Why are OT numbers lower/ higher than log files and other stats programs?


Why are numbers recorded by Opentracker often lower than numbers recorded in log files & returned by log file analysis programs such as webtrends, hitbox, awstats, etc?


The challenge & answer is to determine where the main discrepancies lie when comparing stats programs, and when comparing a tracking program to the process of log file analysis.

The explanation comes down to a comparison of what is being counted. This is a good example of the origin of the expression 'comparing apples & oranges".

Some people are a bit disappointed, when they first scrutinize their Opentracker results, due to the low numbers. This is especially true if "hits" have been used as a measurement in the past. Technically, a "hit" is part of server-browser dialogue, and does not tell you how many actual people were on a site. Opentracker was built to make "real" numbers more accessible.

Opentracker is a very strict counter (which is an intentional feature of the design). Opentracker only counts human visitors, also known as "human events". There is a large amount of traffic on the internet, an estimated 25% or less of which are "human" clicks. The rest is made up of bots, spiders, automated crawlers, etc.

A second point is that because Opentracker tracks unique visitors over long periods of time, visitors are rarely counted twice. Strict definitions of unique and returning visitors are used.

Here is a link to a short article called tracking-vs-log-analyzers on this subject.