fast track article using advance purchase orders to forecast new product sales wendy w moe c15 peter s fader university of texas austin 2100 speedway cba 7 216 tx 78712 pennsylvania wharton jon m huntsman hall suite 700 philadelphia pa 19104 bus utexas edu faderp upenn arketers have long struggled with developing forecasts for products before their launch we focus on one data source that has been available retailers many years but rarely tied together postlaunch put forth a duration model incorporates the basic concepts diffusion mixture two distributions representing behavior innovators ie those who place and fol lowers wait mass market emerge resulting mixed weibull specification can accommodate wide variety possible patterns this flexi bility is what makes well suited an experiential category eg movies music etc in which frequently observe very different ranging from rapid exponential decline most typical
