How a drug company is actually valued, and why ordinary start-up methods don't fit

Why the usual methods break

Most start-up valuation methods lean on one of two things: revenue (today's, or a forecast, times a market multiple) or a comparison with typical start-ups at the same stage. A company developing a new medicine has neither in a useful form. It usually has no product revenue until a regulator approves its drug, and the path from the first trial in people to approval averaged 10.5 years between 2011 and 2020: 2.3 years in Phase I, 3.6 in Phase II, 3.3 in Phase III and 1.3 at the regulator.

For that whole decade, a stage ladder that runs idea → prototype → pre-revenue → early revenue files the company as "pre-revenue", next to a software company a few months from launch. But its value isn't in the company's trajectory. It sits in one or a few assets (the drug candidates), and each one is a bet that either pays or goes to zero.

The odds, phase by phase

The best public measure comes from BIO (the US biotechnology industry body), Informa Pharma Intelligence and QLS Advisors, who tracked 12,728 moves by drug candidates from one trial phase to the next between 2011 and 2020. Of the candidates that entered Phase I (the first small safety trial in people), 7.9% went on to be approved by the US regulator, the FDA. The share that made it through each step:

• Phase I → Phase II: 52.0%
• Phase II → Phase III: 28.9%. The lowest of the four, and where most value dies: Phase II is usually the first trial designed to show the drug works in patients.
• Phase III → filing for approval: 57.8%
• Filing → approval: 90.6%

The odds depend heavily on the disease

That 7.9% is an average across very different fields. From Phase I, drugs for blood disorders (haematology) reached approval 23.9% of the time, cancer drugs 5.3%, and urology drugs 3.6%. Rare-disease programmes reached 17.0%. Apply the headline average to a cancer drug and you overstate its odds by about half; apply it to a blood-disorder drug and you understate them threefold. A credible valuation uses the row for the disease, and says which row it used.

The method: risk-adjusted NPV

Draw the drug's life as a timeline of cash: what each trial phase costs, when it could launch, how sales would build to a peak, and when patent protection ends and cheaper generic copies arrive. Then multiply each year's cash by the chance the drug is still alive in that year. Discount the result back to today and add it up. That is risk-adjusted net present value (rNPV), and it is the common language of drug licensing deals and pharma acquisitions.

A shape, not a forecast: a drug entering Phase II has, on the report's figures, a 15.1% chance of approval. Its future sales count at about one-seventh of their face value. Its Phase II trial costs count in full, because the company spends them whether or not the drug works. That asymmetry is why many early assets come out with a negative rNPV, and a negative answer is a finding, not a mistake.

Three traps in a founder's own model

• Counting the risk twice. Start-up methods often discount at 30% a year or more, because the rate carries the risk that the company fails. rNPV carries that risk in the phase probabilities instead, so it uses a far lower rate. Using a start-up rate AND phase probabilities counts the same risk twice and makes every asset look worthless.
• Peak sales decide the answer. In almost every rNPV model the peak-sales figure moves the result more than anything else. Work out the peak sales the asset needs to be worth anything at all (the break-even), then read your own forecast against that number.
• The odds are averages. They describe thousands of programmes, mostly run by larger companies, not your asset. Whether a small company's odds differ is an open question. Say which figures you used and why.

What actually sets the price

A drug developer's value is often realised before approval, through a licensing deal or a sale to a larger company. Those deals are usually built from an upfront payment, milestone payments as the drug clears each phase, and a royalty on eventual sales, and each piece is priced on the same probability-weighted logic. That is why investors in the sector think in rNPV, and why a valuation that ignores the phase odds won't survive their first question.

Why our valuation tool stops

vc-scan's valuation tool runs five methods built for companies valued on revenue and stage. Pointed at a drug developer they would produce a number that computes, looks reasonable, and is wrong. So when you choose "Therapeutics (drug development)", it stops. Medical devices, diagnostics and health software are valued on revenue and stay in the tool, under "Digital health / medtech".

Instead it sends you to our drug-asset calculator, which does what this post describes for one asset at a time: the published odds for your disease area, your costs and sales, and the peak sales the asset needs to be worth anything. This is an explainer, not investment advice.