Probability-based forecasting for U.S. generic drug sales

Abstract

Most sales forecasts, including those in the generic drug industry, are based on an implicit assumption that the market can be represented as a continuous variable, an approach that works only when activity consists of many independent, incremental customer buying decisions, each of which is too small to substantially affect total sales. These conditions are generally true in branded pharmaceutical markets but they do not reflect the reality of the US generic drug industry where decisions regarding which company’s product is dispensed at the pharmacy are primarily a function of a limited number of binary (yes/no) decisions from drug wholesaling companies, the largest of which control access to one-fourth of the market or more. Unfortunately, binary situations are difficult to model in standard spreadsheet forecasts due to the extremely high number of permutations that are possible even with a small number of variables. In this paper, we explore the binary nature of the US multi-source industry from a single, hypothetical, generic company’s perspective and discuss a probability-based sales forecasting technique that offers a more accurate approach to modeling based on the number of wholesaler contracts available and the degree of generic competition present. The forecast is derived from actual wholesaler market share data and, based on extensive industry experience, the parameters used are believed to be reasonably representative of the US generic market, meaning that the results can be scaled to determine approximate probabilities of achieving certain revenue levels in real-world situations.