Reference
Questions, terms and method
Questions about how Long Range Advisory works, every abbreviation used on this site and in a model, and the conventions behind the numbers.
Answers
About Long Range Advisory
What does Long Range Advisory do?
Long Range Advisory provides long-range planning, valuation modeling and executive decision support to biotechnology and pharmaceutical organizations — building the DCF, NPV, IRR and scenario models leadership uses to allocate capital.
Who is Long Range Advisory for?
Executive teams at emerging biotech and pharma companies: CFOs, VPs of Finance and CEOs facing a pipeline, licensing, manufacturing or commercial investment decision that needs a defensible valuation behind it.
What is the difference between adjusted and unadjusted NPV?
Unadjusted NPV assumes technical success. Adjusted NPV risk-weights every cash flow by the probability of success. In the worked example on this site the same pipeline is worth $5.54B unadjusted and $4.08B adjusted — a $1.46B risk discount.
How much does an engagement cost?
Scope and terms are agreed in writing before any work begins, and sized to the decision. Initial conversations are complimentary and focused on understanding your objectives before recommending next steps.
Methodology
The conventions behind every number
These are not preferences. Each one is enforced by a test that recomputes on every change, and a model that breaks one cannot be exported.
What method is used to value a pipeline?
Risk-adjusted net present value, summed over the parts. Each programme is forecast separately from its eligible population through to unlevered free cash flow, each year is multiplied by the probability of reaching the market and discounted to a single base date, and the results are added. Two portfolio adjustments are then argued separately: comorbidity overlap, for the same patient counted in more than one indication, and correlated failure, for one molecule failing across several indications at once.
Where does risk sit in the calculation?
In the probability of success, applied to the cash flows — once, to every year, and nowhere else. Clinical failure is idiosyncratic and diversifiable, so it does not belong in a discount rate that compensates for systematic risk. A 30% discount rate on top of a 7% probability prices the same failure twice, and is the most common error in a biotech model.
Is there a terminal value?
No. The forecast runs to loss of exclusivity and through an erosion tail to a residual. Perpetual growth on a revenue stream with a contractual end date would be the first thing an investor attacked.
What is the forecast horizon, and how are years handled?
Twenty years, on one calendar clock shared by every programme, discounted to 31 December of a single base year. Adding six programmes each on their own launch-relative clock is a real and easily missed error. A mid-year convention assumes cash arrives evenly through the year and is worth three to five per cent of present value; it can be switched off.
How are losses and tax treated?
Losses accumulate and shelter later profit. Cash tax never goes negative — a pre-revenue company has no prior taxable profit to carry back against, so a loss is not a refund.
Where do the assumptions come from?
From you. Published benchmarks sit beside the inputs with a publisher and a date against each figure, and a value outside the published range is flagged rather than rejected — but nothing in the library feeds the arithmetic. The model has no house view; if it did, changing an assumption would mean arguing with a citation.
How is a model checked?
Thirteen integrity tests recompute on every change: the value bridges have to foot, the per-share line has to reconcile to the equity value, probability has to be applied exactly once, no terminal value may have crept in, and the sensitivity grid’s live cell has to reproduce the headline figure. A failing check blocks the export rather than letting a bad number reach a deck.
Can the numbers be audited outside the platform?
Yes, and that is the point of the Excel export. It contains live formulas, not pasted values: every figure on the calculation sheets points back at an input, so any number can be traced to the assumption behind it and changed in the workbook. The export is verified against the platform’s own engine before release.
Glossary
Every abbreviation on this site
Standard definitions, grouped by where they appear. If a term is used anywhere in a Long Range Advisory model and is not defined here, that is an omission worth reporting.
Valuation
The measures a board argues about, and the difference between them.
- NPV
- The sum of a project’s future cash flows, each discounted back to today. It answers one question: is this worth more than it costs, in money valued at the same date.
- rNPV
- An NPV in which each year’s cash flow is multiplied by the probability the programme reaches the market. It is the standard measure for a drug in development, because a plain NPV values a compound as though approval were certain.
- DCF
- The general method behind both of the above: forecast the cash, then discount it for time and risk. A DCF is a structure, not a number — its answer is only as good as the assumptions put into it.
- IRR
- The discount rate at which a project’s NPV is zero, read as an implied annual return. It must be computed on undiscounted cash flows: running it over already-discounted figures is a common and silent error.
- WACC
- The blended cost of a company’s debt and equity, used as the discount rate. Clinical risk belongs in the probability of success, not here — loading both prices the same failure twice.
