Author: Sam Hope, CEO, Beyond Blue
Demand forecasts have a habit of taking on a life of their own. A peak-share estimate, an uptake curve, a projected patient number – once it is written on a slide, it can start to feel more solid than it really is. Teams debate whether the number is too high or too low, whether the curve is aggressive enough, and whether the market opportunity justifies the next investment decision. And of course, those conversations matter because significant commercial, clinical, and organizational choices are often built around them. But the number itself is rarely the most important thing. What matters more, in my mind, is the quality of thinking that sits underneath it: the assumptions being made, the uncertainties being explored and the confidence a team can genuinely place in the decision it is about to make.
Numbers can bring false security
The danger with demand forecasting is not that forecasts are imperfect. Everyone involved knows they are trying to make sense of a future that has not happened yet. The real danger comes when assumptions begin to solidify into apparent certainty. We make assumptions about how quickly physicians will change behavior, how payers might respond, how guidelines may evolve, how competitors might reposition, how patients could move through the pathway, and how much operational reality will get in the way of theoretical demand. Each of these assumptions may be entirely reasonable on its own. But if they are not surfaced, challenged, and understood, they can create a level of confidence the evidence has not actually earned.
That is why a single, precise number can be so seductive. As humans, a range feels uncomfortable because it reminds us what we do not know. A number with a decimal point feels useful because it appears to offer certainty. But precision is not the same as confidence. A forecast can look beautifully modeled and still be fragile if the beliefs beneath it have not been fully pressure-tested.
Start with the decision, not the forecast
One of the most common traps in demand research is becoming absorbed in the forecast before being completely clear on the decision. The strongest demand projects don’t begin with a spreadsheet, a model structure, or a debate about whether the primary output should be share, volume, or revenue. They begin with the bigger questions by bringing commercial, medical, market access and forecasting stakeholders together to agree on the decisions being made, the assumptions underpinning them and the uncertainties that could materially change the outcome.
Should we prioritize this asset over another? Which indication offers the most credible opportunity? What would need to be true for an early launch plan to work? Where is the greatest uncertainty – in physician willingness to adopt, market access, patient identification, competitive response, or somewhere else entirely?
Those questions matter because demand research is not about producing a forecast for its own sake. It is about helping teams make better, more informed choices before the future can be observed. A launch team may need to understand whether the biggest barrier is clinical conviction or access feasibility. A commercial team may need to know whether high stated interest will translate into real switching behavior once physicians face practical constraints. A leadership team may need to decide whether the opportunity is large enough and robust enough to justify further investment. In each case, the value of the research lies not simply in the answer, but in the clarity it creates around what the team can act on with confidence.
Starting with the decision changes the whole shape of the work. It focuses the research on evidence that will genuinely influence action. It helps distinguish commercially important uncertainties from interesting background noise, and shifts the discussion from whether people ‘like’ the number to whether they understand what is driving it.
Building confidence through challenging conversations
This is where I find the real value of demand work often comes from. The final forecast may be the visible output, but confidence comes from the process. Conversations that challenge received wisdom, expose hidden assumptions, and force teams to decide what they really believe. In our Beyond Demand approach, that confidence is built through four connected disciplines.
- Asking the right question, so the work is anchored in the decision that needs to be made, not just the output people expect to see.
- Pressure-testing the assumptions that matter most, separating the factors that genuinely shape demand from those that add noise without changing the decision.
- Generating evidence that can stand up to scrutiny, with a clear view of where there is genuine confidence in the number.
- Creating alignment across commercial, medical, market access and forecasting teams, so people are not just agreeing on a number, but agreeing on what it means and how it should be used.
We bring research, analytics, therapeutic expertise and behavioral understanding together from the outset, with our analysts embedded within the research team. Intentionally designed to make assumptions visible, challenge them with evidence, and quantify which uncertainties really matter.
The result is not simply a number. It is a clear view of the drivers that support it, the conditions that could change it, the areas where further evidence would add value, and the decisions that remain robust across different scenarios.
More data won’t necessarily give you a better forecast
There is often a temptation, particularly when the stakes are high, to believe that more data will solve the discomfort of uncertainty. Sometimes more evidence is exactly what is needed. But more data does not automatically create a better forecast, and it certainly does not guarantee better decision-making. The strongest projects are rarely the ones that collect the most information. They are the ones that create the most useful challenge: What would have to be true for this forecast to be right or wrong? Which assumptions are driving the outcome? Where does the uncertainty really sit? Which beliefs are we treating as facts simply because they are convenient?
Those questions reveal whether the forecast is rooted in a real understanding of behavior and context, or in assumptions that have become embedded because no one has had the time, structure, or confidence to challenge them.
Where AI fits, and where I believe caution is still needed
AI can support parts of the wider forecasting process, for example, by helping teams explore hypotheses, identify potential analogs, or interrogate existing information. But using AI or synthetic data as a substitute for market-specific evidence is a different proposition. Demand assessments are shaped by the particular asset, indication, pathway, stakeholder behavior, access environment, and competitive context. If the source data, assumptions, or model logic are not transparent and relevant to that specific decision, a faster answer may simply create false confidence.
Current industry discussion that we have been a part of also points to caution: adoption of AI within forecasting remains limited, concerns about reliability, transparency and accuracy persist, and established spreadsheet-based approaches remain common. For now, AI’s strongest role is later in the insight ecosystem, supporting human-led analysis and interrogation rather than replacing the collaborative, evidence-based work needed to build the forecast itself.
Three principles for more confident demand decisions
- Collaborate early. Bring commercial, medical, access, and forecasting together around the decision, the assumptions, required calibrations, and the uncertainties that could materially change the outcome.
- Embed analytics. Use rigorous analysis to identify, challenge, and quantify assumptions and separate genuine demand drivers from background noise.
- Understand adoption in context. Combine therapeutic expertise and behavioral understanding to move beyond stated intent and reflect how decisions are made in practice, including physician archetypes, decision-making styles, access realities, and treatment dynamics.
A demand forecast cannot remove uncertainty, nor should it pretend to. What it can do is help teams understand uncertainty well enough to make better choices. It can show which adoption drivers are most sensitive, which access assumptions could materially change the opportunity, where competitive scenarios have the greatest impact, and which decisions remain robust across different possible futures.
That is why, for me, the most dangerous number in pharma insights is not simply the forecast figure that turns out to be inaccurate. It is the forecast that people trust without fully understanding why it looks the way it does.
Great demand research should do the opposite. It should make assumptions visible, create space for challenge, and build a shared understanding of the evidence, uncertainty, and implications. It should shift the conversation beyond whether the number feels right, and towards whether the team has enough confidence to act.
None of us can predict the future with certainty, but we can be much more disciplined about how we understand it. And that, ultimately, is what demand research should deliver: beyond a forecast – the confidence to make better decisions.