Web Site Statistics (and log analysis)

Whenever a visitor comes to a web site they leave a mark. The web site interrogates their browser behind the scenes and extracts certain information. This information is stored in a "log file" that should record every possible amount of data on every visitor. From that information you can determine your web site statistics if you use special software to analyse the log files - I use Web Trends..

Among other data, records are obtained of

• number of hits (this is not visits, but the number of files that have been downloaded. For example a page with four photos and five buttons would count as 10 hits)
• number of pages looked at (sometimes called page impressions). You want a high number of page views per visitor. This shows that you are getting "quaity" viewers, not just surfers passing through
• number of visitors
• where they came from (the referring site - useful to see what sends you business)
• if they came from a search engine, then what keywords words they used to find your site)
• lots of other information like the time of day, the browser, the resolution of the browser, the pages they looked at the time they spent on the site.

For a number of reasons the data recorded may have "holes" in it

1. a large percentage of visits are logged as "no referrer"
2. different log analysis tools will give different results depending on their strengths and weaknesses

Let us look at these two "holes"

"No Referrer"

It is infuriating that perhaps half your web site visitors cannot be analysed as no record is oftained of their referring site. I looked at a site with a total of 30,000 user sessions, 17,000 are listed as coming from the same site's index page (which does not tell you much) and 5000 listed as "no referrer". This can be for a number of reasons

1. The visitors browser does not carry the information as the site information was typed directly onto the location window of their browser. Probably not many unless you have a well kown site/brand name
2. Some browsers (eg Monzilla and some AOL browsers) do not carry referrer data so its not there
3. some proxy servers and firewalls strip out the referrer data before transmitting the request for data to your site
4. If the visitor has been to your site before and "book marked" it there will be no referrer
5. A number of big sites "cache" your site data - that is hold it on a tempory file in their own computer, then other users on their system get the info directly from their cache, without visiting your site. The number of your "visitors" lost to your own stats via this method has not been quantified anywhere that I can find

Return to Web Facts page