Understanding Webalizer Statistical Reports

11/02/06 4:50:40 PM (Edited 11/02/06 5:56:45 PM)
Posted by Ken Treis, President

Every time another computer requests something from our servers, information about that request is saved to a log. Almost all web servers maintain these sorts of logs. Logged information usually includes:

  • the Internet address of the computer that made the request
  • the type of request that was made
  • the URL from the request
  • The type of response (e.g. OK or Not Found) that was given, if any

Our technicians read these logs periodically, especially if they are working to diagnose a problem in your web application code.

The Webalizer is a program that we run to help make it easier to understand your logs. The reports generated by The Webalizer can help you understand traffic patterns on your site.

Viewing Reports

To view the reports generated by The Webalizer, simply navigate to the link we provided to you when we configured your site statistics package. If your site doesn’t have a statistics package or if you have lost the link for any reason, please contact us.

The first page you see will be a summary of the last 12 months of activity. This summary is updated every night, and includes links to detailed reports for each of those months.

Older Reports

As time goes on, reports from more than a year ago will start to disappear from this list. But even though they’re not listed, they’re still present on our servers. Here’s how to view them:

First, click on any of the current monthly reports. Notice how the URL in the address bar of your web browser looks something like this:

http://www.your-site.com/.../usage_200611.html

The last part of that URL (200611) is an indicator of the month for the report (200611 = 11/2006, November of 2006). To view a report from a different month, simply change the year and month numbers in the URL accordingly and press Enter.

Hits, Pages, Files, and Visitors

Each time any computer requests anything from our servers, a “hit” is registered. People like to count these and brag about the massive number of hits their site receives, but by themselves hit counts mean very little.

Counting hits alone is sort of like counting the boards in your house. You’d get a high number, to be sure, but there are too many varying things that influence the final outcome. How many rooms does the house have? Was it built using solid construction or just the minimum requirements of the building code? In short, it depends on how the house was constructed.

Web page hits are even more complicated than that. A normal visitor using a web browser to view your site usually makes an initial request, asking for a single page. But then it makes several more requests after that, because the page includes pictures, backgrounds, or other components that have to be fetched separately. A single visitor who navigates to our homepage, for example, registers 31 hits. Other pages on our site may register more or fewer hits, depending on how they’re built.

Search engines, on the other hand, don’t usually bother to request images or other parts of the page that they consider to be irrelevant. That means that a search engine visiting a page might only register 1 hit. But the search engine will likely follow all of the links it finds, which means it’ll eventually visit every page it can find on your site. This also skews the hit counts.

To try to make some more meaningful sense out of these numbers, The Webalizer reports a broken-down total in addition to raw hit counts. The categories reported are:

  • Files: This is a count of how many times the server actually sent data back to the computer that made the request. Most hits fall into this category, though some systems (mostly search engines) will ask the server for information about the page without asking for the page itself to be sent.
  • Pages: The Webalizer tries to determine how many times a response represented a “page” on your site, to weed out pictures and other components that aren’t really a request for a page from your site. If you have a simple site, this number is probably a good indicator of how many page views your site has gotten. In a complicated web application, however, The Webalizer isn’t always able to determine whether a particular hit represented a page view without some fine-tuning. If your “pages” numbers are almost as high as your “files” numbers, then your site probably falls into this category—let us know if you would like us to fine-tune The Webalizer to report more meaningful page counts for your site.
  • Sites: Every computer on the Internet has a unique number assigned to it, called its IP(Internet Protocol) address. This count represents the number of unique machines that have visited your site. Firewall devices can skew this number, since they can hide several computers behind a single IP address.
  • Visits: The Webalizer tries to determine how many times your site has been visited by clustering hits together. If lots of hits came from the same address, with not much time between them, then they probably represented a single visit. If the same address was seen hitting your site again later, The Webalizer counts that as a second visit. In some situations, some large firewalls (like those used at AOL) can skew these counts too.

More Information

As you can see, web statistics can be complicated. If you’d like more information about the logs on your site, please contact us. You can also visit http://www.mrunix.net/webalizer for more information.

Back to Blog