weight” at a third. What they have in common is regulation: FINRA Rule 2241, SEC Regulation Fair Disclosure and post-2003 reforms require analysts to disclose conflicts and to publish the same information to everyone at once. The most reliable signal is not any one analyst’s target but the consensus—the aggregated opinion of every firm covering the stock, and, more importantly, how that consensus is changing.
What a rating and a price target really are
Every trading day, brokerage firms publish research reports on companies they cover. Each report contains two headline outputs: a rating (Buy / Hold / Sell or the firm’s equivalent) and a price target (what the analyst thinks the stock should be worth 12 months from now). A change in either is what makes the news wire — “Goldman Sachs upgrades Toast to Buy, raises price target,” “JinkoSolar fair value cut after analyst target changes.”
The rating is an ordinal opinion: it tells you where this stock sits in the analyst’s ranked coverage list. The price target is a numerical anchor: it’s the output of a valuation model, plus a discount rate, minus the current price. The rating usually falls out of the price target: if the target is more than ~10–15% above the current price, the analyst writes Buy; if it’s within a few percent, Hold; if it’s meaningfully below, Sell.
Sell-side vs. buy-side (and why you only see one of them)
There are two distinct populations of analysts, and only one publishes.
- Sell-side analysts work at brokerages and investment banks (Goldman Sachs, Morgan Stanley, JPMorgan, BofA, Barclays, UBS, Wells Fargo, Citi, etc.). Their research is distributed to institutional clients and then quickly reaches the press. They’re paid based on the quality and reach of their research and the trading commissions and banking business it helps generate. In the U.S., sell-side analysts must register with FINRA and pass the Series 86/87 Research Analyst Examination.
- Buy-side analysts work at asset managers, hedge funds and pension funds (Fidelity, BlackRock, Capital Group, Bridgewater, Millennium, and so on). Their research is proprietary: it never leaves the building. They’re paid on performance, not output, so their incentive is to keep good ideas quiet, not to broadcast them.
When you read “Wall Street analysts say” in a headline, that is always the sell-side. The buy-side is silent on purpose. That’s the first thing to internalize: the ratings you see are the ones professional investors pay to see through, not the ones they act on privately.
How the price target actually gets built
Under the hood, a target price comes from one of a small handful of standard valuation methods, usually blended:
- Discounted cash flow (DCF). The analyst forecasts free cash flow for ~5–10 years, adds a terminal value, and discounts everything back to present value using a risk-adjusted rate. Widely considered the most theoretically sound, but very sensitive to assumptions about growth and the discount rate.
- Comparable multiples. The analyst applies a P/E, EV/EBITDA, EV/Sales or P/B multiple — benchmarked against peer companies or the stock’s own history — to a forward earnings estimate. Fastest to build, easiest to defend, easiest to be wrong on when the “comps” break down.
- Sum-of-parts (SOTP). For conglomerates or multi-segment companies, value each segment separately (retail on P/E, cloud on EV/Sales, hardware on EV/EBITDA), add net cash, subtract debt. This is why Amazon, Alphabet and Meta targets can vary so wildly across firms — different analysts value AWS or YouTube on different multiples.
- Dividend discount / residual income. Mostly used for financials and mature dividend-payers.
The point is not that any one method is right. It’s that the target is model output, and different assumptions produce different numbers. When Morgan Stanley says $500 and Wells Fargo says $360 on the same stock, they are running different DCFs with different growth assumptions, not disagreeing about the facts.
Rating scales are not standardized — here’s the map
Every firm defines its own rating language. FINRA Rule 2241 requires firms to clearly define what each rating means in every research report, and to disclose what percentage of their coverage sits in each bucket. That’s how you can decode terms like “Overweight” and “Outperform.” Here’s how the biggest sell-side houses translate the same three ideas — better than the market, roughly like the market, worse than the market:
| Firm | “Buy” equivalent | “Hold” equivalent | “Sell” equivalent |
|---|---|---|---|
| Goldman Sachs | Buy (on Conviction List → higher-conviction Buy) | Neutral | Sell |
| Morgan Stanley | Overweight | Equal-Weight | Underweight |
| JPMorgan | Overweight | Neutral | Underweight |
| Bank of America / Merrill | Buy | Neutral | Underperform |
| Barclays | Overweight | Equal Weight | Underweight |
| UBS | Buy | Neutral | Sell |
| Wells Fargo | Overweight | Equal Weight | Underweight |
| Citigroup | Buy | Neutral | Sell |
Two practical takeaways from the table:
- “Overweight” and “Underweight” come from portfolio construction: they mean the stock should sit at more (or less) than its benchmark index weight in your portfolio, not that its price will rise or fall.
