Summary: first use a method to measure user counts that wrong, then apply voodoo to make the stats seem good, throw in compounding factors that most certainly a

Summary: first use a method to measure user counts that wrong, then apply voodoo to make the stats seem good, throw in compounding factors that most certainly affect the user count, forget about controlling/discussing other factors, and then publish on a high profile blog.