Using Analytics - Overview
Overview
Web analytics is commonly defined as the measurement, collection, and analysis and reporting of internet data for purposes of understanding and optimising website usage. Its usage inside a network for what may rightly be considered to be solely intranet sites may seem superfluous but such analysis can provide much in the way of useful information. This information can then be used to plan the capacity necessary to provide an optimal user experience.
How does it work?
Some JavaScript is added to the pages on a web server and, whenever, a user connects to the website, their activity can therefore be recorded along with the response time of the web pages that are accessed. This method is referred to as page tagging. The information gathered can then be stored in a database and then analysed to determine trends.
Why choose Page Tagging?
It is the most appropriate method of counting activity as it includes cached content and cached content – whether cached locally by the browser or by an intermediate device such as a content delivery network or a WAN optimiser - can account for up to one third of all page views
As the JavaScript runs locally, it also has access, via the browser to local information such as the screen size and events such as mouse clicks.
What are analytics good for?
Analytics are good for measuring what users do when the access a website. By using click tracking – remember you can track mouse and keyboard events, you can follow the steps that a user takes to complete, say, a purchase on an e-commerce site. That allows you to determine if design criteria, such as 5 clicks to make a purchase, has been met. They can also be helpful, especially in an intranet scenario, to identify congestion on a WAN link by providing historical evidence of poor connection times and to determine which type of content is most accessed.
What are analytics not good for?
Analytics are of little use in determining the cause of server specific performance or configuration issues or even backend problems like connections to a SQL database. Tools such as Httpwatch, Fiddler or Visual Roundtrip Analyzer are more appropriate to this task. Likewise, analytics will not help you determine if your website was, in fact, coded by a raccoon high on Pepsi Max.