I came across Nielsen's social media report titled "Global Faces and Networked Places" on Çağlar's blog, and had a chance to skim through it briefly.

It's a useful summary of the state of social media today.  However, it's trying too hard to retrospectively explain the success or failure of various social media properties  It delves into the reasons w-k-w is twice as popular as Facebook in Germany, or why Orkut has succeeded in Brazil rather than other properties.

Social media is an exceedingly complex phenomenon to model.  I think all hindsight attempts to explain the behavior of a large social network's participants are examples of us being fooled by randomness.  We crave patterns and invent them where they don't exist.

Nielsen falls in the same trap.