Putting that aside, I think big chunks of this article are trying to draw conclusions from tiny datasets that might not stand up. For example, the chart showing "Average Number of Employees By Annual Salary" is very clearly just a line connecting the individual data points, that should be a scatter plot with a trend line most likely, and what it shows is that most of the companies consist of only the founding team. Or the bar chart of number of founders vs CEO salary - I think they've sliced that data into so many buckets (4 buckets, 3 groups, N=200) that probably the variance between the groups is just noise.