CORPORATE ASSET -- OR A MERE PRESENCE ON THE WEB?
by Inderpal Bhandari, executive editor at large

With the slow but sure advent of electronic commerce and increasing corporate use of the Internet, people are beginning to view the data in their Web server logs somewhat differently than in the past. Traditionally, the main purpose of Web log analysis was to understand the behavior of visitors to the site, where they went, where they spent the most time, what were the most popular paths through the site, etc. Once that understanding was gained, it can be used to improve the usability of the Web site. However, this view is evolving to something much more.

The new perspective is to look upon one's Web site as an on-line investment. This is a direct result of the increasing use of the Internet by corporations for business activities, the advent of electronic commerce, and finally, the expense associated with creating and maintaining a sophisticated Web site.

Consequently, one now considers more business-like issues in addition to the traditional issues of Web design and usability. How is my investment paying off? What can I do to increase that return? How can I transform the data in my Web logs into a corporate asset or, in some cases, a product that can be sold?

While previously Webmasters and system administrators were the main consumers of the data in Web logs, now it is the business executives who are most interested in that information.

That interest is rapidly creating a new industry in the area of Web traffic analysis, or as it is referred to in fashionable circles, Web mining. I came across a definition of Web mining in a white paper from Marketwave:

"Web mining enables discovery of meaningful business correlations and trends by integrating and analyzing Internet/Intranet Web traffic information and traditional business data."

While the phraseology could be improved, this is as good a working definition as I have seen of this concept. It makes clear that the data in the Web server logs is only the starting point of this process. Those data must be enhanced by adding traditional business data to the mix and enabling discovery of meaningful business information. Several inferences can now be made as to how the server logs should be transformed to a corporate asset.

The logs are usually large text files which must be converted into a suitable form to be loaded into a database. This will facilitate the merging of traditional business data as well as enable querying the data. To facilitate the discovery of meaningful business information, the database must clearly identify the variables that are associated with a business outcome. If the corporation is using the Web site to fulfil some need of the user, this is easy to do. For example, if the Web site supports electronic commerce, such variables appear in the Web log data itself and are explicitly related to the behavior to the point that a user decided to purchase or not to purchase something. If the site does not explicitly fulfil a need for the user, then the outcome information is not present in the Web log data. It must therefore be present in the traditional business data. Then, it must be related to the Web log data, which is easier said than done.

Which brings me to the main point. The pre-requisite to Web mining is that the web site must fulfil the user in some way. It must also be designed in a way that makes it clear whether a user is being fulfilled or not fulfilled. Does your web site do this? If yes, you have laid the foundation to convert your on-line investment into a corporate asset. If not, you merely have a presence on the web.

---

Inderpal Bhandari can be reached via http://www.virtualgold.com.