Biotech Analyst Coverage: Read the Revision, Not the Number

Every biotech headline seems to quote an analyst. “$220 at BofA.” “Initiated at Buy.” “Citi slashes target to $35.” And almost every retail investor reads that number as a forecast: the stock will hit $220. Here’s the problem. That isn’t what a price target is, and in biotech specifically, treating it like a forecast is how you lose money.

The fix is simple: stop reading the number and start reading the revision. The single most useful thing an analyst tells you isn’t the target. It’s the change in the target.

Step 1: Know what a price target actually is

A price target is an analyst’s estimate of where a stock will trade in about 12 months. It’s built one of three ways: a discounted cash flow (DCF) of future profits, a comparable-company multiple (what similar biotechs trade for), or a risk-adjusted peak-sales model (how much the drug could sell, times the odds of approval, times a 3-8x multiple). None of these are forecasts. They are snapshots of one analyst’s assumptions on the day they hit publish.

In biotech, single-point targets are nearly meaningless because the risk is binary. A drug either gets FDA approval or it doesn’t. If a target assumes approval and the FDA issues a CRL (complete response letter, a rejection), the target doesn’t drop 10%. It vaporizes. This is why you see a $120 target on a stock trading at $15. The analyst is modeling the approval scenario. The market is pricing the coin flip. Both are rational. Neither is a prediction.

Step 2: Decode the rating language

A rating and a price target are different things. Buy/Hold/Sell is the plain version. Overweight/Neutral/Underweight means the same thing, measured against a benchmark. But the skew matters. A genuine “Sell” is rare, because it burns the analyst’s relationship with the company’s management and its bankers. In practice, “Hold” often means “Sell, politely.” A “Hold” with a price target 30% below the current price is a sell signal wearing a nicer name. Read the target against the current price, not just the word.

Step 3: Read the revision, not the number

The initiation, the upgrade, the downgrade, the target raise or cut: these are the signal. A $100 target that was $100 last month tells you nothing new. A target cut from $100 to $60 means the analyst just learned something that changed their model, and that’s the information. You can watch the revision feed in real time on MarketWatch’s upgrades and downgrades page.

Watch the pre-catalyst and post-catalyst cycles. Analysts rarely move targets into an FDA decision, because nobody wants to be wrong on a binary event. Then they pile on afterward. That revision cluster after a readout is the market digesting the result in real time, and it usually tells you more than the headline number did.

Step 4: Understand why consensus is always late in biotech

Consensus estimates and targets are systematically slow around FDA decisions and trial readouts. The reason is structural: analysts update models on schedules, not on news, and a binary catalyst doesn’t fit a quarterly model. By the time the average target moves, the stock has already moved. This is why trading the catalyst itself (see how to trade FDA catalysts) is a different game from holding and waiting for the revision to catch up. Consensus is a lagging indicator here. Treat it that way. The consensus target you see on a stock page is just an average of every covering analyst, which makes it the slowest-moving number of all. Track the individual revisions, not the average, because the average lags even the laggards.

Step 5: Check the conflicts

Most coverage has incentives you should know about. If the analyst’s bank underwrote the company’s IPO or follow-on offering, or is advising on a merger, the coverage isn’t purely thesis-driven. There’s also the quiet period: in the weeks around an IPO, underwriter analysts can’t publish, which is why coverage “initiates” with a splash right after. And the initiate-to-curry-favor dynamic is real: a bank initiates coverage on a biotech partly hoping to win future banking business. None of this makes a target wrong. It makes it worth knowing who’s talking.

Step 6: Triangulate, don’t outsource

The right way to use a price target is as one input, not the thesis. Put it next to your own valuation (see how to value a biotech in three methods) and your own read of the last earnings call (see the five numbers that matter). Then watch the spread. When targets are tightly clustered, the market agrees. When they’re wildly dispersed, say $10 to $38 on the same stock, the dispersion itself is the signal: the binary event is unresolved and nobody actually knows.

Here’s the pattern in practice. Sarepta (SRPT) is the Duchenne muscular dystrophy company behind Elevidys, a gene therapy approved in 2023. When two patients died of liver failure in 2025 and the FDA paused distribution, the analyst targets told the story faster than the stock did. Targets that had been $120 got cut to $70, then to $36. One analyst went all the way to a $10 target with an Underperform rating, while another held a Buy at $38. Peak-sales estimates for Elevidys dropped from $2.4 billion to $1.7 billion. The absolute number at any single moment was close to useless. The revision sequence, a steady one-way slide, was the actual forecast. And the $28 gap between the bull and the bear was the honest read on the risk.

Common mistakes

Buying a stock because the target implies 200% upside. The target assumes the best case, and you just paid for it in advance.

Reading the target but not the revision. A stale high target is worse than no target at all.

Treating consensus as a forecast. By the time the average moves, the move is over.

Trusting a target from the lead underwriter without asking what they’re hoping to win.

Final checklist

• What is the rating, and what does it mean next to the current price?
• Has the target changed recently, and in which direction?
• Is there a binary catalyst between now and the target date?
• Who’s the analyst, and does their bank have a relationship with the company?
• How wide is the dispersion between the highest and lowest target?
• Have you run your own valuation, or are you outsourcing the thesis?