- Terminal value
- A lump sum standing in for all cash flows beyond the forecast. Long Range Advisory models do not use one: a pharmaceutical revenue stream has a contractual end date, and perpetual growth on it is the first thing an investor attacks.
- Sum of the parts
- Valuing each programme separately and adding the results, rather than modelling the company as one revenue line. It makes the concentration of value visible, which is usually the finding.
- Unlevered free cash flow
- Cash generated by the business before financing, after tax, capital expenditure and movements in working capital. Not the same as net income, and it is the figure that reaches a valuation.
Development and regulatory
What has to happen before a compound earns anything.
- PoS
- The cumulative chance a programme reaches the market from where it stands today. Published transition data puts a Phase 1 asset near 8% and a Phase 3 asset near 50%.
- Phase 1 / 2 / 3
- The clinical stages. Phase 1 tests safety in a small group, Phase 2 tests dose and a first efficacy signal, Phase 3 confirms it at scale. Each is a gate a programme can fail at.
- NDA / BLA
- The submissions that ask the FDA to approve a small molecule and a biologic respectively. "Filed" means submitted, not approved.
- CRL
- The FDA declining to approve a submission as it stands. About 47% of first-cycle failures are eventually approved, with a median delay of well over a year.
- LOE
- The point at which a product loses its protection and competitors enter. A small molecule facing generics can lose half its volume in the first year; a biologic facing biosimilars erodes far more slowly.
- Orphan designation
- A regulatory status for treatments of rare conditions, carrying incentives including exemption from Medicare price negotiation for products whose approved indications are all orphan.
- PRV
- A transferable voucher for expedited FDA review, awarded on certain approvals and commonly sold. Treated as one-off other income, not revenue.
- CRO
- A company that runs clinical trials on a sponsor’s behalf — site selection, monitoring, data management.
- CMC
- The manufacturing and quality work a submission has to document. A real and often underestimated share of the spend between a positive trial and a first sale.
Commercial
How an approval turns into money, and what stands in the way.
- Eligible population
- People diagnosed, treated, and eligible under the label expected — not everyone with the condition. The single figure nobody else can supply for you.
- Peak penetration
- The share of that eligible population reached in the best year, across every competitor, payer restriction and undiagnosed patient.
- Gross-to-net
- The gap between list price and the money that actually arrives, after rebates, chargebacks, discounts and fees. It runs from roughly 30% to 70% of gross across the industry, so net price is well below any published list price.
- COGS
- The direct cost of making each course of treatment. Low for a small molecule, materially higher for a biologic, and higher again for an autologous cell therapy.
- S&M / G&A
- The two operating cost blocks after cost of goods. S&M scales with revenue; G&A is largely fixed and runs at roughly 30–41% of R&D for a clinical-stage company.
- FP&A
- The finance function that builds the budget, the forecast and the long-range plan, and explains the variance between them.
- LRP
- The multi-year plan a board approves, typically five to ten years out and reconciling to the valuation rather than sitting beside it.
- Delivery capacity
- The ceiling on how many patients can physically be treated — certified sites, apheresis and manufacturing slots, isotope supply. Where it binds, it rather than demand sets the revenue line.
- MFP
- The price a selected drug is sold at to Medicare following negotiation under the Inflation Reduction Act, from nine years after approval for a small molecule and thirteen for a biologic.
Statistics and modelling
The terms that appear on the distribution and sensitivity screens.
- Monte Carlo simulation
- Running a model thousands of times with assumptions drawn from a range, to produce a distribution of outcomes instead of a single figure.
- P10 / P50 / P90
- Percentiles. P50 is the median — half of outcomes fall below it. P10 and P90 bracket the middle eight in ten, which is the range worth planning against.
- PERT distribution
- A three-point distribution taking a low, likely and high estimate, weighting the likely case four times the extremes. Preferred to a triangular distribution, which weights all three equally and overstates the tails.
- Tornado chart
- One bar per assumption, running from its low case to its high case around the base, sorted by width. It shows which assumption matters and whether its downside is worse than its upside.
- S-curve
- For any figure on the horizontal axis, the probability of ending at or below it. The chart to use when the audience should choose their own threshold.
- Sensitivity analysis
- Changing one assumption at a time, everything else held still, to see how far the answer moves. Two bars on a tornado are separate experiments and do not add together.
Long Range Advisory. Illustrative outputs based on user-entered assumptions. Not investment, securities, accounting, tax, or legal advice. Read the full legal disclaimer.