Lately I have seen a lot of tech sites talk about the popularity of different sites based on information from ComScore. ComScore is an internet marketing research company and it uses surveyed computers to study online behavior.
This made me enquire into how ComScore was collecting data and how it was monitoring the sites. I also wondered about the sample size that ComScore uses to report to its clients. ComScore is popularly used by a few tech blogs like GigaOm, TechCrunch and hence I wondered if these tech sites have reliably tried to find out how ComScore accumulates its data.
ComScore states that it installs monitoring software on computers with user's consent and tracks their behavior. In return the users get security software, virus-scanning and even cash prizes and goodies. This is the good way of monitoring, but then it makes me wonder how many computers can have this software installed ?? :-0
In the past, ComScore have also used not-so-righteous ways for monitoring user behavior. They have used spyware from DollarRevenue and also had something embedded into iMesh for reporting. They even gave away proxy servers to the world, so that they can be monitored. College dorms, libraries and public internet cafes are popular places where the tracking/monitoring softwares could be installed... This means that obviously sites where college/young students visit like Facebook and Orkut would get higher pageviews.
I have used many different types of web-tracking softwares and was once part of a web tracking application project. During that I realized how impossible it was to create a meaningful and true picture of the net-surfer's behavior.
Cookies distributed through popular sites is one way to track users, but is an expensive affair. Google easily does it due to its rule over the web and nearly everyone using analytics or urchin gives Google all the data. I'm wondering when Google would start selling its tracking data and become an internet marketing research provider!!