Designing Pharmaceutical Forecast Models That Actually Work

Every big – and expensive – decision that’s made in biopharma is built on a forecast. From launch decisions to portfolio prioritization and investment planning, to strategic direction – everything uses input from a pharmaceutical forecast model. Yet model design is often where things go wrong. A model that is poorly designed does not just produce a bad number, it undermines the credibility of every decision built on top of it, which ultimately impacts the patient.

Share:

Register today








About this webinar

This webinar covers what separates a good forecast model from a bad one, the mistakes that come up repeatedly in pharmaceutical forecasting, and the design principles that address them. Whether you are building a new model or evaluating one you have inherited, the session gives you a practical framework to apply immediately.

What you’ll learn

  • Why model design matters and the consequences when it goes wrong
  • The most common mistakes seen in pharma forecast models
  • The CATS framework: Consistency, Applicability, Transparency, Simplicity
  • Practical design principles you can apply to your next model build or use to evaluate existing ones

At the end of this webinar, you’ll come away with a practical, case-study-backed framework to build comprehensive, accessible, and consistent forecasting models that your teams will actually use to drive critical decisions.

Speakers

andrew-ward

Andrew Ward

Head of Implementation, J+D Forecasting
Abi Spedding

Abigail Spedding

Implementation Director, J+D Forecasting

Frequently Asked Questions

A forecast model is the foundation every commercial decision rests on. Resource allocation, launch planning, portfolio prioritization, supply chain, strategic direction: all depend on what the model produces. A poorly designed model does not just generate a bad number. It erodes stakeholder trust, slows decision-making, and can corrupt outputs across the entire planning cycle. Even the best assumptions produce unreliable results if the underlying model is structurally flawed.

Four mistakes come up repeatedly. First, models built for the analyst rather than the user: complex tabs, hidden logic, and no clear navigation mean only the original builder can use them. Second, confusing complexity with accuracy: more inputs and methodology do not improve a forecast the data does not support. Third, invisible assumptions: when stakeholders cannot see or challenge what drives a number, bias fills the gap and credibility suffers. Fourth, designing for today without thinking 12 to 24 months ahead: models that cannot accommodate future needs get rebuilt from scratch, which is expensive and disruptive.

CATS stands for Consistency, Applicability, Transparency, and Simplicity. Consistency means models across brands and markets share a common structural backbone, making outputs easier to consolidate and compare. Applicability means methodology is chosen to fit the decision, product, and available data rather than habit. Transparency means every assumption is visible, every output is traceable, and there are no hidden calculations. Simplicity means complexity is only added where it is genuinely justified by the decisions the model needs to support.

They tend to manifest differently. Model problems look like assumptions that cannot be challenged, a structure only the original builder can operate, complexity the data does not justify, or outputs that cannot answer the decisions. Process problems look different: a technically sound model nobody uses consistently, scenario outputs produced but never fed into decisions, or stakeholders bypassing transparent assumptions to plug in numbers they want. That last one is a governance failure, not a modeling one. The diagnostic question is whether the function has the right model with appropriate methodology, standardized outputs, and transparent assumptions. If not, the model needs rebuilding. If it does but those standards are not followed, the problem is ownership and governance, and rebuilding will not fix it.

The session is relevant to anyone who builds, uses, commissions, or relies on pharmaceutical forecast models. That includes forecasting and commercial analytics teams, finance functions involved in AOP and LBE planning, and senior leaders who use outputs to make resource and strategic decisions. If you have ever inherited a model you could not trust, or built one nobody used, this session is for you.

Explore all webinars