Domain name and site hosting preferences: empirical evidence

At the Multimedia University, Malaysia, a first‐year undergraduate course in electronic commerce was given the task of evaluating commercial Web sites. The existence of the host was verified when it was added to the sample, and again six months later. The sample was then analysed for correlation between domain names and regional allocation of IP addresses. Three characteristics of the sample stand out. First, in many cases the top‐level domain name does not correspond to the location of the host; there is a noticeable bias toward hosting sites with regional domain names in North America. Second, dot COM and certain regional domains seem to dominate the Internet commercial landscape. Finally, there is a measurable rate of attrition over time, which may or may not be truly significant.