EBizinsights XL 3.0 is the most recent release of Visual Insights' visual Web-site performance analysis tool. By providing a highly graphical display of data managed by an online analytic processing (OLAP) data store, Visual Insights attempts to make it easy for business managers to make sense of their Web efforts. Specifically, Visual Insights aims to improve its customers' understanding of site performance and activity, visitor behavior, and promotional effectiveness by enabling them to use a process of visual correlation to learn about visitor behavior and visitor responses to promotional stimuli.

The product line, for which eBizinsights is the base product, includes three extensions: Business Performance Pac 1 enables eBizinsights to produce reports that include product, order, and customer demographic information; Business Performance Pac 2 adds tools that let developers get under the hood and extend the eBizinsights analysis capabilities and create custom reports; Business Performance Pac for Microsoft Commerce Server provides a set of SQL Server Data Transformation Services tasks to extract Commerce Server transaction information and populate the eBizsights XL data set.

The eBizinsights product includes several client and server components. The client side has two components: the Visual Portal, which lets users select and tailor the presentation of about 200 graphical reports (called "Insights"), and the Visual Path Analysis module, which graphically depicts the paths that users have followed through a Web site. We looked at eBizinsights XL 3.0 running in its base configuration and augmented with Business Performance Pac 1 to facilitate evaluating its use as an e-commerce site analysis tool.

Functioning as the client in a client/server deployment model is the eBizinsights Visual Portal, with server components that build on Microsoft SQL Server 2000 performing the server function. EBizinsights' installations typically involve on-site participation of vendor or partner personnel. These partners plan the installation and initiate the processing with the eBizMiner utility that parses the client's Web logs into a set of relational tables and associated OLAP cubes in the Microsoft Analysis Services Data Warehouse; Visual Insights calls this data set its webhouse.

Once the webhouse is established, users can configure their systems to run update scripts regularly. In our case, Visual Insights personnel gave us a few sets of sample data and walked us through the installation via phone. When that was done, we were able to experience and review the Visual Portal from a businessperson's perspective.

When it starts up, the Visual Portal presents a three-column display. (See Figure 1). The most salient feature of the left-most column is a tree-structured list of Insights, which roughly represent predefined queries or views of the data in the webhouse. EBizinsights provides a variety of predefined Insights, organized into subject areas that include Advertising, Visitor, Site Activity Insights, Product Sales, and Site Operation. Below the Insights tree is an area that presents some statistics about the current data set, and a panel that displays any currently active filters.

The middle column displays three bar charts arranged vertically. On top, the Focus bar chart displays a conventional two-dimensional (2D) bar graph of the current Insight's primary dependent variable against what the Insight designer considered the most interesting independent variable. In the middle, the Correlation bar chart graphs the dependent variable against what the Insight designer considered the second most interesting independent variable; the viewer can change which independent variable this chart displays. The bottom-most bar chart graphs the dependent variable against time. The right-most portion of the Visual Portal displays the main graph, which correlates the information presented in the Focus and Time bar charts.

Drill up, drill down, and filter controls accompany each of the bar charts. The drill up and down controls allow you to display the selected chart at higher or lower levels of aggregation. The filter control lets you assign a name to the currently viewed set of data so that you can easily return to it. You can dictate that a filter remain available only during the current session, that it persist but remain private, or that it become available to all users.