Keeping track of web usage data

The use of website statistics is extremely valuable and can provide essential data regarding visits to a website and usage, which in turn can help to develop the site’s content and marketing activities. However, the wide range of statistical packages also represent a variety of different methods of recording and analysing online traffic data, plus there can also be discrepancies between this information and statistics provided by online advertising activities, such as pay-per-click advertising or banner adverts.

Every action taken on the web leaves a trail of data that can be harnessed to provide valuable information on the way a website works. There is a wide choice of statistical software packages that will analyse this data and present the results in a user-friendly format – we shall review some of these in a future issue of this newsletter. What is important, however, is that each of these packages will analyse data in different ways, which can lead to discrepancies between figures from different packages and also those reported by advertising tracking tools.

It is generally accepted that measurement errors between advertising tracking and website statistics can be in the range of 10-20%. This is not very helpful when you’re trying to analyse the impact of an advertising campaign accurately, but it is also worth remembering that the data that is provided as part of a website marketing campaign is still much more detailed and accurate than most other media. By analysing the data trends, as well as other activity data that may contribute to the overall return on investment, such information available through the web is still a very powerful and valuable resource.

Ideally you need to get into the workings of a particular statistics package and also any advertising recording software to try to understand and identify the technical issues that may create distortions in the figures. For example, some of the most common factors are the following:

Tracking users by IP address: many users on the Internet may share the same IP address, whether accessing the web through a company server or from one of the large ISPs, such as AOL, Wanadoo or Tiscali. Therefore statistics may understate individual visits as it can appear that the same person has visited more than once.

The use of cookies: cookies are small files that are placed on a user’s computer and are often used by advertising tracking software or by web statistics packages to record and remember activity data. However, depending on the browser security settings or the preferences of users, these cookies may be blocked or deleted, which means that data on visits and repeat visits could be lost.

Cached pages: Internet browsers will often ‘cache’ pages in their memory so that they can sometimes take the fastest route to a web page by delivering these previously visited versions of the page rather than the latest version. When this happens, such pages cannot be measured or tracked effectively.

Onsite tagging: if the tracking software or web statistics package requires tags to be placed on each page of the website, then data can be affected when new pages are added or if pages are amended and tags are not included correctly. In addition, if pages are slow to download and the user clicks on to another page or a different site instead, the tags are often the last part of the page to be loaded and therefore these visits or page views may not be counted correctly.

Click fraud: this is a common concern for advertisers and can happen in a number of ways. There are some controls in place but unscrupulous companies could using software to deliver a large number of false clicks to an advertiser. Unusual traffic patterns need to be monitored and a percentage of fraudulent or automated traffic will always need to be considered within the overall statistics.

So, all of these above issues need to be considered within any statistical analysis, but at the same time the trend data will continue to provide valuable information as usually the same issues will apply on an ongoing basis. Also web activity data is still a very detailed and incomparable level of analysis that should be used as part of any web marketing campaign, albeit with an awareness of the potential issues within the figures.

If you’re using statistical data as part of your website marketing activity and would like help with interpreting this information effectively, please contact us for more information.

Leave a Reply