What Is "Demand" When There Are No Sales?

By Tom Walls, Principal and Founder, Axon Bridge Consulting

In a commercial company, demand planning starts with history. Someone pulls three years of shipments, adjusts for seasonality and promotions, layers in the sales team's view, and the number that comes out is the thing supply plans against. Most of the machinery of sales and operations planning, from statistical forecasting to consensus demand reviews, exists to make that number better.

A clinical-stage company has none of that. There are no shipments to customers, no order history, no sales force with a quota. Ask a program team what next year's demand is and you will get a puzzled look, followed by a reference to the protocol. And yet the demand exists. It is written down in the clinical development plan, the enrollment projections, the stability protocol, the submission strategy, and a dozen spreadsheets owned by people who would never describe themselves as demand owners. The first job of planning in a clinical-stage company is to go and find it.

The Four Streams That Replace Sales

The demand for a clinical-stage product comes from four places, and only one of them looks anything like a sales forecast.

Clinical demand is the one everyone thinks of. It starts with the protocol: the dose, the regimen, the treatment duration, the number of arms, the randomization ratio. It becomes a forecast when you add the enrollment plan: which countries, how many sites, when each site activates, how fast sites screen, how many screened subjects fail. For weight-based dosing you also need an assumed patient weight, and that single assumption alone can produce a large swing in demand. Clinical demand is real patient consumption, but it is consumption of kits at investigator sites, not sales to customers, and the person who owns the assumptions behind it sits in clinical operations.

Analytical and development demand is the stream that gets missed. Method development and validation consume material. Stability studies consume material at every timepoint for the life of the program, and a new formulation or a new site means a new study. Process development and engineering runs consume drug substance that never becomes a kit. What makes this stream different from the others is that it does not land at one level of the bill of materials. Clinical demand is expressed at the finished kit and cascades down. Analytical demand pulls directly at every level: drug substance for characterization and reference standards, drug product for stability, finished kits for in-use and shipping studies, and raw materials and components for their own testing. Each of those pulls has to be time-phased against the supply at that level, not netted at the top and pushed through. In an early program this stream can be larger than clinical demand, and it pulls from exactly the same constrained batches.

Regulatory demand is smaller but nonnegotiable. Submission samples, retention samples, and reference standards all have quantities and dates attached to them, usually set by a filing timeline that regulatory affairs is managing without any reference to what is on the shelf.

Commercial demand arrives last, and only if the program works. It is the prelaunch build, the process performance qualification batches, and the launch inventory. It is the only stream that will ever look like a traditional forecast, and by the time it shows up the clinical streams are still running alongside it.

Who Owns What, And Why They Belong In The Room

The point of listing the streams is not taxonomy. It is that each one has a different owner, and none of those owners report to supply chain. Clinical operations owns the enrollment curve. Analytical development and quality own the stability and method schedules. Regulatory affairs owns the submission plan. Commercial, when it exists, owns the launch forecast. A planner who builds the demand picture alone, by reading documents, will get it wrong, because the documents are stale the week after they are written.

This is the first place where the clinical-stage company diverges from the commercial S&OP model, and it matters for everything that follows in this series. In a commercial company the demand review is a negotiation with a sales organization that has a direct interest in the number. In a clinical-stage company the demand review is a conversation with four functions that have no particular interest in supply at all. Clinical operations wants sites open. Regulatory wants the filing on time. Nobody in that room is going to volunteer that their study is likely to enroll six months late or that the stability protocol just added two timepoints. They have to be asked, on a cadence, with the plan in front of them.

Regulatory affairs deserves a specific mention, because in my experience it is the function planners meet last and should meet first. Regulatory holds unseen demand: samples and retains that are committed to health authorities and cannot be traded away when supply gets tight. The time to learn about that demand is when the submission strategy is written, not when a batch is being allocated.

Getting The Basics Right Before The Arithmetic

Once the streams are identified, three decisions turn them into something supply can use. They sound mundane. Every planning failure I have seen in an emerging biotech traces back to at least one of them.

First, one unit of measure. Clinical operations counts subjects. The CDMO counts liters and grams. Packaging counts kits. Finance counts dollars. The unit that reconciles all of them is the physical package delivered to the dosing site, whatever that is for your product: a vial, a bag, a blister wallet, a kit. Every stream, including the analytical and regulatory ones, should be expressed in that unit, with the conversions to grams and liters held in the bill of materials rather than in someone's head. A stability study expressed as "12 timepoints" is not demand. Forty-eight vials in March is.

Second, the right time bucket for the right purpose. For the finished product forecast that drives depot resupply, weekly buckets at the depot level are the most useful, because enrollment is a curve rather than a line and monthly buckets hide the shape of it. For the cross-functional planning conversation and for anything upstream of packaging, monthly buckets are enough and quarterly is acceptable before a study starts. The failure mode is not choosing wrong; it is having three functions each using a different one.

Third, a horizon that outlasts the trial. A trial may run 18 months. Drug substance to finished kit can take 12 to 18 months on its own, and a CDMO will want a commitment a year before a campaign starts. The planning horizon therefore needs to be at least 36 months for every stream, which means forecasting studies that are still in the development plan and not yet in a protocol. A rough guesstimate for a study two years out is worth more than a precise forecast for the one enrolling now, because the near-term study is already supplied and the far one is what the next campaign decision depends on.

The Hidden Competitor

If there is one message to take from this first piece, it is about the analytical stream. In a commercial company, quality control samples are a rounding error against sales. In a Phase 1 or Phase 2 program they are frequently the largest single consumer of drug substance, and they are time-phased against a schedule that nobody in supply chain has seen. The typical way this surfaces is late: a cohort or a site activation is weeks away, the planner goes to allocate material and finds that stability pulls or a method validation have already consumed vials that were assumed to be available. Nobody has done anything wrong. The demand simply had no owner in the plan.

Demand in a clinical-stage company is not a number. It is a set of assumptions, each with an owner, each of which will change. The number falls out once the assumptions are collected in one place, in one unit, on one horizon. The next article in this series looks at what happens when that number meets supply, and why the arithmetic is different from anything a commercial planner would recognize.