Me, Me, Me
By Amit Asaravala
In 1998, personalization was a hot topic among retailers and media outlets with active Web sites. Companies were looking for ways to enhance their online customer service and experience, and numerous sites began copying the my.yahoo.com model. Pretty soon, most prominent B2C sites had some sort of "my site" feature that let customers set preferences for the content on pages they were served.
Of course, no generation of technology appears without pundits instantly looking forward to the next generation of technology. Industry leaders have talked about the day when sites would automatically learn your preferences by recording and analyzing your habits. No longer would you have to scan through newspaper headlines in search of interesting articles—your personal news site would automatically serve up recent articles about your favorite sports team.
While academics were researching so-called personal learning agents long before 1998, few sites had successfully implemented such a system. Commercial software for intelligent personalization was even scarcer. Surprisingly, the technology isn't complex: Use cookies or a login prompt to identify each unique visitor, modify the URLs on a site so that each click-through generates a new database record, and serve the main Web page via a script that checks the database and compiles a listing of recommended content based on a user's previous history.
Despite this, only a handful of companies are developing personalization software that infers user preferences. Few new companies have entered the space, and many of the initial vendors have disappeared, consolidated, or started selling content management solutions.
