Hybrid model building for forecasting new product demand in retail chains

Abstract

Predicting the demand of new products is extremely crucial as it has a strong bearing on manufacturing decisions, marketing efforts, sales strategies, financial planning and profitability. Predicting new product growth poses the challenge of having limited historical data points to model. In addition, they have highly uncertain future demand patterns and are heavily influenced by a host of external factors that are not completely known in the initial stages of the product life cycle. This solution adopts a hybrid methodology by analyzing the influence of new product adoption, its replacement cycle, behaviour of innovators and imitators purchasing this product and the impact of leading macroeconomic indicators on its demand.