How to underwrite a market-share loser: size the pond before counting the fish, sort the causes into fixable and structural, and check whether the derating happened to the price or to the earnings.
1. Separate share of a market from size of a market before judging a share loss
The repeatable method
- Size the total addressable market in units of customers, not dollars — how many people have the condition, how many are currently served.
- Establish the market's growth rate independently of the company's.
- Multiply: falling share × faster-growing market can still mean rising revenue. Check whether it does, in the reported numbers.
- Only treat share loss as thesis-breaking once the market's growth stops covering it — and state in advance what penetration level that happens at.
Here: "934 million people worldwide have obesity. Only 2.2 million are currently treated with branded medication. This means over 97% of the market is still untapped. Demand for obesity drugs is growing at more than 100% (!) per year." Against GLP-1 share falling 59% → 50% and obesity share 74% → 53%, the conclusion is checked against the accounts: "even with less market share, Novo's revenue is still hitting record highs."
Watch for
- The argument outliving its condition. It works while penetration is 0.2%; it stops working as the market matures, and no trigger level is named here.
- Share of new patients versus share of the installed base. The first is the leading indicator, and it is the one Lilly is winning.
2. Sort the causes of a setback into fixable, fixed, and structural
The repeatable method
- List every cause separately rather than accepting a single narrative.
- Label each: already fixed (with the evidence and date), fixable with capital or time, or structural.
- Size how much of the damage each caused, so the structural share of the problem is explicit.
- Underwrite only the structural one — the rest is timing.
Here: three causes. Capacity, now fixed and dated — DKK 47.2bn of factory investment plus "three manufacturing sites from Catalent for $11 billion", and "as of early 2025, the FDA declared the shortage resolved." Compounding pharmacies, now closing — "over 130 lawsuits" and "the FDA has also ruled that most compounding is illegal again", though "up to 30% of patients" had switched. A better rival drug, structural — "20% weight loss with Zepbound compared to 14% for Novo's Wegovy." Two of three are behind; the third is the actual bet.
Watch for
- Fixed problems being counted twice — once as damage done, once as future recovery. Recovery requires patients to switch back, which the piece does not evidence.
- A structural disadvantage answered with pipeline breadth rather than with a better product. That is a real strategy and a slower one.
3. Decompose a derating: did the multiple fall, or did the earnings?
The repeatable method
- Chart the forward multiple and EPS on the same timeline from a common start date.
- Quantify each leg in percentage terms, so the source of the share-price move is unambiguous.
- A multiple falling against rising earnings is a sentiment problem; both falling is a business problem.
- Cross-check with a peer's multiple on the same date to see how much of the gap is company-specific.
Here: "Since 2020: the Forward PE declined from 20.9x to 12.9x (−38%); EPS rose from 9 DKK to 23.5 DKK (+161%)," described as "the entire investment case in one image." The peer check: LLY at 34.1x against NVO at 16.9x next year's earnings — "you can buy more than two shares of Novo Nordisk for 1 Eli Lilly."
Watch for
- Figures struck on different dates inside one article. The general-information block here is stamped 29 July 2025 while the catalyst section is January 2026 — which is why the intro says "3x as cheap" and the conclusion says "more than two times".
- A cheap multiple on forward earnings that are about to be cut. The same price concessions listed as catalysts are also margin concessions.
The repeatable method
- Before running a reverse DCF, check whether free cash flow is currently depressed by growth capital expenditure.
- If it is, substitute earnings for free cash flow and state the substitution explicitly rather than silently.
- Solve for the growth rate the current price requires at your required return.
- Compare that hurdle with the company's realised growth and with the end-market's growth — not with the analysts' forecast.
Here: "
We use EPS instead of FCF because Novo Nordisk is investing heavily in future growth (high CAPEX). Our Reverse DCF shows Novo Nordisk only needs to grow its EPS by
2.9% per year to return 10% per year to shareholders. This looks very reasonable to me. The market doesn't seem to have a lot of expectations from Novo Nordisk right now." Compare the FCF-based versions used for
LeMaitre (16.1% required) and across the Buy-Hold-Sell sheets.
Watch for
- Switching the input in the direction that flatters the answer. The test is whether the same substitution would be made on a name you wanted to reject.
- A hurdle so low it stops discriminating. At 2.9%, the reverse DCF passes almost anything — it is a floor test, not a valuation.
5. Separate the priced business from the unpriced optionality, and do not pay for the second
The repeatable method
- Identify the indications, geographies or products the current price plausibly assumes.
- List the ones that are clinically or commercially live but not in anyone's model, with their evidence stage.
- Check the valuation still works with all of them valued at zero.
- Treat each as a separate catalyst to monitor, with its own read-out date.
Here: the base case is obesity and diabetes at 16.9x. On top, unpriced: "Semaglutide is showing promise for heart failure, kidney disease, and even alcohol addiction. Each of these is a multi-billion dollar opportunity on its own"; a licensed triple agonist at 24% weight loss after 48 weeks; and emerging-market volume — "China, India, and Africa with local production… by 2030, they could serve millions of patients." The reverse DCF's 2.9% hurdle is what makes it possible to value all of that at nothing and still own the shares.
Watch for
- Volume optionality that arrives at lower prices. "Even at lower prices, the volume creates a massive advantage" is asserted, and no margin bridge is shown.
- Pipeline breadth counted as strength when the rival has depth. Several second-place drugs are not equal to one leading one.
6. Find the party that actually controls the customer, and check which way they leaned
The repeatable method
- Trace the purchase decision to whoever really makes it — a formulary committee, a benefit manager, a regulator, a procurement office.
- Track that party's published decisions as the leading indicator, ahead of prescription data.
- Note the price that was paid for the decision, because access is usually bought with margin.
- Watch for the political layer above it, which can reset the terms.
Here: the US fight is settled by intermediaries, not patients — "In May 2025, CVS Caremark (a large pharmacy benefit manager) made Wegovy its preferred drug over Lilly's Zepbound" — alongside direct-to-consumer cash pricing "as low as $349 per month" and "a deal with the White House to lower prices in exchange for broader access to Medicare and Medicaid starting in 2026." Each is a volume win bought with price.
Watch for
- Formulary wins that reverse. A preferred listing is a contract, not a moat.
- The political attack that produced the concession (a $1,300 US price against under $200 in Europe) recurring in another jurisdiction.
Methods distilled from the archived Compounding Quality post for personal study. The underlying investment case is the work of Steven Van Den Burg. Not investment advice.