Understanding Site Logs, Part 1

This document should help explain what all those crazy numbers on your site logs actually mean. The main page shows you a rolling twelve month summary of traffic for your site. The links for each month will bring you to a page that shows the details of traffic specifically for that month. Understanding these pages can be quite a daunting task. Hopefully, this series of documents will make that somewhat easier.

Definition of terms

Hits

A "hit" is, technically, one HTTP request made to your site. A hit is counted for every file that the client browser attempts to retreive from the web server. Note that each image on a page is a separate file -- and that often, single images are broken up into multiple smaller images in order to aid in page formatting. So, a when a page with many images is retreived, it will generate more than one hit. Don't be fooled by this number, it is a very poor indicator of overall site traffic.

Files

A "file" is an HTTP request that resulted in a successful transfer. For example, if a browser requested a file that didn't exist, it would count as one hit but zero files. If a browser requested a file that did exist, it would count as one hit and one file. So, the "files" number will always be less than or equal to the "hits" number.

Pages

A "page" is one complete web page, including all text and images. It is the whole of what you see in your browser window. If a page contains ten images, then requesting that page will result in one "page", eleven "files", and eleven "hits" (one for the page itself, and one for each of the ten images). This is the statistic that is usually counted by those "site counters" that you see. However, they are mostly useless in determining actual site usage.

Visits

Now we're getting tricky. This is the number you want to pay attention to. A "visit" is defined as activity from a single source with no interruptions longer than thirty minutes. What does that mean exactly? Well, if you go to a site, and plod around for a while, and then leave -- that counts as one "visit". If somebody else from another source is doing the same thing at the same time, that would be another visit. If you come back two hours later and browse around some more, that will count as a third visit. If you leave and come back 29 minutes later, you won't generate a new visit -- since you weren't gone that long, you'll be counted as still being within your second visit. During your sessions, you may have loaded thirty pages, and two hundred files, but you've only really "visited the site" twice. This is the most accurate statistic we can provide to give you an idea of your overall site traffic. You might think of it as telling you "how many people used my site". However, this number is still not infallible. Many larger sites will conglomerate connections through a single pipeline, such as a firewall. These devices will typically replace the originating address with its own -- in effect, making requests from all users behind it look as if they're coming from the same source. In this case, if there are five different users behind a firewall browsing different pages on your site at the same time, it will only be counted as one visit because the web server only sees a single source address -- there's no way to tell that the requests are actually coming from different people. For this reason, it is really impossible for anyone to claim statistics about "unique visitors", and any such statements should be taken with a grain of salt.

Sites

The number of different addresses that traffic to your site originated from. This is not the number of visitors, but the number of locations that visitors came from.

KBytes

The amount of raw data served from your site. A byte is one character of information, such as one letter or digit. A "KByte" is a kilobyte, or 1024 bytes. Most of the raw data served for your site consists of graphics, since even a small image can contain many KBytes of data.

URLs

URL stands for Uniform Resource Locator. It's a fancy way of saying "address". The number of "unique URLs" can be thought of as the number of different pages that visitors viewed on your site. This number may seem artificially high however. This is because our software is very dynamic -- i.e., page addresses change with the options you've selected. For example, multiple home searches will show up as different pages, even though they're not, because they (probably) always return different results.

Referrers

A referrer is a site that links to your site. For example, if www.abc.com has a link to your site on their site, and a visitor on their site clicks that link to go to your site, then www.abc.com will show up as a referrer in your logs. This information can be a very powerful tool, as it shows you how the majority of your visitors found your site.

Usernames

This is the number of different users that have logged in to your Web Assistant.

User Agents

A user agent is another way to say "browser". Different brands and versions of browsers are tracked. You'll be able to tell, for example, that 37% of visitors to your site are using Microsoft Internet Explorer version 5.0.