July 19, 1999 DESIGN AND EVALUATION OF A TRUST BASED ADVISOR ON THE INTERNET

We propose that the Internet can enable trust based marketing through accurate and complete information and advice. We concentrate on shopping advice and develop a virtual personal advisor as an important component of trust building. Bayesian utility and segmentation models are used to make recommendations and a dialogue-based computer graphical user interface (GUI) is designed to instill confidence in the advisor's recommendations. A prototype system, called Truck Town, is developed for pickup truck purchasing and the proposed site is evaluated in an empirical concept and use test. The key research questions examined are: Can a site be developed on the Internet that develops trust via a virtual advisor? Can information intensive and virtual advisor sites do as well as face-to-face selling in terms of establishing trust? Which segments of consumers prefer a virtual advisor-based Internet site to objective, information intensive sites with no advisors? The results indicate that the proposed site can build trust; most customers would consider buying a vehicle through Truck Town and would be willing to pay for the service. A combination of the advisor site and an information intensive site together dominated the existing auto dealer system in terms of customer acceptance. The preference for the two Internet sites was equal, but analysis indicates segmentation in the site preferences. Consumers who are not very knowledgeable about trucks, who visited more dealers, and who are younger and more frequent Internet users have the highest preference for the virtual personal advisor. This paper closes with a discussion of future research and strategic marketing issues raised by a virtual trust based advisor system.