Home / Finance and Gross-to-Net / Financial Planning and Demand Forecasting
Financial Planning and Demand Forecasting
Written, approved, filled and persistent demand are four different numbers.
Patient-based forecasting built on the full conversion path, long-range revenue scenarios, and clinical trial financial management that carries Phase 1 teams from trial costs to patient demand.
The problem
A forecast built on written demand overstates the patients who stay.
Traditional forecasts start from how many patients exist. The question that matters is how many can be found, reached, started and kept, and a single-point forecast hides which assumptions have to hold. Before launch, the same team is often still accounting for clinical trial costs.
Scroll sideways to see the whole diagram.
The forecast is not the deliverable. The decision is.
Phase 1 · Pre-Launch and Launch Readiness
The team that managed trial costs now has to plan for patient demand.
Before launch, R&D spend is the most material line on the P&L, and Finance is built to account for it. Launch hands the same team a different problem: forecasting patients, access and the deductions that come with them.
Where Finance starts
Clinical trial costs
- Rapid growth, maturity gaps. 80% of trials are sponsored by small biopharma, with an average of 5.8 years from founding to exit.
- Highly complex, highly outsourced. Most R&D runs through third-party vendors, sub-vendors and sites, and the burden of coordination falls on the sponsor.
- Constant change. Scope changes and contract change orders delay milestones and push budgets over.
Where launch takes it
Patient demand
- How many patients reach therapy, and when
- What access and affordability cost to get them there
- What the product nets once the deductions land
Automate the accruals that carry across
- 01Configuration
- 02Confirmation
- 03Calculation
- 04Refinement
Model costs with configured logic, give reviewers the workflow and evidence to verify them, translate operating data into financial impact as it arrives, and compare baseline forecasts and accruals to actuals quickly. The discipline built for trial accruals is the one launch forecasting needs.
Demand prediction
Not how many patients exist. How many can be found, reached and started.
Demand prediction answers the question forecasting models skip: where along the way patients are likely to stop. AI and machine learning run against a representative proprietary database, indicative, not exhaustive of the larger data universe, to move a forecast from a historical projection to a map of opportunity.
The models are only as good as the question. They start from the label, the diagnosed and treated population it actually defines, and the transitions between diagnosis and therapy initiation.
Oncology epidemiology
Persistence, not a start count
Incidence, mortality and treatment outcome modeling by tumor type and line of therapy, with survival analysis that estimates how long patients remain on treatment.
Outbreak dynamics
Ready before demand arrives
Time series forecasting and epidemic models that anticipate spread and utilization, supporting supply planning and channel readiness ahead of demand.
Validation
Every model shows its error
Accuracy and reliability metrics on every model, with automated cross-validation so the evaluation is repeatable. A forecast that cannot say how wrong it might be is not a forecast.
The signals that feed it
The objective is not forecast accuracy for its own sake. It is knowing which decisions change when the numbers shift.
Financial planning
Find the critical intersections on the Roadmap to Revenue.
Strategic planning reduces uncertainty before it turns into forecast variance. Learn from the past to predict the future, and put resources where they drive outcomes.
Plan to report
One Gross-to-Net cycle, from budget to fiscal close.
Reimagine Finance to reduce inefficiency, promote transparency and improve predictability. Every quarter runs the same loop: reconcile the actuals posted, refresh the accrual rates, assess market conditions, predict the access impacts, and carry it into the next forecast.
Long-range planning
A range with probabilities, not a single line.
Long-range plans break when one number carries every assumption. We build revenue scenarios from the drivers that actually move, each with a value range and a probability, so the plan shows which outcomes depend on what.
- 01Sanity-check the drivers: product profile, launch timing, commercial uncertainty
- 02Assign each driver a value range and a probability
- 03Map the risks and opportunities the drivers missed, by probability and impact
- 04Revenue scenarios that state what has to be true for each outcome
What we deliver
Plans that state what has to be true.
Where it applies
The exits it works, and when.
10 Exits of Therapy™
Product Lifecycle Phases
From the work
Where this came from.
Ultra-Rare Pre-Launch Patient Demand Modeling and Launch Readiness
A tiny observable population and thin claims data, where an epidemiology forecast was not enough.
Read the use case →Key conceptThe Forecast Is Not the Deliverable. The Decision Is.
Understanding decision impact matters more than forecasting accuracy in a pharmaceutical launch.
Read on patrickrcoyle.com →