Launch Excellence in Pharma: Building the Commercial Model
A pharmaceutical launch concentrates several years of investment into a window of a few months, and the organisation running it usually gets one attempt. That combination attracts a large advisory literature, most of which opens by asserting how often launches fail and then offers a framework for avoiding it. The assertion deserves more scrutiny than it usually receives, and the framework deserves less deference — because the same body of research contains a finding that is both better supported and more useful.
Missing an analyst forecast is not the same as failing. A profitable, clinically valuable launch that fell short of a bullish consensus counts as a miss, which means the statistic measures forecast optimism as much as commercial performance — and the cohort it describes launched before most of the access architecture that now governs launch existed.
The useful finding is in the same research and is quoted far less. Early trajectory persists: 78% of launches lagging forecast in year one continued to lag in year two, and 70% of those lagged again in year three, while 65% of year-one outperformers kept outperforming. That is what justifies disproportionate investment before launch rather than after it — and it means budget reserved for a year-two corrective push is being reserved for the moment it will work least well. This guide covers that evidence, the eight decisions a commercial model consists of, and the 2026 access dates that now set the launch calendar.
What the launch failure statistic actually says
Almost every piece of launch excellence content opens with a version of the same number — that most launches fail. The figure circulates as 56%, 60%, 67% and two-thirds depending on the source, which is the first sign that it deserves inspection rather than repetition.
| What is claimed | What the research actually found | Why the difference matters |
|---|---|---|
| "Most drug launches fail" | About two-thirds of new drugs failed to meet pre-launch consensus sales expectations for their first year | Missing an analyst forecast is not failing. A launch can be profitable, clinically important and commercially sound while missing a bullish consensus — and the statistic counts it as a miss |
| An industry-wide, current figure | A sample of 210 new molecular entities launched between 2003 and 2009, measured against consensus forecasts made one year before launch | That cohort launched before the current access environment existed — before the EU joint assessment, before Medicare negotiation, and before most of the payer architecture that now governs launch |
| Multiple independent confirmations | The circulating variants — 56%, 60%, 67% — largely trace back to this work or to re-analyses of it | Repetition is not replication. Four numbers from one lineage is one number, and quoting a specific variant implies a precision the evidence base does not have |
| A statement about execution quality | It is jointly a statement about forecast optimism, since the benchmark is a prediction made by people with an interest in the asset | Half the gap may sit on the forecasting side of the equation, which is a different problem with a different owner |
Why this matters more than a pedantic correction
A launch team told that two-thirds of launches fail draws a different conclusion from one told that two-thirds miss an optimistic forecast made a year out.
The first framing invites fatalism, or an expensive scramble to do everything at once. The second points at something actionable: the forecast itself is part of the system being managed, and a launch judged against a number nobody stress-tested is set up to be reported as a failure regardless of how it performs.
The practical instruction is to interrogate the forecast with the same rigour as the plan. What penetration does it assume, over what period, against which competitors, with what access position and at what price? A launch plan built to hit a number whose assumptions were never examined is not a plan — it is a commitment to someone else's arithmetic.
The finding that does survive: early trajectory persists
Strip out the headline and one result in the same research is both robust in shape and considerably more useful. It is the reason launch deserves disproportionate investment, and it is quoted far less often than the number that does not support that conclusion.
| Observation | Figure | What follows from it |
|---|---|---|
| Launches lagging forecast in year one | 78% continued to lag in year two | A slow start is not a slow start. It is the beginning of a pattern, and the assumption that year two provides a correction opportunity is not supported |
| Of those still lagging in year two | 70% lagged again in year three | Recovery becomes less likely as time passes, not more — the opposite of how most recovery plans are argued internally |
| Launches exceeding forecast in year one | 65% continued outperforming in year two | Momentum is real in both directions. The asymmetry is that the downside persists slightly more strongly than the upside |
The commercial consequence is specific and it changes how a budget should be phased. If trajectory is largely set in the first year and rarely reverses, then spending held back for a year-two corrective push is being reserved for a moment at which it will be least effective. The conventional launch curve — build slowly, evaluate, then invest behind evidence of traction — is precisely inverted relative to what the persistence data implies.
