Creators of
Web pages may wish to know how many times their pages are visited.
While the two methods described below will provide statistics on Web
page access, you should be aware that their usefulness may be
limited. For example, overcounting can result from access by
automated processes, while undercounting can happen because access to
cached pages is not counted or the visitor has image-loading disabled.
MIT Information Systems recently introduced a Web Reporting service. Through a small (pixel-size) invisible image on the Web page, the service records when the page is accessed. The owner or maintainer of the page can retrieve reports on access that include date and time of access, whether access was from inside or outside of MIT, the domain name of the accessing computer and its operating system, the browser name and its version.
Counters, like the one at the bottom of this page, simply tote up the number of times a Web page has been accessed. MIT's Student Information Processing Board (SIPB) provides a counter service through their web server. For details, see http://www.mit.edu/doc/counter-howto.html. Keep in mind that many Web users find page counters annoying to look at, as well as slowing the loading of the page.
For another look at this topic, see "Why web usage statistics are (worse than) meaningless".