An analysis of frequency and duration of search on the Internet

While seeking information online, consumers exhibit different patterns of search, that is, frequency of visit and duration of visit per search episode. This research empirically examines if consumers' search patterns are affected by the type of information that they seek. A discrete hazard model is used to study the time spent per trip, and an ordinal probit model is used to study the frequency of searching. The baseline hazard is estimated without imposing any parametric restrictions. We study the role of seven different types of information that consumers seek. The model is calibrated on data collected through online surveys.