Professional Microsoft® SQL Server® Analysis Services 2008 with MDX

25.2. Collecting Data

In most Web Analytics endeavors, there are typically three types of data sources:

• Web log data

• Commerce data

• Campaign advertising data

Most of the data that you will be analyzing is actually based on the web log data, because it describes the actions and patterns of your web site visitors. The commerce data identifies what purchases were made, but the web log data allows you to identify the actions that customers made for those purchases. This allows you to possibly simplify the flow from viewing to buying. Banner ad data (one example of campaign advertising data) allows you to know which customers visited your web site, but the web log data allows you to determine how effective the banner ads were. Customers who click banner ads and see only the first page hardly indicate a successful banner ad campaign, whereas customers who click banner ads and make purchases (based on the clicked banner ads) reveal something that is being done right.

25.2.1. Web Log Data

The primary source for Web Analytics is the web server log files. Depending on how the web server is configured, each time it has been accessed, the web server records each instance to a log file. This information is also known as click-stream data, because each click of the mouse produces data. These log files contain valuable information including the visiting patterns, pages viewed, browser and operating system information, and the length of time your customers spent online.

Issues with click-stream ...