Basic Modeling Methodology (1.5 hr)
Goals and expectations for a forecast
Patient-Based Models and Rx-Based Models
Modeling Methodology – the basic flowchart of a simple model
Basic Epidemiology (1.0 hr)
Definitions – incidence/prevalence/mortality
Rating epidemiology sources
Meta-analysis – discrete to continuous conversions
Patient Flow Analysis
Microsoft Excel Basics (0.5 hr)
Basic Excel skills necessary to complete the exercises
Future Event Modeling (2.0 hr)
The Diffusion Process for Market Events
Source of Volume Issues
Current Market Data Analysis (1.5 hr)
Conversion of Sales/TRx/EUTRx to patient-equivalents
Persistence and Compliance
Patient turnover issues (patient-equivalents vs. unique patients)
Scenario Generation (1.0 hr)
Real-world Forecasting Issues
Internal Consistency for Assumptions
Curve-Fitting (1.0 hr)
Discussion of standard techniques and their shortcomings
Ordinary Least-Squares (OLS) fitting techniques made simple
Fitting to multiple curves
Creating a Sample Model (2.0 hr)
This section works in among the other sections throughout the workshop, and allows participants to complete a simple model over the course of the seminar.
Capabilities Assessment / Staffing Assistance
Over the years, we have acquired significant experience in the development of modeling and forecasting capabilities, in the management capacity of developing our own departments, and in an external capacity, helping client companies build their own modeling and forecasting departments. The process by which we do this has run all the way from auditing and refining the in-house capabilities of the client company to building an internal forecasting department from the ground up for companies who were previously entirely reliant on outside help for all their forecasting activities. The steps we follow can be broadly summarized, although the specific needs of each company differ. The process is highly collaborative, and not all steps are needed for any one company, but the following is a general roadmap of the framework within which we work:
Scope Assessment: We begin by identifying the way forecasting will be implemented throughout the company. In some cases, a standalone forecasting department that supplies global marketing teams with an independent forecast is preferred, in other cases, the responsibility for the forecast should reside with the marketing team, and the forecasting skillset is on-loan to, or resident within, the group, with a separate forecasting expert group responsible for process control. There are other popular configurations, but identifying the best fit with the corporate culture and overall organization is critical.
Design: Once we know what the forecasting capability needs to be, we can design a broad specification for the department, in terms of headcount and skills.
Recruitment: Over the years, we have acquired significant experience in identifying the right (and wrong) blend of education, experience, skillset, and personality for building a successful team. The key challenge is finding individuals with the technical capacity to build accurate, error-free forecasts, who can also integrate with the marketing teams to ensure that the forecast is well-supported within the organization. Even the hypothetical ‘perfect’ model is useless if the end-users of its output don’t trust it. We typically fit in with the HR practices already in place at the client company, and sit-in on the hiring process (we’re not a headhunting organization), looking for red flags, as well as green flags.
Training: It’s unusual for a new team to coalesce with the perfect mesh of skills, so we work with the team to build a core skillset that enables the group to work near-autonomously. Business competence is typically easier to find than strong quantitative model-building skills. Experience working with other people’s models is still a long way from being able to build one from scratch, so we tend to devote time early on to technical skills training (business and market knowledge is best acquired in-house). The training involves both the ability to sit down in front of a blank spreadsheet and start coding, and the use of forecasting tools.
Tools: For clients where we are working to develop an independent in-house modeling capability, we supply many of the tools we use ourselves as part of our consulting business. These tools range from spreadsheet add-ins to standalone forecasting systems. Both are important to enhancing the efficiency of the team. The less time it takes to code a model, the more time there is to think about the market, and refine the forecast to encompass the collective wisdom of the marketing team. For many forecasts, the market logic is simple enough (or, more precisely, can be treated as being simple enough) to fit within a broad template approach. Our work in modeling systems has allowed us to develop a number of templates that are flexible enough to handle the majority of forecasting situations, and these can be made available to the client.
Ongoing Assistance: There are always situations that are sufficiently unique or complex that outside expert assistance is needed. However much we work towards autonomy, there are times when a particular forecast needs a specialist. For these situations, we function as a member of the forecasting team, and work alongside the client company as we do on any other custom consulting project.