Designing Multiple TPPs with Multiple Ends in Mind
Why Optionality and Proper Commercial Evaluation Are the Real Keys to Value (The GLP-1 Story)
I’m mindful that my philosophy of optionality can seem attractive, but perhaps too big a change to the current state. I’d like to argue that many of our success stories are actually stories of better, earlier, learning, and that pattern is one we should see and pivot towards - so I’ve pulled out what mattered for one of our more interesting recent case studies.
I’ve been reflecting again on one of our industry’s favourite mantras: “design with the end in mind.” Borrowed from Stephen Covey, it sounds wise - until you realise most of us interpret it as “lock in one rigid Target Product Profile (TPP) as early as possible and march toward it.”
That approach has never made sense. The worst number of options in pharma? Zero. The second worst? One.
Biology doesn’t care about your Gantt chart or your single-scenario spreadsheet. It hands out wild cards. The winners aren’t the ones who predict the future perfectly (no one can). They’re the ones who learn asymmetrically - faster, cheaper, and with outsized upside when the data aligns. And the mechanism that turns that learning into competitive advantage is optionality, built on proper commercial evaluation from the very first morning of discovery.
Let me show you what that actually looks like, using the story everyone is still talking about: GLP-1 receptor agonists.
The Early Trap: One TPP, One Path, Symmetric Failure
When Jens Juul Holst and colleagues (working with Michael Nauck) showed in 1993 that GLP-1 infusions could normalise fasting blood glucose in poorly controlled type 2 diabetes patients - without hypoglycaemia risk - it was a high-five moment. Novo Nordisk held the patents. The hormone clearly suppressed appetite too.
But the initial TPP was narrow and linear: diabetes only, short-acting molecule, modest expectations. Native GLP-1 has a 1–2 minute half-life. Early analogs flopped on practicality. Even liraglutide (Victoza, approved 2010) hit walls - nausea at therapeutic doses, disappointing weight loss when doses were dialled back. The first approved GLP-1RA, exenatide (Byetta, 2005), was twice-daily with only modest efficacy (~1% HbA1c drop, 2–3 kg weight loss).
Classic one-option thinking. A single end in mind. Symmetric learning: everyone chasing the same diabetes endpoint, the same injection frequency, the same modest profile. The result? A niche product, not a blockbuster.
The Pivot: Multiple Ends + Commercial Foresight = Asymmetric Explosion
Then Novo did something different. Lotte Bjerre Knudsen’s team didn’t abandon the molecule - they engineered optionality into it.
They acylated liraglutide for once-daily dosing via albumin binding. More importantly, they re-evaluated the commercial landscape properly: payers weren’t buying simple HbA1c drops - they mostly had that already; physicians needed convenience and adherence; patients (and the market) had obesity written all over the rodent data. So they tested higher doses with gradual escalation to manage nausea. That small, cheap probe unlocked Saxenda for obesity in 2014.
Here’s the part worth digging into: developing for 15–20% weight loss could have seemed obvious sooner. Appetite effects had been known since the mid-1990s. Knudsen herself had seen dramatic rodent weight loss in 1994 and pushed obesity development for liraglutide alongside diabetes. Yet the true blockbuster path waited until semaglutide.
Why the delay? Diabetes was the lower-risk commercial beachhead - clearer reimbursement, existing infrastructure, and proven incretin science. Obesity indications carried higher regulatory hurdles, longer trials, and payer scepticism (decades of failed weight-loss drugs, or drugs that had made it to market, and then underwhelmed).
Modest weight loss at diabetes doses didn’t justify the investment until a more potent molecule existed. Semaglutide (compound #217 from a screened library) changed everything: C18 fatty di-acid, hydrophilic γGlu-2×OEG linker, and an Aib substitution for DPP-4 resistance stretched the half-life to a week and delivered stronger obesity signals.
Crucially, they didn’t stop at the original diabetes TPP. They ran parallel paths - basket-like exploration across populations and endpoints. The SELECT trial (non-diabetic overweight/obese patients with CVD) revealed a 20% MACE reduction. Emerging real-world data showed potential drops in risks of depression, anxiety, substance use, and even some neurodegenerative signals. Suddenly GLP-1s weren’t just for obesity or diabetes; they were metabolic blockbusters modifying multiple chronic diseases.
