Share
Web statistics such as Google Analytics and Nedstat are important instruments for measuring aspects such as the success of your website or the effectiveness of your marketing campaigns. But the differences in the measurements produced by the packages often give rise to discussions about the reliability of the data.
In March 2009, TNO completed a project, the purpose of which was to assess the differences in web statistics and find out how they arise. An extensive benchmark study involving major Dutch websites and a controlled test environment was carried out to establish what degree of variance in website statistics is acceptable. Bart Gijsen of TNO will present the results of the project on 25 March.
In the second half of 2008, in co-operation with online marketing experts from Blue Mango, ClickValue, Maximum and Netprofiler, TNO launched a project designed to analyse the differences in and measurements of website statistics. This involved examining web-statistics data together with STIR and DART data from various websites. Participants in the study included Agis, Ilse Media, Typhone, TNO and Univé Verzekeringen.
Tests were also carried out with various packages in a closed website environment. Among other things, the results provide concrete guidelines for determining whether the differences in statistics for a website fall within an acceptable range, or whether there is a reason to doubt the statistics.
TNO will present the research results at the next meeting of the Web Analytics Association Nederland. The meeting will be held on Wednesday 25 March 2009 at LBi Lostboys in Amsterdam. There is no admission charge and everyone is welcome. For more information, visit the WAA website (see link below) or contact Martijn Staal.
