Comment by edwin

We looked at the same data that Rand Fishkin used and definitely came to a different conclusion.