The article describes a mix research model named MOSS used in Slovenia for in­ternet audience measurement. MOSS is a commercial project that was performed in 2005 as an experimental measurement with the aim to define a methodology for the future measurements. In the period from 2006 and 2008, the project will be carried out twice a year.

Data collection methods are client initiated log files collected on independent ser­ver (automatic measurement), computer assisted telephone interviewing and web self administered interviewing. The main purpose of the study is to esti mate and correct errors caused by simply using log files for audience estimation.

Major Slovenian sites (53 in 2005) were automatically monitored for 1 mo nth and CATI was conducted simultaneously on a random sample of population aged between 10 and 75 years (N = 1028). The survey addressed the following issu­es: internet usage frequency, geography of usage (various locations), number of computers a person is using, number of persons using internet on the same computer at home, and attitudes towards cookies. Data collected by web survey (N = 4.442 in 2005) were used to define profiles of a siteis consumers.

ln spring 2005 58% Slovenians were using internet on at least monthly basis. Among them 37% were deleting cookies. Automatically 225 millions of impres­sions connected with 1.4 million of different cookies and half a million browsers with out cookies were recorded. On the basis of CATI data the cookie weight was figured to finally obtain audience estimations from automatic measurement data. ln the spring 2006 measurement we tried to rectify some calculation defects of the experimental measurement. We included the profile and browser usage in CATI survey and excluded the extreme behaviour from the automatic traffic measure­ment. In the future we will try to solve the problem of cookie deletion, by reason of which the number of automatically measured visitors is overestimated. However, we will not be able to remedy all defects of the automatic measurement and new limitations of the automatic measurements may be expected to arise.