How do web site statistics work?
Here are definitions and tips for those looking at a reporting software such as WebTrends:
Tips:
Visitor Summary
In WebTrends go straight to this stats grouping, it tells you the best information about your site and does so using Unique Visitors (see definition below). This grouping contains the following information:
Visitors Who Visited Once
Visitors Who Visited More Than Once
Average Visits Per User
Definitions:
Traffic Reporting Software - Software the analyzes a web site to determine how many people visited, how long they stayed, and what they did.
Hits - This is a terribly useless statistic unless you are a cyber-geek. It tells us how many individual files were downloaded from a site including images and supporting files, understanding that a given webpage can vary containing 1 - 1000 images/files. Therefore Hits do not actually indicate much from a marketing perspective. Ignore this statistic.
Page Views - This is the total number of pages that were viewed on a site. Page Views also give us an indication of how many pages an average Unique Visitor viewed. For example if in one month there were 1000 Unique Visitors and 4000 Page Views then we know that each Unique Visitor saw an average of 4 pages over the month. However this does not tell us if each unique visitor saw 4 different pages or if the same page 4 times.
Spider / WebCrawler / Robot - A variety of names are used to describe automated non-human programs that search websites. The name Robot is a good representation because these programs are robots with a single function - Find All Content on Web All Pages. The terms Spider and WebCrawler describe the behavior of these robots as they crawl around your site looking in every nook, cranny, and corner (LINKS) of your website. Typically webmasters like to exclude Robots from reports because they inflate the results somewhat. However it is also very interesting to run reports of Robot activity in order to make sure they are finding everything easily.
Top Content Groups:
Content Groups are groupings of content. Well, maybe that wasn't too helpful. Think of it this way, there are times where we want to count groups of pages together, for example if wanted to count all pages in Grad Admissions including Grad Ed, Grad Psych, Grad MBA, and Grad Religion. In those situations the web reports can be set up to group all those directories together.
Top Directories:
If your website has multiple directories then it is often useful to know which directories are the most popular. For example, Athletics has numerous sport directories including Baseball, Softball, Volleyball, Men's Tennis, Women's Tennis, etc. Top Directories helps you understand which directories are getting the most activity.
Unique Visitor - This is the main term that you want to monitor in web reporting. It tells you how many unique people visited your website. "Unique" signifies that each visitor has only been counted once. *See Waiver.
Visitor - Without the adjective "Unique" this term implies that visitors could be counted more than once.
Visit - This is the number of actual visits to your site. For example, consider a person entering a clothing store. Their visit begins when they enter the door and ends when they exit. On a website it is much harder to calculate a visit, so Traffic Reporting Softwares mark a visit as "over" when the user stops interacting with the site*. Remember that this statistic does not tell us what a person did or how many pages they saw, only that they visited, spending a period of time on our site. In combination with the Unique Visitor statistic Visits can tell us how many times each Unique Visitor visited to our site.
WebTrends - This is a brand of web traffic reporting software.
*Waiver:
Due to the way computers work, there is no perfect system for counting Unique Visits, Visits, Page Views, Hits , etc. The numbers returned by softwares like WebTrends are very good but never perfect. For those interested in a technical definition, the following things can interfere with reporting:
1. Proxy Servers - result in a deflated count
2. IP Address changes - result in an inflated count
3. Internet Cafes - result in a deflated count
4. The same user using multiple computers - results in a deflated count
5. Spiders - result in an inflated count but we can filter spiders out
6. The problem with how we define things. E.g:
If a person gets up in the middle of a visit, eats lunch, and comes back - is that one visit or two visits? If they move from their kitchen computer to the bedroom computer is that one or two visits? Traffic reporting software generally answers these questions by combining the IP address with a time out value. I.e. if a person with the same IP address is inactive for more than 30 minutes then their visit is over.