Web 2.0 has ushered in a new era of democratic usage of the Web. It is more focused on the user than its earlier version. This has pushed much more information, in many more formats, on the Web.
Web is a major source of information today. However, it is also a source of information overload. It is not only the user generated content, but also professional publications like newspapers and magazines taking the online route. In addition to the stiff competition in the online businesses, Web is continuously changing and adapting to the demand by diverse users to display more relevant content. What is the best way to handle this? The answer lies in the not frequently mentioned concept of Adaptive Websites.
The First Bite
Carolyn Wei explains the concept by using Amazon.com as an example.
Adaptive websites use data provided by users and monitor their actions on the website to customize the content and layout that will interest the user, e.g., Lonely Planet could display more relevant weekend getaways by considering the user’s location or by understanding the typesof getaways prefered.
No, the My Yahoo! like portals are not adaptive websites, they allow the user to personalize content, but the onus is on the user. A website, by being adaptive, learns from its usage, learns from user’s experiences and adapts itself. The biggest difference in the two is the ability to learn and adapt. The result can transform the view including changes in the layout along with the content.
To be able to serve a user, adaptive websites build the user information within themselves, called user models. The user models depend on two types of information:
- information provided by the user voluntarily
- information gathered by the website over a period of time of usage
The earlier one might comprise age, location, gender, profession or other deterministic factors, whereas the latter is more of an experience out of multiple interactions with the website. Users leave breadcrumbs of their visits, which can used to build information about their interests or likes. Sometimes the navigation options used by them can provide more information about related content or popular content.
However, it is difficult to track every single user like this. Adaptive websites use the technique of clustering to group users and build user models for that. Whenever a user visits the website, the cluster is identified and the corresponding user model is loaded.
To keep on improving the user model, continuous monitoring, data logging and mining of the log is required, which can affect the performance. It is possible that the user does not provide accurate voluntary information, in which case the adaptation will fail. It is therefore important that the user is explicitly told about significance of the information.
To think of it, this concept is applied in lot of places – some websites display the appropriate language depending on my language preference or the IP address in the HTTP header (internationalization and localization). Rojo, a feed reader, asks the user about his/her interests and provides the popular feeds in those interest areas. However this concept has to be applied to a much wider aspect of the website – its design and information architecture.
Today, with online overtaking print, blogs being used for businesses and networking, and websites getting as common as the common man, it is important that the websites now start understanding the users rather than just recording them. The current path leads to Adaptive Websites which will adapt themselves for the user.
Technorati tags: adaptive websites, portals, clustering, user model
Copyright Abhijit Nadgouda.