Jeremy Smith's blog

Entry Is Labelled

A Note About Web Server Statistics

Because I need to write this down.

Web server stats are inaccurate and misleading. They do not tell you how many people view the content on your web site in a given month, in a given day, or in a given year. There are only two things web server stats are good for:

  1. A number to show managers who tell you, "I need to see numbers to determine how synergized our forward-looking web costumer relations vision is!!!"
  2. Determining server load.

Many web savvy users use news aggregators, like Bloglines, to consume content. Using such software, the viewer never actually visits your site. The software does but not the reader of the content. In the case of Bloglines, 9700 people might read your web site's content via the service; but it only appears in your web server stats once every couple of hours or so as it polls your XML feed.

Additionally, your content may be duplicated in other places —, as an example pertaining to the Case Blog system. So, according to my stats for the month of September, I have had 2685 unique visitors. This doesn't include:

Just had to get that out of my system.


And, there you have it. Statistics for group and organizational blogs. Hopefully I have not performed a detrimental service to the system for I agree with Kieran below.


  1. gravatar

    I know what you mean Jeremy. We have deliberately not put in stats for our Warwick Blogs users because we don't want to mislead them. We fear that if we give someone some stats which tell them they have no readers, it might put them off when they actually get all of their content read through out aggregation page which is actually where we get most traffic.