Learning from Logs

By Janine Dalton
Web Developer

In these times of advanced conversion tracking and complex cookies, the humble log file may be overlooked. Do not dismiss it! Do you want to know which sites and search engines referred visitors to your site? Or how about when Google last crawled your home page? Perhaps you would like to know the most popular pages on your site or what browsers people access your site with? Your log file contains the answers to these questions and many more.

The first step is to ensure that your host has enabled logging for your site. This is standard practice but there are no guarantees, so double check with your host.

While the log file itself contains all the raw data you need regarding visitors to your website, it makes for slow and tedious reading. Luckily, there is plenty of web log analysis software available that can turn all the text into intuitive charts and meaningful lists. Different analysers excel at certain areas e.g. referrer analysis or search engine spider statistics. It is worthwhile installing a few on your web server and trying them out to find which best meets your requirements. Two popular log analysers are AWStats and Webalizer. They can run on both Apache and IIS web servers and have the added bonus of being free. Your web host can best advise on which log file analysis options are available with your hosting package.

Whichever analyser you choose, you will soon discover that web analytics is a field rich in buzzwords. Here is an explanation of the key terms:

Hits:
Total number of requests for any file (web page, graphic, JavaScript file etc) from the server. If your page contains a style sheet and 3 images, the total number of hits to view that page would be 5. So although this is perhaps the most common metric quoted in web statistics it is not informative.

Page Views:
Total number of web page requests.Since only requests for actual pages are counted, this is more meaningful than the Hits statistic, and is usually considerably lower than it.

Visits:
Requests recognised as being from the same user (host) within a certain time interval (usually 30 minutes) are included as part of one visit or session. For example, the same user may visit 5 pages on your site within a few minutes. This would count as 5 Page Views but only one Visit.

Unique Visits:
This is the Visits statistic with duplicate hosts removed. It gives the total number of distinct hosts that have visited your site for a given period.

Page Views will be higher than Visits, and Visits will be higher than Unique Visits. Which figure is most important depends on your online objectives. For some sites, getting users to visit single pages is the aim. For others, repeat visits and the "stickiness" factor (how long users stay at your site) are crucial.

There are a few caveats with log files. All logs will contain inaccuracies to some degree e.g. due to browser caching. Do not expect to find an explanation for every difference among reports from other analysers and tracking systems. You can only compare like with like. However, if you want to study your site's effectiveness, you can learn a lot from your log file.