Traffic reports & analysis

The world of website analytics is a bit of a murky one, and this section of the DyadStats website is an attempt to shine some light on it.

You would be forgiven for thinking that that only system administrators and bearded techies would have an interest in the contents of server log-files, and to a large degree you'd probably be quite right. However, if your business is online or you use the internet to generate enquiries, you've probably already spent some time looking at where your visitors come from and which search terms or keywords have been used to find your site, and in many cases this will have involved analysing server logs.

Cookie tracking

You can produce a vast array of largely unhelpful and generally unreliable reports from standard web-server logs, and it's quite easy to see details of all the websites that have referred visitors, and produce lists of the keywords that have been used.

There is quite a big problem though. A large proportion of e-commerce transactions do not take please on a visitor's first time time to a website. It's easy to shop around on the Internet and there can easily be 3 or 4 or more visits in between the first visit and the one where the purchase or enquiry happened, and there's no reliable way using log-file analysis to show which search terms specifically lead to actions.

To get round this, many visitor tracking systems issue unique reference numbers and then store this in a database along with a variety of information similar to what you'd find in a log-file.

Pros and cons

Apart from making it possible to analyse which websites, search engines and keywords actually drive enquiries and sales, filtering important metrics into a database with visitor reference numbers has a number of other advantages:

Unlike log-files that record every request for every page and all of its components, tracking systems can happily ignore a lot of the background noise and focus on the important stuff. With information already in a database, producing reports can be much quicker and easier, often this means reports can be 'live', i.e. updated as you look at them.

The accuracy of the reports produced and the reliability of any recommendations they make or infer will vary quite significantly from one system to another, but the same is true of old-school log munching systems too.