Order forecasts, retail sales, and the marketing mix for consumer durables

The paper examines the problem of forecasting ongoing factory orders and monitoring retail demand, with specific reference to high-technology consumer durables. We present evidence of the managerial importance of the problem and, using a case study of a computer peripheral manufacturer, we describe how different data sources and models can be used to increase prediction accuracy. First we examine the order placement and retail demand process using extrapolative methods that focus on identifying short- versus long-run movements in orders. We then introduce marketing-mix data for improved retail demand tracking and forecasting, and we propose the use of conjoint measurement data to simulate a product's utility over time and include that information in the demand model. Lastly, we describe the forecasting and marketing planning use of these models and discuss their implications.