Information Accessed or Information Available? The Impact on Consumer Preferences Inferred at a Durable Product E-commerce Website

Most previous choice modeling research infers preferences by assuming that consumers consider all the information available at the point-of-purchase. Because e-commerce sites increasingly incorporate tracking technologies that can monitor consumer behavior on their site, our research studies how incorporating the information accessed by consumers into a choice model impacts model performance and inferred preferences. We use data from an electronic goods manufacturer that monitored the attribute information accessed by 582 shoppers while they made Customize and Buy decisions at the firm's website. We find that incorporating the information accessed by consumers into the choice model provides more valid estimates of attribute preferences and better fitting choice models than models based on information available. Because firms can easily obtain this type of information as a by-product of their online operations, we propose that managers who monitor information acquisition and apply the information accessed model will have a useful methodology to gain a better understanding of consumer preferences.