The point is, he is making generalizations, not based on "personal experience" with 7 computers, but based on objective data from a sample space of tens of thousands of computers. Yes, such a decision is also motivated by cost and reliability factors. And with so many computers, yes they have figures on support calls, trouble tickets, number of RMAs, etc. Since the user base did not change between the use of HPs and Dells (they didn't fire all employees and hire a whole new set who are just gentler with PCs), it provides an excellent set of data to draw conclusions about laptop reliability between the two manufacturers.