Deciding how much of the R&D budget chases novel-mechanism breakthroughs versus extending and defending the value of products already on the market
Before a company can debate how much it should spend on breakthrough science versus defending its existing franchises, it first needs a consistent, auditable way to sort every active program into one of two buckets. This classification exercise sounds administrative, but the definitions chosen determine everything downstream — the reported split, the incentive structure of R&D leadership, and ultimately how the market perceives the durability of the pipeline.
The Innovation bucket is reserved for programs pursuing genuinely novel biology or engineering: a first-in-class mechanism of action, a new therapeutic modality the company has not previously commercialized (e.g. a small-molecule company's first cell-therapy or RNA program), or a target with no validated precedent in the clinic. The defining feature is not scientific ambition alone but genuine novelty risk — the program could fail not because execution went wrong, but because the underlying biological hypothesis itself turns out to be incorrect.
This bucket also typically includes platform-building investment — discovery-stage tooling, target-identification infrastructure, and early translational science — that has no near-term product attached to it but is intended to seed the innovation pipeline several cycles out. Because these programs sit far from an approved, revenue-generating asset, their value is almost entirely optionality: a call option on a scientific hypothesis paying off.
A useful classification test: if the program failed completely, would competitors reasonably say "the science just didn't work" rather than "the execution was poor"? If yes, it belongs in Innovation — its risk is scientific, not operational.
The Lifecycle Management (LCM) bucket captures investment that extends, defends, or deepens the value of an asset whose core mechanism is already clinically and commercially validated. Common LCM program types include:
• New indications — testing an approved drug in an adjacent disease where the same mechanism plausibly works, leveraging existing safety data to shorten the clinical path • New formulations — extended-release, subcutaneous-from-IV conversion, fixed-dose combinations that improve convenience, adherence, or differentiate against biosimilar/generic entrants • Combination products — pairing the marketed asset with another agent (own or partnered) to expand its addressable use • Pediatric and geographic extensions — trials required or incentivized to extend exclusivity or open new regulatory territories
The defining feature of LCM work is that the core mechanistic and safety risk has already been substantially retired by the original approval — what remains is largely execution risk: can the trial be run well, will payers reimburse the new indication, will physicians adopt the improved formulation.
In real portfolio reviews, a meaningful fraction of programs sit uncomfortably close to the boundary. A combination product pairing a marketed asset with a genuinely novel investigational molecule is arguably both — LCM from the marketed drug's perspective, Innovation from the new molecule's perspective. A next-generation formulation using a materially new delivery platform (e.g. a novel long-acting depot technology) may carry real technical risk even though the drug substance itself is proven.
Because the bucket a program is assigned to affects budget visibility, career incentives for the team running it, and how the split is reported to the board and investors, companies with mature portfolio-governance processes assign this classification to a cross-functional portfolio committee rather than letting individual program teams self-classify — self-classification predictably drifts toward whichever bucket currently enjoys more favorable budget treatment or executive attention.
Once every dollar is classified, the resulting Innovation:LCM ratio is only meaningful in context. The same 40:60 split that looks dangerously innovation-starved for an early-stage biotech might look aggressively risk-tolerant for a diversified large-cap — because the two archetypes are solving structurally different problems with structurally different constraints.
A diversified large-cap pharmaceutical company typically has a substantial base of marketed revenue to protect — often concentrated in a handful of products approaching loss-of-exclusivity. For that company, LCM investment is not merely opportunistic; it is a direct, relatively low-risk lever for defending near-term cash flow that funds everything else, including the Innovation bucket itself. A large-cap company under-investing in LCM is effectively choosing to let defensible, already-de-risked revenue erode faster than necessary.
An emerging, pre-commercial biotech company faces the opposite structural reality: it typically has no marketed base to defend, so there is little for an LCM bucket to act on. Nearly all of its R&D spend is, by definition, Innovation-bucket spend — pursuing the novel mechanism that is its entire reason for existing. The comparison is not really "biotech is bolder than pharma" so much as "biotech has nothing yet to extend, and pharma has something worth protecting."
Comparing a single company's split to an "industry average" without adjusting for portfolio maturity is one of the most common analytical errors in portfolio strategy reviews — the right benchmark is a band conditioned on how much marketed revenue the company already has exposed to loss-of-exclusivity risk.
