Posted by 0 Comments Others

Dixon Jones, managing director of Receptional Ltd and an expert in web site promotion and traffic tracking, explores the confused world of ‘hits’, ‘unique visits’ and ‘page impressions’.

Anyone who has a web site will be interested in how many visitors it attracts. In fact the frequency of visitors may be the difference between success or failure for the entire business. But what do we mean when we say that someone has visited a site?

The word ‘visitor’ implies one person viewing a series of pages. You might think that figures for visitors are extremely easy to come by and interpret. Unfortunately nothing could be further from the truth. At the birth of the internet one file

equalled one page and a visitor could be defined as a ‘hit’ when that particular page was called up and downloaded onto the viewer’s computer. Now that one page may call up many different files, just recording ‘hits’ will produce totally misleading figures. For instance, 3600 ‘hits’ a day might appear to be a good figure if each ‘hit’

equalled one visitor. However should the web page being monitored call up several different files, then those 3600 ‘hits’ must be divided by the number of files to give the correct number of visitors. So 3600 ‘hits’ on a page which comprises ten separate files, would meant 360 visitors, one tenth of the original figure.

To overcome these difficulties with ‘hits’, some tracking programs use the term ‘page impressions’ which means just what it says. A user downloads a particular web page and the action is counted. This might at first seem a sensible solution until you consider the action of a typical visitor who returns several times to the same page either during one visit or on subsequent visits.

The accuracy of the data will depend upon whether or not the user’s PC is ‘caching’ the files integral to that page, or whether the user clears the cache after each session; in other words, whether the page re-loads each time. The term ‘page impression’ becomes meaningless on framed sites. If a framed page has a separate frame for the header, the left border and the main text area the visitor will record a total of four different ‘page impressions’ rather than one. So again we have a discrepancy.

‘User session’ or ‘unique visit’ is the terminology for counting the number of times a particular user visits the site. This is determined by the visitor’s IP address and thereby solves the problems of repeat visits to pages. It is a far more reliable way of assembling the statistics. However even this method is fraught with anomalies. Not everyone has a unique IP address, and some may use more than one IP address if they use more than one computer or dial in through an ISP. A home user who logs on every night to the same site will be leaving a trail of different IP addresses. In a work situation the company may have a leased line, therefore one IP address, but several different employees might log to the same site each day. These visits can in no way be described as the same ‘user session’ or ‘unique visit’.

So what is the point of monitoring the site if the terminology is so confusing and inaccurate? Well, the value of web site statistics lies in the interpretation of trends rather than concentration on precise figures. Whatever the unit of measurement, if it doubles, then you can make a reasonable assumption that the amount of traffic to the site has doubled.

An off-line promotion – maybe a radio or TV advertisement – might generate significant traffic at specific times on particular days. Surges of activity on the site can be measured to show how well or badly a specific promotional activity has performed. For e-commerce sites ratios of sales against visits will be vital market intelligence.

Individual page monitoring related to dwell time and referring pages can provide an insight into the efficiency of the navigational structure. Similarly lack of visitors to specific pages might highlight complete failures of navigation or incredibly slow loading times of link pages.

Monitoring can also indicate the key words which visitors use to find the site. Frequency of particular key words will provide an indication of which words should be used in titles, descriptions, alt and meta tags, to provide optimum search ability. All statistics will in fact point the way for adjustment of the site and the generation of even more traffic.

The internet is still such a new medium that the rule book is still being written. It is only by studying statistics of current usage that patterns of visitor behaviour can be established, and the medium can be refined and fine-tuned. So assuming that everyone wants more traffic to their web site, good monitoring will continue to be essential as a site research and development tool – regardless of which terminology is used. As the technology advances still further no doubt ways will be found to iron out the anomalies of language and interpretation. But until that day comes it pays to be just a little sceptical of the figures.