Forecasting

Forecasting
Our Forecasting group provides high-level analytical expertise in a variety of capacities. For companies that internalize the forecasting process, we offer training and model auditing services. For companies that value objective external assessments of their pipeline, we create complete forecast models, and partner with the client to populate the model with assumptions.

Supporting services that we also offer include:

  • Auditing existing forecasts, for both technical accuracy, and plausibility of assumptions
  • Retrospective analysis of promotional campaigns
  • Analysis of physician-level data
  • Therapeutic Area Strategic Planning
We have extensive experience across virtually all therapeutic areas, and experience forecasting at all stages of product development from preclinical to end-of-lifecycle. There are certain key elements that are present in the majority of our forecasts:

Hybrid Patient-/Rx-Based Structure – Patient-based models are great for assessing the potential size of a market, and allowing for demographic trends to exert their proper effect, but they are very poor at assessing the current size and trends in the market today. Rx-based models are wonderfully detailed measures of the current market, but they provide very few clues about the future development potential for the market. Humanumeric has united them in a single approach creating a much more robust model. It is now possible to see short-term accuracy and long-term predictive value in one place.

Analysis of Historical Data – Important market metrics are buried within the sales data. The proper approach will reveal them, and supply our clients with tremendous insight into the structure of the market. Unifying these insights with lessons learned from the epidemiology of the disease yields crucial information.

Curve-Fitting – Having assessed the market trends in the current market, those trends can be extended to the end of the forecast period to serve as a “Default Forecast”. This is the “What if nothing happens?” scenario, and is very important, because it serves as the basis for all future events. Standard statistical techniques (Exponential Smoothing, Single/Multiple Regression, ARIMA, etc.) have proven to be less effective than we would like, so we have developed our own approaches, which are more suited to the kind of timeframes of interest to these markets.

Future Market Events – How the model handles future events is of critical importance to the robustness and flexibility of the forecast. There are several questions that need to be addressed before we can assert that the methodology used for future events is appropriate.

  • Auditing existing forecasts, for both technical accuracy, and plausibility of assumptions
  • Retrospective analysis of promotional campaigns
  • Analysis of physician-level data
  • Therapeutic Area Strategic Planning
Our approach addresses each of these constraints, and allows model users full control over the dynamics of an event with a compact and manageable set of parameters.