Thanks to Dan Carr for the image! Understanding Web Usage:
Finding the Person Behind the Browser

Background

Corporate Web sites are now performing business-critical functions like e-commerce, project management, and sales support. A Fortune 1000 company is estimated to spend $12M per year to produce, maintain, and promote its Internet and Intranet sites. Clearly, their impact must be evaluated routinely.

However, existing software packages for analyzing Web usage do little more than computing the number of hits per page or a geographic load distribution. These tools do not address high-level business concerns about a site, nor do they offer insight into the browsing behavior of its visitors.

UWU is designed to answer questions like: Are visitors able to quickly navigate my site and find the information they require? Can visitors be classified by their browsing behavior? Is my Web advertising strategy working? How are my customers and competitors using my site?


Features of UWU
  • An interactive environment for exploring patterns in visitor behavior. By browsing through a site, the UWU user receives synchronized statistical information about the pages visited.
  • Immediate access to data for any page and sub-area in a site. Existing packages focus only on the most frequently requested pages.
  • Visualization of page-by-page usage information through dynamically generated overlays. Statistical information can be superimposed directly on a Web page, indicating the popularity of each link on a page, for example.
  • Drill-down access to every visitor to a page as well as data mining algorithms to identify typical browsing patterns.
  • Statistical comparison between visit patterns of user-defined groups of visitors or collections of pages.
  • Support for either standard server log files or data collected by network traffic monitors.

Possible Business Applications of UWU
  • Internet Call Centers: Improve routing by analyzing visitor click streams.
  • E-commerce: Dynamically generate content to maximize sales; assess ease-of-use.
  • Pre-fetching: Suggest the next most likely page to be visited, accelerating browsing. Existing schemes download all links on the current page.
  • Targeting Advertising: Tailor banners to the interests of a particular type of visitor.

Try the UWU Data Browser:
For Further Information Contact:
Mark Hansen and Wim Sweldens
Bell Laboratories
uwu@bell-labs.com