New Product Forecasting

Abstract

Abstract New product forecasting refers to the process of forecasting the demand for products that have yet to be launched or which have recently been launched. Forecasts are often based on consumer intentions surveys, management judgment or mathematical diffusion models. Data from consumer intentions surveys is usually more reliable when the new product will be purchased soon after the survey has been carried out, but the data usually has to be de‐biased to improve forecast accuracy. Management judgment is subject to cognitive and motivational biases but its role can be improved by aggregating the judgments of groups of managers using methods like Delphi or prediction markets. Diffusion models require historic data on demand. When this is unavailable models can be fitted to the demand history of previously launched analogous products on the assumption that these will have similar sales patterns to the new product.