Two honest qualifications keep this from becoming an argument for spending everything at once. The finding is a pattern across a cohort, not a law governing any individual product — some launches do recover, usually because something material changed such as a label expansion or a competitor withdrawal. And persistence is partly a statement about the underlying asset: a product that underperforms because it is undifferentiated will keep underperforming, and no phasing of promotional spend corrects that. The finding argues for front-loading effort on launches that deserve it, not for front-loading everything.
What a pharma commercial model actually consists of
The phrase is used loosely enough to mean almost anything. In practice it is a set of eight decisions, most of which are difficult to reverse after launch, and which is why they belong in the pre-launch period where the trajectory evidence says the value sits.
| # | Decision | Why it is hard to reverse | When it must be settled |
|---|---|---|---|
| 1 | Which customers the model serves — prescriber, institution, payer, patient, or a combination | It determines the entire field structure. A model built for prescribers cannot be repointed at institutions without rehiring | Pre-launch, and earliest of all |
| 2 | Access position and pricing | Price anchors on entry and moves down far more easily than up. In assessed markets the position is fixed by a process with its own calendar | Well before launch — the assessment timeline governs it |
| 3 | Channel mix and field size | Headcount decisions carry notice periods, hiring lead times and territory design | T-12 months at the latest, because recruitment and training precede launch |
| 4 | Evidence plan beyond the label | Real-world and health-economic evidence takes years, and the need for it is known before launch | Pre-launch. Evidence commissioned after a gap appears arrives too late to change trajectory |
| 5 | Which geographies, in what order | Sequencing interacts with reference pricing and assessment timetables across markets | Pre-launch, and it is a portfolio decision rather than a brand one |
| 6 | Medical and commercial split of responsibility | It shapes hiring, compliance boundaries and who holds key relationships | Pre-launch, and frequently left ambiguous |
| 7 | Measurement design and the control comparison | A baseline cannot be established retrospectively. Without it, year-one performance cannot be attributed to anything | Before the first activity, or not at all |
| 8 | The forecast and its assumptions | It becomes the standard everything is judged against, and it is rarely revisited once agreed | Pre-launch, and it should be stress-tested rather than accepted |
The two decisions most often deferred are the two that cannot be
Measurement design and forecast interrogation are the items that get postponed, because neither produces anything visible and both are uncomfortable.
Measurement is deferred because there is nothing to measure yet — which is exactly why it must be designed now. A baseline and a control comparison established before the first activity is the only way year-one performance can be attributed to the launch rather than to the market. After launch, the counterfactual is gone permanently.
The forecast is deferred because challenging it is politically awkward at the moment everyone needs the number approved. But the trajectory evidence means the organisation will be judged against that number within twelve months, on a basis that will not reverse.
Both cost analyst time and no capital, and both are the cheapest items in the launch budget.
The 2026 access environment: dates that change the plan
Most launch excellence frameworks in circulation predate the current access architecture. Two developments in particular now govern launch planning before any commercial decision is made, and both carry hard dates.
Europe — the Joint Clinical Assessment
| Element | Detail | Launch planning consequence |
|---|---|---|
| Operational from | 12 January 2025 | Already live. A launch plan that treats European assessment as a country-by-country sequence is describing the previous system |
| Current scope | Oncology medicines and advanced therapy medicinal products | Applies now to the two categories with the most launch activity |
| January 2028 | Extends to rare disease medicines | Orphan portfolios need the capability built before this date, not at it |
| 2030 | All medicinal products | Eventually universal. Building the capability once, early, is cheaper than building it under deadline |
| Geographic reach | The EU plus Norway, Switzerland, Ukraine and the UK | Wider than the EU itself, which affects sequencing assumptions |
| The operational pressure | A 90-day window between release of the PICO framework and dossier submission | The single most demanding element. Ninety days is not enough time to generate evidence — only to assemble it, which means it must already exist |
| First-year throughput | 10 joint clinical assessments and seven joint scientific consultations in 2025 | Modest against the pipeline. Capacity is a real constraint and early slots are scarce |
| Organisational shift | Responsibility moved from local country teams to global market access and health economics functions | A structural change to who owns launch evidence, and a common source of internal confusion |
The first-year experience also identified recurring evidence problems: risk of bias in single-blinded trials, concerns about non-European trial locations, and inadequate handling of missing data. All three are trial design decisions taken years before launch, which is the point — the assessment reaches back into development, and a launch team inheriting a dossier cannot fix them.
