The intelligent e-store: easy interactive product selection and comparison

E-commerce is one of the most active and important Internet application areas, yet selecting a product to buy is normally quite a frustrating experience. In this paper, we identify the principal user tasks: the thinning-game and the end game. The thinning-game is used to find a suitably small set of candidate items on the basis of personal specifications. The end game is used to compare the features of a set of candidate items in order to find a single "right" item to purchase. We derive requirements for these two tasks and use them to review traditional techniques and to propose efficient solutions. The thinning game is solved by dynamic taxonomies, a powerful knowledge management model that also provides for multilingual access and easy user preference tracking. For the end game, which is inherently an information presentation problem, a color-coding scheme is used.