User modeling of online consumers: Betweengender differences in Click Path Data

Given people's differences in preferences, personalities, or demographic factors, building useradaptive systems can aid in providing each individual user with the interface that helps to best achieve the desired outcome. For example, online retailers now have the possibility to offer customized Web sites to individual customers, so as to meet each customer's specific needs. However, online retailers typically only have little data available to segment customers in real-Time. In this study, we propose that observable usage data (such as click path data) can be used to distinguish female and male Web site visitors. Drawing on research in psychology and consumer behavior, we hypothesize that observable differences in Web surfing behavior can be indicative of a user's gender. Our results show that women tend to need less time and have shorter click paths than men in online product configuration tasks, and thus confirm the utility of using observable usage data to distinguish Web site visitors based on their gender.