The United States — the negotiation clock
| Product type | Eligible for selection | Negotiated price effective | Strategic consequence |
|---|---|---|---|
| Small molecule drugs | 7 years after FDA approval | Year 9 | A nine-year commercial runway before price control. It compresses the payback period and changes indication sequencing |
| Biologics | 11 years after FDA approval | Year 13 | Four additional years, and the source of the disparity |
| The gap | — | Four years | Known as the 'pill penalty'. It is a portfolio-level input, affecting which modality is developed as much as how a launch is run |
Reform has been proposed rather than enacted. Executive Order 14273 of 15 April 2025 directed the Department of Health and Human Services to work with Congress on aligning the treatment of small molecules with biologics, and the EPIC Act (H.R. 1492) would equalise both at eleven years. Claims about the consequences — including an estimate of 188 fewer small molecule medicines reaching market and a reported 70% fall in small molecule funding since 2021 — come from advocacy and policy sources with a position, and should be attributed rather than repeated as settled findings.
What both of these mean for a launch team specifically
Access is no longer a workstream that runs alongside the commercial plan. It sets the calendar the commercial plan has to fit inside.
In Europe, the ninety-day window between PICO and dossier means the evidence must exist before the assessment starts — which pushes the decisive work back into development. In the United States, the negotiation clock means the commercial runway is a known finite number from the day of approval, and it differs by modality.
The practical instruction is that the launch plan should be built backwards from the access calendar rather than forwards from the approval date. Those are different plans, and only one of them survives contact with the process.
Verify both sets of dates before relying on them. The European scope expands on schedule, but the American position is under active legislative and executive pressure and may not hold
Launch archetypes and the trap attached to each
The same research that produced the trajectory finding also identified four launch situations, each with a characteristic failure. The value of the framework is that it names the trap a team is most likely to walk into given its own situation.
| Archetype | The situation | The characteristic trap |
|---|---|---|
| Go for gold | A strong product with genuinely good data in an established market | The good data trap — assuming product quality guarantees uptake, and under-investing in removing the practical barriers to prescribing. The most dangerous archetype precisely because confidence is justified |
| Stand out from the crowd | A credible product entering a crowded, competitive market | Failing to differentiate because the insight work was not done. Differentiation is discovered, not asserted, and crowded markets punish generic positioning fastest |
| Category creator | A product establishing a new treatment category | Under-investing in establishing the unmet need itself. The market has to be built before the product can be sold into it, and that work is usually under-resourced because it does not look like selling |
| Market shaper | An undifferentiated product in an unestablished disease area | Both problems at once. The hardest position, and the one where an honest pre-launch assessment most often argues for a smaller commitment than the plan assumes |
The framework is useful diagnostically rather than prescriptively. Its practical value is in the conversation it forces: a team that cannot agree which archetype it is in has not agreed what the launch is for, and that disagreement will surface later as conflicting decisions about spend, evidence and messaging.
India: launching under price control
Almost none of the above transfers directly. Indian launch economics are shaped by a different constraint, and the constraint is unusually explicit.
For any formulation on the National List of Essential Medicines, price is not a launch variable. The 2026 revision under the Drug Price Control Order covered 907 scheduled formulations with a wholesale price index adjustment of approximately 0.65%, effective 1 April 2026. A ceiling price applies, quarterly returns are filed, and the annual movement is inflation-linked and small. Launch is therefore a volume, coverage and speed problem rather than a pricing one — and a commercial model imported from a market where price is the primary lever will optimise something that cannot move.
