Richard MacManus has published some fascinating data from Compete that shows how the social nets are approaching the traffic of the big portals. They found in June, two out of every three Americans visited a social networking site.
I don't doubt that the trend is true. However, the research methodology appears to be flawed. So I question the statistics.
The problem lies in how Compete defines what's a social network. They identify ten in their chart. One of them is Blogger, which is part of Google. So technically they counted some of Google's traffic twice.
::Later: Compete writes in that they only counted Blogger once and did not count Orkut either when figuring Google's footprint
