Web analytics has always been about monitoring behavior in the internet. An off-site web analysis does analysis on the behavior of common or public internet surfers in order to determine the potential performance of a website. Measures could include such as those that indicate the potential audience a particular website has, its visibility and the amount of buzz it could create.
On the other hand, on-site web analysis deals with web analysis done on a website by its owners. It could be also be done by third party vendors hired by the site owners. This type of web analysis is more geared towards the commercial performance of the site. It measures the success of particular marketing campaigns. For example, on-site web analysis measures conversion rate of landing pages. From such analysis, marketing techniques can be improved so that commercial performance can be improved.
There are two major on-site web analytics. One is the page tagging and the other is web server logfile analysis. While the first requires hiring third party vendors to perform, the latter is quite simpler since web servers do record all transactions anyway. It uses two types of measure namely page views and visits or sessions. Page view is basically the number of request for a particular page. Visits/ Sessions are a series of requests from one unique user which typically expires after a given time of inactivity.
The Visit/Session measure used to be quite effective. But with the emergence of such technologies such as web proxies and search engine spiders and robots, its efficiency was diminished. To combat this problem, cookies were used and spider requests were ignored.