| Constraint | What it means | What the commercial model must do instead |
|---|---|---|
| Ceiling pricing on scheduled formulations | Price set by regulation, revised annually by a WPI adjustment that was about 0.65% for the latest revision | Compete on availability, coverage and prescriber relationship. Margin comes from volume and cost, not from price realisation |
| Brand proliferation | Multiple brands of the same molecule from different companies competing simultaneously | Speed to coverage is the differentiator. The window between launch and competitive saturation is short and it does not reopen |
| Field deployment as the launch mechanism | The launch largely is the field plan — territory coverage, doctor lists and call frequency from day one | Recruitment and territory design are on the critical path, months before the commercial launch date |
| Trade and distribution reach | Product availability at the point of prescription determines whether a prescription converts | Stocking coverage before promotional activity, not alongside it. A prescription that cannot be filled trains the prescriber not to write it again |
The trajectory finding transfers to India even though the pricing framework does not — arguably with more force. In a market where several brands of the same molecule launch into the same prescriber base, the prescriber's first choice becomes habitual quickly, and a brand that is unavailable or unpromoted in the opening months is competing against an established habit rather than against an alternative. The persistence of early position is the one part of the international evidence that translates without adjustment.
A worked example: a pre-launch readiness assessment
Readiness is usually assessed against a checklist of deliverables. A more useful assessment asks which decisions are settled, which are reversible, and how much time remains to change them. The figures are illustrative.
| Decision | Status at T-9 months | Reversible? | Assessment |
|---|---|---|---|
| Customer definition | Settled — prescriber-led with institutional overlay | No | Fine. Settled early, as it must be |
| Access position | Assessment dossier in preparation; PICO not yet received | No — externally governed | The critical path item. Ninety days from PICO to submission means the evidence must already be assembled |
| Field size and structure | Proposed, not approved | Partially — hiring lead time is roughly four months | Approaching the point of no return. At T-9 this is the last quarter in which it can change without slipping the launch |
| Evidence plan | Label-based only; no health-economic package | No — generation takes years | The gap that cannot be closed. Plan around it honestly rather than commissioning work that will arrive after trajectory is set |
| Measurement design | Not started | Yes — but only until the first activity | The cheapest available fix and it closes permanently at launch. Highest-value action this quarter |
| Forecast assumptions | Agreed twelve months ago; not revisited since | Yes, and politically difficult | Stress-test now. The organisation will be judged against this number within a year, on a basis that will not reverse |
What this assessment produces that a checklist does not
A ranked list of what can still be changed, rather than a list of what is incomplete.
On a conventional readiness checklist, the health-economic evidence gap would be flagged red and generate an urgent commissioning request. The assessment shows that work cannot arrive in time to affect trajectory, so the honest response is to plan around it and start it for the next indication rather than to spend into a closed window.
Meanwhile measurement design — which no checklist treats as urgent because it produces no deliverable — is the highest-value item on the page, because it is free, it is reversible only until the first activity, and without it the launch cannot later prove what worked.
The general rule: at T-9 months, prioritise by reversibility rather than by size of gap. The largest gaps are frequently the ones nothing can now be done about, and treating them as urgent consumes the attention that the closing windows need.
Where launches fail
- Planning forwards from approval rather than backwards from the access calendar. In assessed markets the process sets the timetable, and a ninety-day evidence window cannot be met by work started when the window opens.
- Reserving budget for a year-two corrective push. The persistence evidence says year two is when correction is least likely to work. Money held back for the recovery moment is held back for the wrong moment.
- Accepting the forecast without interrogating it. The organisation will be judged against a number set a year out by people with an interest in the asset. Half the reported launch gap may sit on the forecasting side.
- Skipping measurement design because there is nothing to measure yet. The counterfactual disappears at first activity and cannot be reconstructed. This is the only item on the list that is free and permanently closing.
- The good data trap. Strong evidence creates confidence that the product will sell itself, and the practical barriers to prescribing go unaddressed. It is the characteristic failure of the strongest launches, which is what makes it dangerous.
- Importing a price-led model into a price-controlled market. Where a ceiling price applies and annual revision is inflation-linked at under one percent, a model optimising price realisation is optimising a variable that cannot move.
- Leaving the medical and commercial split ambiguous. It determines hiring, compliance boundaries and relationship ownership, and resolving it after launch means renegotiating all three under time pressure.
A launch sequence, working backwards
Anchored to the access calendar rather than to the approval date, because that is the constraint that actually binds.