- Sell ratings are structurally rare. Sell-side firms cover companies they hope to do banking business with, and while Rule 2241 requires them to disclose the Sell percentage of their coverage, that percentage typically runs in the single digits.
The pipeline: from valuation model to the tape
Here’s what actually happens between an analyst’s spreadsheet and a headline crossing your screen.
The regulatory guardrails, quickly
- FINRA Rule 2241 (equity research): requires disclosure of the analyst’s personal holdings, the firm’s 1%+ ownership, whether the firm makes a market in the stock, and any investment-banking relationship in the past 12 months. Compensation cannot be tied to a specific banking transaction. Firms must also enforce a 10-day quiet period after an IPO if they were an underwriter.
- FINRA Rule 2242 (debt research): parallel framework for bond analysts, with information-barrier requirements between debt research and banking / trading.
- SEC Regulation Analyst Certification (Reg AC): every analyst must personally certify that the views expressed in the report reflect their honest opinion and that their compensation is not tied to the specific recommendation.
- SEC Regulation FD (2000): companies cannot selectively disclose material non-public information to favored analysts. Everyone gets it at the same time — which is why quarterly earnings calls are now public webcasts.
These rules exist because the post-dot-com enforcement actions of the early 2000s—the Jack Grubman and Henry Blodget cases in particular—demonstrated what happens when analysts privately mock stocks they publicly rate Buy. The Global Analyst Research Settlement that followed forced structural separation between research and investment banking at the ten largest firms.
Why consensus beats any single call
The single most useful thing a retail investor can do with analyst data is stop obsessing over one bank’s price target and start watching the consensus. The consensus is the mean (or median) of every covering analyst’s estimate, aggregated by data services like LSEG’s I/B/E/S (which tracks over 40,000 companies across 70 markets from 900+ contributing firms), FactSet, Bloomberg or Zacks.
The reason is simple: when you average many independent forecasts, individual errors partially cancel out. A large body of forecasting research has found that the equal-weighted average of independent forecasts is usually more accurate than any single forecaster picked in advance — a result that holds up across macro, weather, and, yes, corporate earnings.
Even more useful than the level of the consensus is the direction. When the consensus target is being revised up and multiple firms are moving in the same direction inside of a few weeks, that’s a meaningful signal — it usually reflects new information about earnings, guidance, or the industry. When the consensus is flat and only one firm moves, that’s noise.
Common mistakes retail investors make
- Treating the price target as a forecast, not a fair value. A 12-month target is what the analyst thinks the stock should be worth using their model. It is not a prediction of where the stock will trade — realized returns depend on flows, sentiment, macro, and every earnings print between now and then.
- Reading a rating change without reading the report. A downgrade from Buy to Hold with a price target still 15% above the current price is very different from a downgrade to Sell with the target below the current price. The headline flattens that.
- Ignoring the timeframe. Almost every rating and target is 12 months. Nothing in a research report is a signal about tomorrow, next week or next quarter.
- Confusing “Overweight” with “will go up.” Overweight means “own more of this than your benchmark says,” which is a portfolio-construction instruction, not a directional prediction.
- Chasing single-analyst calls over the consensus. One firm’s outlier is a story; the consensus is the anchor. If Wedbush has a $500 target and everyone else is at $320, the market is trading closer to $320 for a reason.
Related concepts and what to learn next
- How to read an earnings report — the primary input for every model that produces a rating.
- Whisper numbers and post-earnings drift — the empirical evidence that revisions matter more than levels.
- The P/E ratio, when it misleads — the most common valuation multiple that shows up in every target model.
- WACC — the discount rate in the DCF that produces the target.
Sources
- FINRA Rule 2241 — Research Analysts and Research Reports
- FINRA Rule 2242 — Debt Research Analysts and Debt Research Reports
- SEC Regulation FD (Fair Disclosure)
- Financial analyst — buy-side vs sell-side, Series 86/87
- Buy-side analyst — incentives and visibility
- Sell-side analyst — regulatory history
- Stock valuation methods (DCF, multiples, SOTP)
- Consensus forecast — why averaging beats picking one
- Institutional Brokers’ Estimate System (I/B/E/S)
Disclosure: This article is for informational purposes only and is not investment advice.