The obesity market isn’t an obesity market anymore. It’s an $80B+ (and growing rapidly toward $150B+) platform for disease modification. Payers see hospitalisations avoided and QALYs gained. Physicians get precision tools. Patients get life-changing holistic benefits. That’s what proper commercial evaluation from day one - payers, physicians, patients - plus deliberate optionality delivers.
Keytruda did the same with its adaptive basket trials and PD-L1 biomarker from Phase 1. Sildenafil pivoted from angina to Viagra on an “off-path” signal that a rigid TPP would have killed. The same pattern.
From Rigid TPPs to Target Opportunity Profiles
This is why traditional single-scenario TPPs are dead. Or at least, they should be.
A single TPP set too early - often by one or two departments, who are perhaps less worried about the commercial business case - creates the illusion of certainty. It symmetrises learning. Everyone ends up with the same endpoints, same populations, the same commoditised label.
Instead, treat early phases as a garden of possibilities. Plant 3–5 scenarios via a Possibility Tree. Use adaptive designs, parallel PoC studies, biomarkers, and real options analysis (ROA) to value the flexibility. Turn New Product Planning into New Product Possibilities.
Build commercial evaluation in from the start - not as a late-stage checkbox, but as the lens that decides which branches to nurture. What payers will actually pay for. What physicians will prescribe. What patients will adhere to long-term. Quantify the asymmetric upside of those extra paths.
The GLP-1 story proves it: the “delay” on the high-weight-loss obesity TPP wasn’t stupidity - it was disciplined risk management plus kept options alive. When the data finally aligned, Novo captured billions in upside that a rigid diabetes-only plan would have left on the table. Imagine that insight had come even one year earlier - an upside bigger than most drugs ever make.
The 2026 Imperative
We’re in the middle of patent cliffs, AI acceleration, and payer demands for outcomes guarantees. The companies that will win aren’t the ones worrying about one perfect TPP. They’re the ones designing with multiple ends in mind - marking out constellations of possibilities, engineering planned serendipity, and learning asymmetrically.
Biology still holds the wild cards. But now we have the tools to play them.
So the next time your team sits down to draft a TPP, ask the uncomfortable question: “Are we building one end… or a garden?”
Because in pharma, optionality plus proper commercial evaluation isn’t a nice-to-have. It’s the only way a TPP actually delivers its true utility.
Really compelling perspective. The case for optionality, especially through the GLP-1 lens, is hard to argue with.
What resonates most is the idea that early certainty can limit learning. We see this often in Medical Affairs when initial framing narrows how evidence is generated and how value is ultimately understood.
Where I would add some nuance is in how this plays out in practice.
Optionality is powerful, but it requires discipline. Most organizations do not struggle with lack of ideas. They struggle with focus and execution across too many parallel paths. Without clear prioritization, optionality can quickly become noise rather than advantage.
I also think there is an important balance in how early commercial evaluation is introduced. Understanding payer, physician, and patient needs early is critical. At the same time, scientific credibility has to lead. In Medical Affairs, that trust is what ultimately allows optionality to translate into meaningful impact.
For me, the opportunity sits in the middle. Using optionality to expand what is possible, while staying grounded in evidence and clear choices about where to invest.
That is where this becomes less about designing multiple ends, and more about enabling the right ones to emerge with confidence.
The SELECT trial shows market underestimates the impact of GLP-1s, with a 20% reduction in MACE in non-diabetics shifting the narrative from metabolic to cardiovascular, affecting reimbursement. Investors note infrastructure lagging adoption but catch-up is rapid once mass market is reached. The Keytruda analogy fits: trial approach creates enduring advantages in clinical evidence not easily copied by biosimilars. The question is whether Lilly's retatrutide (a triple agonist) marks a molecule-level break like Novo's semaglutide or if future advances will focus on formulation and delivery rather than molecule design.