- T-24 months and earlier — settle customer definition and the evidence plan. These reach back into development and cannot be added later. If the health-economic package is not commissioned here, it will not exist when it is needed. Confirm trial design against the assessment criteria that have proved difficult: blinding, trial location and missing-data handling.
- T-18 months — build the access position and the geographic sequence. Assessment timelines, reference pricing interactions and market order. In Europe the joint assessment now governs this, and the capability sits with global market access rather than with country teams — confirm who owns it before the calendar starts moving.
- T-12 months — approve field size, structure and territory design, and stress-test the forecast. Recruitment and training lead times mean this is effectively the last point at which structure can change. Interrogate the forecast in the same meeting, because both decisions rest on the same assumptions and reviewing them separately hides the inconsistencies.
- T-6 months — design the measurement, including the comparison. Baseline, control geographies or matched territories, and the agreed read-out date. This is the cheapest item in the plan and the only one that closes permanently at first activity. Agree the read-out with finance now, while it is a design question rather than a defence.
- T-0 to T+6 — execute at full intensity and monitor against the baseline, not against the forecast. The trajectory evidence means this window carries disproportionate weight. Monitor for early divergence and be willing to act inside it, because the correction opportunity is now rather than at the year-one review.
One expectation to set with the launch team. Three of these five steps produce nothing anyone can present, which is why they are the ones that slip. Evidence planning, forecast interrogation and measurement design have no visible output and no launch-day moment — and between them they determine whether the visible work in the final six months lands on a trajectory that persists.
Key takeaways
The seven actions this article argues for, separated from the evidence that supports them.
- Plan backwards from the access calendar, not forwards from approval. In assessed markets the process sets the timetable, and a 90-day evidence window cannot be met by work started when the window opens.
- Front-load the effort rather than reserving it. Early trajectory persists, so budget held back for a year-two corrective push is held back for the moment it will work least well.
- Interrogate the forecast in the same meeting as the field structure. Both rest on the same assumptions, and reviewing them separately hides the inconsistencies.
- Design the measurement before the first activity. The counterfactual disappears at launch and cannot be reconstructed — this is the only free item in the plan and the only one that closes permanently.
- Watch for the good data trap. Strong evidence creates confidence that the product will sell itself, and it is the characteristic failure of the strongest launches.
- Prioritise by reversibility at six to nine months out, not by size of gap. The largest gaps are often the ones nothing can now be done about.
- Adapt the model for price-controlled markets. Where a ceiling price applies and annual revision runs under one percent, a strategy optimising price realisation is optimising a variable that cannot move.
Conclusion
Launch excellence is usually framed as doing more, faster, in the months around approval. The evidence points somewhere less energetic and more demanding: the decisions that determine a launch are largely made before anyone is watching, and several of them close permanently long before launch day. Customer definition and the evidence plan reach back into development. Field structure closes at roughly twelve months. The access calendar belongs to someone else entirely.
The finding that justifies all of this is not the failure statistic everyone quotes but the persistence one almost nobody does. Launches that lag in year one overwhelmingly continue to lag, and the probability of recovery falls rather than rises as time passes. That is what makes the pre-launch period disproportionately valuable, and it is also what makes the conventional launch curve — build slowly, evaluate, then invest behind evidence of traction — exactly the wrong shape.
The reordering worth taking away is a change of question. Most launch planning asks what has to be ready by launch day. The better question is which decisions stop being changeable, and when — because measurement design, the cheapest item in the entire budget, closes the moment the first activity happens, and it is the one deadline nobody writes down.
Frequently Asked Questions For Launch Excellence in Pharma: Building the Commercial Model
The discipline of settling the commercial model before launch rather than iterating it afterwards, because early trajectory persists. It covers eight decisions — customer definition, access and pricing, channel mix and field size, the evidence plan, geographic sequencing, the medical and commercial split, measurement design, and the forecast and its assumptions — most of which are difficult or impossible to reverse once launch has happened. The reason it matters is not that most launches fail, which is a weaker claim than usually presented, but that launches which start slowly rarely recover: in the most-cited research, 78% of launches lagging forecast in year one continued to lag in year two, and 70% of those lagged again in year three.
Not as the statistic is usually stated. The most-cited research examined 210 new molecular entities launched between 2003 and 2009 and found that about two-thirds failed to meet pre-launch consensus sales expectations for their first year. Two qualifications matter. Missing an analyst forecast is not the same as failing — a profitable, clinically valuable launch that missed a bullish consensus counts as a miss, so the figure measures forecast optimism as much as commercial performance. And the cohort launched before the current access environment existed, predating the EU joint assessment and Medicare price negotiation. The variants in circulation — 56%, 60%, 67% — largely trace back to this work or re-analyses of it, so repetition across sources is not independent confirmation.
Sometimes, but the base rate is against it and it does not improve with time. In the cited research, 78% of launches that lagged forecast in year one continued lagging in year two, and 70% of those lagged again in year three — meaning the probability of recovery falls as the launch progresses rather than rising. Outperformance was similarly persistent, with 65% of year-one outperformers continuing in year two. Two qualifications keep this honest: it describes a cohort pattern rather than a law governing any single product, and recovery, when it happens, usually follows something material such as a label expansion or a competitor withdrawal rather than increased promotional effort. The practical implication is that budget reserved for a year-two corrective push is reserved for the moment it will work least well.
A single clinical assessment conducted at European level under the Health Technology Assessment Regulation, which became operational on 12 January 2025. Its current scope is oncology medicines and advanced therapy medicinal products; it extends to rare disease medicines in January 2028 and to all medicinal products in 2030, and it applies across the EU plus Norway, Switzerland, Ukraine and the UK. The most demanding operational feature is a 90-day window between release of the PICO framework and dossier submission — enough time to assemble evidence but not to generate it. Only 10 assessments and seven joint scientific consultations were completed in the first year, so capacity is a genuine constraint, and responsibility has shifted from local country teams to global market access and health economics functions.
It puts a known clock on the US commercial runway from the day of approval, and the clock differs by modality. Small molecule drugs become eligible for Medicare price negotiation selection 7 years after FDA approval, with negotiated prices effective at year 9. Biologics become eligible at 11 years, with prices effective at year 13 — a four-year structural difference known as the pill penalty. The practical consequence is that payback periods and indication sequencing must be planned against a fixed horizon rather than an open-ended one. Reform has been proposed but not enacted: Executive Order 14273 of 15 April 2025 directed work with Congress on alignment, and the EPIC Act would equalise both at eleven years. Verify the current position before relying on it, because this is under active legislative pressure.
The set of decisions determining how a product reaches and is paid for by its customers: which customers the model serves, the access and pricing position, channel mix and field size, the evidence plan beyond the label, geographic sequencing, the medical and commercial split of responsibility, measurement design, and the forecast it will be judged against. What distinguishes it from a marketing plan is reversibility — most of these decisions have long lead times, structural consequences or external calendars attached, and cannot be adjusted once launch is underway. The two most often deferred are measurement design and forecast interrogation, both of which cost analyst time rather than capital and are therefore the cheapest items in the launch budget.
The binding constraint is price control rather than price strategy. For any formulation on the National List of Essential Medicines, a ceiling price applies — the 2026 revision under the Drug Price Control Order covered 907 scheduled formulations with a wholesale price index adjustment of approximately 0.65%, effective 1 April 2026, with quarterly returns filed to the regulator. Price is therefore not a launch variable, and launch becomes a volume, coverage and speed problem. Three consequences follow: brand proliferation means several brands of the same molecule compete simultaneously, so speed to coverage is the differentiator; field recruitment and territory design sit on the critical path months before launch; and stocking coverage must precede promotion, because a prescription that cannot be filled teaches the prescriber not to write it again.
Prioritise by reversibility rather than by size of gap, which is the opposite of what a conventional readiness checklist produces. At that point some decisions are already fixed — customer definition, the evidence plan, and anything governed by an external assessment calendar — and flagging them as urgent consumes attention that closing windows need. Field structure is typically in its final quarter of adjustability because recruitment and training lead times run to several months. Measurement design is usually the highest-value action available: it costs nothing, it is the only item that closes permanently at first activity, and without it year-one performance cannot be attributed to the launch rather than to the market. Forecast interrogation belongs in the same period, because the organisation will be judged against that number within twelve months on a basis that will not reverse.
Let's Discuss Your Requirements
