With the growth and wide availability of the Internet, most retailers have successfully added the Web to their other, more traditional distribution channels (e.g., stores, mailings). For many companies, the Web channel starts off as little more than an online catalog tied to a secure electronic point of sale. Although valuable in its own right, such use of the Web falls short of some of the unique possibilities it offers for intelligent marketing. Consider the following intrinsic differences between physical, brick-and-mortar stores, and online, Webbased stores. Physical stores are rather static and mostly customer-blind. In particular, 1) the store’s layout and content are the same for all customers, 2) changes to layout and/or content are generally costly, and 3) visits are not traceable except for limited sale’s data, such as what was bought, when it was bought and by what method of payment. Online stores or commercial Web sites, on the other hand, are naturally dynamic and customer-aware. Indeed, 1) layout and content can be modified easily and cheaply, 2) layout and content can be tailored to individual visitors, and 3) every visit automatically generates a rich trail of information about the customer’s experience (e.g., visit duration, pages viewed, items bought if any, etc.), and possibly about the customer’s persona (e.g., demographics gathered through an online questionnaire at registration time). With such flexibility and nearly everything traceable and measurable, the Web is a marketer’s dream come true. Although data-independent initiatives, such as offering social interactions (e.g., user forums) or providing virtual versions of physical stores (e.g., displays, lighting, music) (Oberbeck, 2004), can clearly enhance the user experience, the full benefit of the emerging and growing Web channel belongs to those who both gather and adequately leverage the rich information it provides.
In the context of e-commerce, Web mining lends itself naturally to a staged approach, where moving from one stage to the next requires increasing sophistication, but also produces increasing return-on-investment. The first stage is limited to the analysis of the direct interaction of the user with the site; the second stage introduces behavioral information; and the third stage enables personalization of the user’s experience. We examine each in turn.