Marketing approaches to forecasting problems

✍ Scribed by Professor John H. Roberts

• Publisher: John Wiley and Sons
• Year: 1998
• Tongue: English
• Weight: 90 KB
• Volume: 17
• Category: Article
• ISSN: 0277-6693
• DOI: 10.1002/(sici)1099-131x(199806/07)17:3/4<169::aid-for689>3.0.co;2-n

✦ Synopsis

Much research in marketing revolves around an understanding of the performance of products in the marketplace and the determinants of that performance. By understanding how the marketing mix aects consumer behaviour, marketers can both predict the results of future management action and provide diagnostic information to enable such actions to be further re®ned. Thus, the development of quantitative models to explain the relationship between management control variables and consumer response is a critical area of research in marketing. In this Special Issue of the Journal of Forecasting we examine how well marketers understand what it is going to happen, as well as how to in¯uence it. The Special Issue presents new research in marketing applications of forecasting techniques. In this new research I have tried to present a broad cross-section of approaches that marketers use. Therefore, the purpose of this Special Issue is twofold; ®rst, to provide a framework of the forecasting approaches used in marketing and exemplars of each, and second, to provide advances in thinking in each of these areas.

In organizing the contents of this issue I have arranged the forecasting techniques marketers use into four categories. While the boundaries of these categories may be blurred, they do form distinct approaches to the forecasting problem. The four techniques in this classi®cation include econometric and time-series analysis modelling at the aggregate level, longitudinal analysis of individual purchase patterns using scanner data, dynamic models of the product life cycle of new products at the aggregate level, and new product prelaunch models calibrated on individual-level data. These four traditional types of modelling can be represented as follows: