Visitor Intelligence News - April 2001 ======= ============ ==== ===== ==== In this issue: - The strange case of the phantom website visitors - New whitepapers from Site Intelligence - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - | This is Site Intelligence's electronic newsletter. | | Newsletter contents are copyright (c) 2001 Site Intelligence Ltd. | | You are receiving this email because you subscribed via our website:| | http://www.site-intelligence.co.uk/ | | To unsubscribe or to change your subscription details, contact us | | at: newsletter@site-intelligence.co.uk | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - The strange case of the phantom website visitors --- ------- ---- -- --- ------- ------- -------- The recent controversy around e-district's visitor numbers (see, for example, http://www.theregister.co.uk/content/23/17229.html) underlines the importance of accurate visitor tracking. This issue of Visitor Intelligence News takes a look at the technical background to the issue. Much of the activity recorded by a webserver represents "phantom" visits - that is, visits that do not genuinely contribute to the business goals of the website. These phantom visits have to be filtered and ignored to obtain reliable measures. Essentially there are four potential problem areas which can give rise to phantom visitors. We'll consider each in turn, starting with the most basic and ending with the most subtle. - Problem area 1: hits or clicks? - Companies often talk about "hits" to give an idea of how popular a website is. But this can be very misleading. The term "hit" is used in different ways by different people. Most commonly a "hit" means a single request, of any type, received by a webserver. So the action of downloading a homepage which contained 9 graphical images would generate 10 "hits" on the server. With this definition the number of "hits" has very little to do with the number of visitors to a site - one could increase the number of hits simply by adding more graphics to the home page! A better term is "page impressions" - one page impression being a request for a full page, inclusive of any graphics and other embedded items. One page impression closely corresponds to one "click" by a real person visiting the site. Or one "click" by a machine... - Problem area 2: human or machine? - Search engines rely on automated "crawlers" to index website content. These "robots" visit a site periodically and may check every page. This generates a large number of "phantom" page impressions. Most crawlers can be identified easily by their browser type, but some masquerade as a standard browser (IE or Netscape). There are other automated process that generate hits as well - for example, many sites use automated performance monitoring tools which "ping" the site every few minutes. These have to be recognized and filtered. Once robot activity has been excluded, what remains represents visits by real people. But there are still problems... - Problem area 3: visitor or employee? - The activity generated by a company's own employees in testing and maintaining the site may be significant. Many companies have their own site set as default homepage on their employees' browsers - so a "hit" is generated every time a new browser window is opened. These "staff" visits are another type of "phantom" visit. They can be filtered out using IP address. - Problem area 4: genuine or superficial interest? - A final area to consider is whether all visitors are equally valuable. A person who visits the homepage, scans it and leaves the site within 5 seconds is less valuable than a visitor who spends several minutes on the site reading a range of pages. Although visitors with only superficial interest are not strictly "phantoms", smart e-businesses will wish to segment visitors with different levels of interest and to concentrate activity measures on the most valuable visitors. - Putting it all together - Addressing all of these issues to separate real visits from "phantoms" is a complex technical task. Fortunately software systems such as VBIS can automate the process. Conversion from hits to clicks and filtering of robots and employees can be made automatic. Companies making use of of website visitor statistics for management information purposes need to be aware of these issues and should make sure that whatever software systems they have in place are taking the necessary steps to banish the "phantom" visitors. New whitepapers from Site Intelligence --- ----------- ---- ---- ------------ Site Intelligence has recently released two new whitepapers - available free of charge from our website. Visit the library at www.site-intelligence.co.uk to request these whitepapers. The titles are: - "A strategic overview of website visitor intelligence": a useful introduction for all readers, and ideal for busy executives! - "Strategies for website visitor segmentation": how to combine old- and new-economy approaches to segmentation to improve e-business marketing effectiveness. Feedback -------- We aim to make Visitor Intelligence News interesting and informative for all readers with an interest in website visitor intelligence, including but not limited to Site Intelligence's clients. We welcome your feedback on this issue and ideas for articles you would like to see in future issues. Please send comments and suggestions to: newsletter@site-intelligence.co.uk. Our next issue will be in around 6 weeks' time.