How do we collect data?
The ratings are created basing on Serbian web sites for which at least a hundred views of particular characteristics are noted. An exception to the rule are the search engines and directories, where the calculations take into account all web sites, regardless of the traffic.
The data is gathered by the means of tracking scripts which are embedded in source codes of the pages enrolled in gemiusTraffic, the collected information is subsequently transferred to a central calculating unit. Web pages monitored by gemiusTraffic act as internet "receptors" – they gather data which, after it is aggregated, constitute a basis for calculating the rankings.
Each time a browser loads a page included in the research, the information on web browser and operating system version, screen resolution and IP of the user's computer (for the purpose geo-location) is automatically sent to the gemiusTraffic server.
Data obtained from the Serbian web sites included in the gemiusTraffic research is divided into two groups: data from internet users connecting from Serbia and from internet users visiting Serbian web sites from outside Serbia. The data generated by the users visiting Serbian web sites from abroad is used only to prepare rankings for countries from which internet users visit Serbian web sites. The data is later calculated and published on gemiusRanking RS.
Type of research: site-centric
While the user-centric model consists in creating user panels with the data used as a basis for statistical analyses are provided by the users themselves, the site-centric model involves creation of web sites panels and information on a user's behaviour is collected by the server hosting the relevant web site.
How is it done? When you go to a web site, either by typing its URL in the address bar of a web browser or through a hyperlink the browser sends a request for the HTML code of the web site to the server on which it is located. As a next step, the elements of the web site (for example graphic files) are loaded. With every such connection, the browser transfers some technical information back to the server - it is this very data that is the basis for the study.
Who is measured?
The group subject to research is composed of all internet users who visit randomly selected Serbian web sites enrolled in the gemiusTraffic.
Visitors are identified basing on the cookie files assigned to their browsers' profiles. This means that a visitor (cookie) does not translate into a real person using the internet. If a person uses more than one browser on one computer, than he/she will be identified as an individual visitor (cookie) with each of such browsers. If a user connects from several different computers, e.g. at school, work, home, etc., in each of these locations he/she will be seen as a separate visitor (cookie). A reverse situation is as well possible, with a few persons using one browser profile in which case their behaviour, computer and browser technical data will be identified as characteristics of a single visitor (cookie).
Noteworthy is the fact that the data being the basis for the rankings are generated by the human internet users only – visits made by web crawlers are excluded.