Even though his own anecdotal evidence isn't reliable since it's a small sample size, multiple independent sources coming up with similar numbers over much larger sample sizes is pretty good evidence of a huge problem. I'll grant that theirs probably aren't properly set up statistical studies so the numbers could be off by a decent amount, but if they were even remotely close to the acceptable industry average failure rate of 1-2% or even less then you'd never get numbers as high as 50% to 60% if your random sample was even as small as 50 people.
We can argue back and forth all day over what the actual failure rate is, but at the end of the day, the only answer that matters is that it is far too high.





Reply With Quote