is a PhD candidate at Macquarie University. His research focuses on direct and database marketing, empirical model comparison and modelling diffusion of innovations. His research has been published in academic journals and proceedings of various international conferences. Currently Bakher holds managing position at a multinational telecommunication company. Greg Elliottis a professor of business (marketing) at Macquarie University. He has also held academic positions at a number of other Australian and overseas universities. He has published widely and his current research interests are in the fi elds of fi nancial services, international, cross-cultural and social marketing. He currently teaches marketing management, strategic marketing and services marketing. He has also had extensive experience in management education and training in Australian and Asia for major international and local companies.ABSTRACT The use of demographics by researchers in the online shopping literature is common, however, they are typically constructed as either moderators or control factors. Little attention has been given to explicitly modelling the predictive utility of demographics. The present research models the impact of nine demographics, six social connectedness measures and fi ve prior online experience variables on consumers ' actual online purchases. A large and representative data set was used. Our results show that a model on the basis demographic data alone explains 22.6 per cent of the variance in the consumers ' overall online shopping behaviour. The model ' s utility increased to 45.4 per cent once social connectedness and prior internet experience were added to the model. Furthermore, analysing 14 online product categories, we found that the predictive power of demographic variables is product specifi c. Overall, our results strongly support the use by practitioners of demographics as powerful predictors for direct targeting of online shoppers.

is currently employed by a major multinational telecommunications company in a senior management role. His research has been published in international academic journals and proceedings of international conferences. His research and professional interests are focused on direct and database marketing, data mining, empirical model comparison and modelling the diffusion of new products and technologies. Greg Elliottis a Professor of Business (Marketing) at Macquarie University, Australia. He has held academic positions at a number of Australian and overseas universities. His research has been concentrated in the fields of marketing theory, financial services, services marketing and international marketing. He is also active in researching Chinese consumer behaviour.Correspondence: Greg Elliott, Marketing and Management, Macquarie University, Herring Road, North Ryde, NSW, 2109, Australia ABSTRACT The results of past studies that compared the performance of alternative growth models are generally inconclusive. The objective of the current study is to provide further empirical evidence regarding the performance of three popular growth curves, namely, the Bass, Logistic and Gompertz models in the context of online shopping diffusion in Australia. The results of model fitting to an online shopping time series (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009) show that all three models represent the diffusion curve quite well and adequately; however, the Bass model described the online shopping diffusion curve more accurately than the other two models. Forecasting with early diffusion data (1998)(1999)(2000)(2001)(2002) suggests that the Bass, Logistic and Gompertz models are unable to adequately describe the diffusion curve from limited data. Nevertheless, the Bass model with the adjusted market potential coefficient (m) produced forecast accuracy that is comparable to the Bass model fitted to the full data (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009). Overall, our results suggest that the Bass model outperforms the Logistic and Gompertz models.

No abstract