[Air-l] web browsing patterns and distributions?

BEAUVISAGE Thomas RD-TECH-ISS thomas.beauvisage at francetelecom.com
Tue Apr 26 05:30:41 PDT 2005

Hello Allan,

The question you raise was, among others, one of those I tried to answer in my PhD on "Semantics of users' paths through the Web", so it might help you. It was based on user-centric traffic data collected by on-computer probes at home, for 3400 people for jan-oct 2002 and 600 people for jan 2000 - oct 2002.
You can get it at this address : http://thomas.beauvisage.free.fr/publis.en.php
but except the abstrat, it is in French...

So shortly, for your precise question: I observed that for the 34-months panel, on average, for 100 sites visited by individual over the 34 months, 75% appear only in one session, while 1.5 % appear in more than 40% of his sessions. The one-visit sites appear all along the internauts' "life on the Net", which means they continuously view sites that they will never visit again. These sites occure mainly in long, complex sessions (avg. 42% of one's all sessions), containing unusual thematics or services for the user, where search engines are over-represented. On the contrary, they hardly appear in sessions dedicated to frequently revisited sites ; in this context, it consists in following a pointer from a well-known site to discovering a new site, as a kind of "step-aside" movement in a straight Web path.
All detailed information and calculus, as well as classification of Web sessions on the basis of their contents and on personnal web territories, can be found in the PhD thesis.


-----Message d'origine-----
De : air-l-aoir.org-bounces at listserv.aoir.org [mailto:air-l-aoir.org-bounces at listserv.aoir.org] De la part de Allan A Friedman
Envoyé : jeudi 21 avril 2005 06:54
À : air-l-aoir.org at listserv.aoir.org
Objet : [Air-l] web browsing patterns and distributions?

Hi All,

I've been looking for information on web browsing patterns in terms of page exposure.  Ideally, I'd be interested in roughly how many pages people visit in a given time period and the shape of that distribution (i.e. most people look at a few pages, but a long tail look at many pages).  If anyone could point me to some any broadly applicable information about how people "acquire" new pages (search engines, links, email), I would be a happy modeler of user behavior.  I've found plenty of user studies on browsing behavior *inside* a website, but I'm looking for domain-level use for a project on trust mechanism simulation.



Allan Friedman
PhD Student, Public Policy
Kennedy School of Government

The Air-l-aoir.org at listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org

Join the Association of Internet Researchers: 

More information about the Air-L mailing list