Personalization - Web Pages

Web Pages

See also: Web pages and Adaptive hypermedia

Web pages are personalized based on the characteristics (interests, social category, context, ...) of an individual. Personalization implies that the changes are based on implicit data, such as items purchased or pages viewed. The term customization is used instead when the site only uses explicit data such as ratings or preferences.

On an intranet or B2E Enterprise Web portals, personalization is often based on user attributes such as department, functional area, or role. The term customization in this context refers to the ability of users to modify the page layout or specify what content should be displayed.

There are three categories of personalization:

1. Profile / Group based
2. Behaviour based (also known as Wisdom of the Crowds)
3. Collaboration based

Web personalization models include rules-based filtering, based on "if this, then that" rules processing, and collaborative filtering, which serves relevant material to customers by combining their own personal preferences with the preferences of like-minded others. Collaborative filtering works well for books, music, video, etc. However, it does not work well for a number of categories such as apparel, jewelry, cosmetics, etc. Recently, another method, "Prediction Based on Benefit", has been proposed for products with complex attributes such as apparel.

There are three broad methods of personalization:

1. Implicit
2. Explicit
3. Hybrid

With implicit personalization the personalization is performed by the web page (or information system) based on the different categories mentioned above. With explicit personalization, the web page (or information system) is changed by the user using the features provided by the system. Hybrid personalization combines the above two approaches to leverage the best of both worlds.

Many companies offer services for web recommendation and email recommendation that are based on personalization or anonymously collected user behaviors. Following the example of Certona, the online retailing industry has been early adopters of 3rd party personalization tools offered by companies like PredictiveIntent, MyBuys, Barilliance, iGoDigital, 4-Tell and Amazon.com. Other 3rd party vendors like Choice Stream or Vignette bring personalization to content and display advertising.

Web personalization is closely linked to the notion of Adaptive hypermedia (AH). The main difference is that the former would usually work on what is considered an Open Corpus Hypermedia, whilst the latter would traditionally work on Closed Corpus Hypermedia. However, recent research directions in the AH domain take both closed and open corpus into account. Thus, the two fields are closely inter-related.

Personalization is also being considered for use in less overtly commercial applications to improve the user experience online. Remote control manufacturer Ruwido developed an interactive IPTV platform in 2010 called Voco Media, which controls digital media in the living room using web personalization. It uses personalization as a tool that supports modern forms of TV usage, by allowing users to create different profiles for each family member, personalized menu structures and fingerprint recognition.

Internet activist Eli Pariser has documented that search engines like Google and Yahoo News give different results to different people (even when logged out). He also points out social media site Facebook changes user's friend feeds based on what it thinks they want to see. Pariser warns that these algorithms can create a "filter bubble" that prevents people from encountering a diversity of viewpoints beyond their own, or which only presents facts which confirm their existing views.