This is Sci7’s comments on an article in press entitled Optimal Web site reorganization considering information overload and search depth by
Chang-Chun Lin, currently in press and due to be published in the European Journal of Operational Research.
The paper discusses adaptive websites, which are defined by reference to earlier work as: “sites which automatically improve their organization and presentations by learning from visitor access patterns”. There is also a discussion of the challenges of organising an expanding website, the research focuses on browsing rather than searching, though search logs are considered as one of the sources of information on how users are interacting with a site can be determined.
Personalisation is one incarnation of an adaptive site, architects of websites are advised to take care given the fact that users may have different goals, a user may use a search engine for both work and personal reasons for example. The way in which a site adapts to a particular user may be as simple as changing the links which are presented on pages to present personalised paths through a site, or may be more involved, pages may be specially compiled by bringing particular chunks of content together. Those using adaptive techniques should be aware that this approach may make some information difficult or impossible for particular users to access.
The study is highly theoretical, and shows that such techniques can be used to minimise the number of clicks required to reach information, however no examples of real sites which could benefit from such an approach are cited. It is likely that an adaptive approach would only be one element of a broader navigation strategy.
Sci7 has experience in developing both consumer and scientific sites with adaptively generated content and links. One advantage of such sites is that they can respond to user activity trends without manual intervention. A simple example of an adaptive site would be an ecommerce site presenting links to currently popular products on all pages - note not “from the home page”, to account for those arriving via deep links, including those from RSS feeds and search engines.
There is no comment on how heavily personalised sites could be designed to work well with search engines. Thought has to be given to what version of the site a search engine would be served, and how any “adaptation” mechanism would respond to the behaviours of the various search engine crawlers. There is an important consideration of avoiding a penalty for displaying different content to crawlers than human users.
Tags: Usability.
DOI Link to original article: 10.1016/j.ejor.2005.05.029