A single company's own historical split is rarely static — it shifts predictably as its portfolio matures:
• Early stage: the split is almost entirely Innovation, since there is no marketed base yet • Post-first-approval: LCM spend appears for the first time, funding label expansions and formulation work for the newly approved asset • Multi-product maturity: as more assets reach the market and approach patent expiry in overlapping windows, LCM's share of the budget typically rises, sometimes substantially, as multiple defense programs run concurrently • Post-major-loss-of-exclusivity: after a large asset genericizes, LCM spend on that franchise drops sharply (there is often little left worth defending), which can mechanically swing the reported split back toward Innovation even without any deliberate strategic shift
Because of this last effect, a rising Innovation share in the reported numbers does not always signal a deliberate strategic pivot — it can simply reflect a large LCM program rolling off after its target asset lost exclusivity.
Comparative benchmarking is most useful as a sanity check, not a target to hit mechanically. Boards and portfolio committees typically use industry-pattern comparisons to ask two questions: is our LCM investment sufficient to responsibly defend the revenue we are structurally exposed to losing, and is our Innovation investment sufficient to replace that revenue on a timeline consistent with when it will actually erode?
A company that mechanically targets "the industry average split" without reference to its own patent-cliff exposure or pipeline maturity risks either under-defending near-term revenue it can least afford to lose, or over-defending a franchise that was never going to be worth the incremental spend, at the direct expense of the innovation pipeline that determines the company's value a decade out.
Innovation and Lifecycle Management investment are not simply "riskier" and "safer" versions of the same activity — they sit on structurally different probability distributions, and comparing them on a single blended return metric obscures the strategic tradeoff a portfolio committee actually needs to see.
The Innovation bucket's expected-value distribution is best described as wide, flat, and long-tailed: most individual programs return little or nothing (the underlying mechanistic hypothesis simply does not pan out), but the distribution has a long right tail representing the rare program that becomes a genuinely new standard of care, or opens an entirely new addressable market the company did not previously compete in. The probability of success for a single first-in-class program from early discovery through approval is low by design — that is the price of pursuing something no one has validated in the clinic before.
The LCM bucket's distribution looks essentially the opposite: tall, narrow, and concentrated near the present. Because the core mechanism, safety profile, and often much of the manufacturing and regulatory pathway are already established, individual LCM programs succeed at a much higher rate and read out much faster. But the payoff for any single program is capped — a new indication or formulation extends and reshapes existing value, it does not create a fundamentally new source of it.
Judging the two buckets on the same "expected NPV per program" metric systematically undervalues Innovation, because a handful of high-payoff outliers carry nearly all of the bucket's aggregate value — averaging across programs (rather than modeling the distribution) makes the bucket look worse than it is.
A portfolio built entirely from the LCM distribution would be highly capital-efficient in the near term — predictable returns, short cycle times, low variance — but by construction it can never produce a category-defining asset, because it never funds the genuinely novel bets that could become one. A portfolio built entirely from the Innovation distribution would have the highest theoretical long-run expected value, but the near-term cash flow volatility and probability of a multi-year stretch with no successful readout would be difficult for almost any organization, public or private, to sustain operationally and financially.
Holding both buckets deliberately is a portfolio-construction decision analogous to a barbell strategy in finance: a base of high-probability, moderate-return LCM programs funds current operations and investor confidence, while a satellite of high-variance, high-ceiling Innovation programs is what determines whether the company still has a compelling pipeline in ten to fifteen years.
The table below is illustrative rather than drawn from any single company's actual figures, but it reflects the structural pattern portfolio teams generally observe when they model the two buckets separately: Innovation programs take years longer, succeed far less often, but the rare winner dwarfs what any single LCM program can return; LCM programs read out quickly, succeed at a high rate, and deliver dependable but bounded value.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| New Innovation | |||
| Lifecycle Management |
With both buckets classified, benchmarked, and understood on their own risk/return terms, the portfolio committee's live decision is where to move the dial — and that decision should be driven by which structural threat or opportunity is most pressing this planning cycle, not by inertia or last year's number.
When a material share of current revenue is concentrated in assets approaching loss of exclusivity within the planning horizon, near-term LCM investment becomes a direct, comparatively low-risk lever for defending cash flow the organization is otherwise structurally certain to lose. New indications extend a franchise's addressable population before generic or biosimilar competition arrives; improved formulations can meaningfully slow erosion even after a base patent expires, by giving prescribers and patients a differentiated reason to stay on the branded product.
Because LCM programs read out quickly and build on already-validated mechanisms, a portfolio committee facing a concentrated cliff can meaningfully move the needle within the same planning cycle it makes the decision — a rare property in an industry where most R&D decisions take a decade to show results. This immediacy is precisely why LCM investment is the natural first response to a near-term revenue-defense problem, even though it does nothing to solve the underlying long-term pipeline question.
