Time-series models in marketing: Past, present and future

Time-series methods have been available to explain and forecast the behavior of longitudinal variables for several decades. We first discuss why, at first, these methods received relatively little attention from marketing model builders and users. We then show how a number of obstacles to their more widespread use have recently been attenuated. Finally, we identify four developments that may significantly affect the future use of time-series techniques in marketing: the ever-increasing size of marketing data sets, the rate of change in the market environment, a growing interest in exploring the finance-marketing interface, and the emergence of Internet data sources. (C) 2000 Elsevier Science B.V. All rights reserved.