When the long-range pipeline view shows too little novel, differentiated science entering clinical development relative to what will be needed to replace revenue a decade or more out, the correct response is a shift toward Innovation — even though it will not show up in reported results for years and will increase near-term variance in R&D output. This is the harder call organizationally, because the benefit is diffuse, distant, and unprovable in advance, while the near-term cost (higher failure rate, longer time to any readout at all) is immediate and visible every single portfolio review.
Organizations that consistently under-invest in Innovation during comfortable periods — when current revenue is strong and there is no immediate cliff pressure — are the ones most likely to face a severe, compounding crisis later, because Innovation-bucket assets that were never started cannot be accelerated into existence once the need becomes urgent.
The two triggers are not mutually exclusive and often coexist: a company can simultaneously face a near-term cliff that argues for more LCM and a thin long-term pipeline that argues for more Innovation — which is exactly why rebalancing is a genuine tradeoff decision rather than a formula with one correct answer.
A critical and frequently underweighted feature of this decision is timing lag. A shift toward more LCM investment produces visible results within the same multi-year planning cycle — new indications and formulations read out and launch relatively quickly. A shift toward more Innovation investment produces essentially no visible output for years, sometimes most of a decade, because discovery-to-approval timelines for genuinely novel mechanisms are long by nature.
This asymmetry creates a persistent organizational bias toward LCM: it is far easier to justify a budget allocation whose payoff is visible before the next planning cycle than one whose payoff, if any, will not be visible until well after the executives making the decision may have moved on. Recognizing and deliberately correcting for this bias — rather than letting it operate silently — is one of the central disciplines of long-horizon portfolio governance.
The final test of any Innovation/LCM split is not how it looks in this year's budget review, but how the resulting pipeline looks five to ten years out — and the two strategies diverge in a way that is easy to miss in the short run and difficult to reverse once the long run arrives.
A portfolio weighted heavily toward Lifecycle Management produces exactly what it is designed to produce in the near term: high program-success rates, predictable revenue defense, and a pipeline that reads as productive on every near-term scorecard. The structural problem only becomes visible on a longer horizon — because LCM investment by definition extends the value of assets that already exist, it does nothing to replace that value once the underlying franchises eventually genericize regardless of how well they were defended.
Modeled forward, an LCM-heavy portfolio's projected value tends to hold up well for the first several years of a ten-year projection — sometimes outperforming an Innovation-heavy alternative for most of the window — before beginning a visible decline as the extended-but-still-finite-life franchises reach the end of what defense can achieve, with too few novel assets having been started early enough to be ready to replace them.
The danger of an LCM-heavy strategy is specifically that it is invisible in the metrics executives are typically judged on quarter to quarter and year to year — the erosion shows up only once it is largely too late to correct within the same leadership cycle.
A portfolio weighted more heavily toward Innovation produces a visibly noisier near-term trajectory — more programs failing outright, longer stretches without a readout, and less predictable year-to-year output, which is uncomfortable for any organization's external reporting and internal morale. Modeled forward, however, the same long-tailed distribution that makes individual Innovation programs unattractive in isolation is what gives the aggregate portfolio a meaningfully higher projected ceiling by the end of a five-to-ten-year window, provided even a modest fraction of the novel-mechanism bets clear the bar.
This is the core reason long-range portfolio models nearly always show Innovation-heavy and LCM-heavy trajectories crossing rather than running parallel: the LCM-heavy line starts higher and flatter, the Innovation-heavy line starts lower and more volatile, and depending on how many years out the projection extends and how the novel bets actually resolve, the lines can cross well before or well after the horizon a given planning exercise cares about.
Because Innovation-bucket outcomes are dominated by a small number of high-variance events, any single projected line for an Innovation-heavy scenario should be read as one draw from a wide distribution of possible outcomes, not a forecast in the way an LCM projection — built on much more predictable, higher-probability programs — more reasonably can be. Sophisticated portfolio teams present these projections as a fan of plausible trajectories rather than a single line precisely because a single number invites false confidence in a bucket whose entire strategic rationale is that most individual bets will not pay off.
The practical takeaway for a portfolio committee is not that one split is unconditionally correct, but that the planning horizon chosen for the projection matters enormously to which strategy looks better — a five-year window will tend to favor LCM-heavy allocations, while a ten-year-or-longer window increasingly rewards a portfolio that started funding its Innovation bucket early enough for the long tail to have a chance to pay off.