So that wasn't a good example. But I can and do get 40/5 for $79 a month so that still doesn't mean that this set of statistics shouldn't be called "The most speed people are willing their isp's for." Not the fastest isp's.
Unfortunately, Ookla doesn't get terribly specific when it comes to sectioning out the highest-speed users on a given ISP, in a given area. Probably because the sample size would be very small at that point, judging from some speed test data dumps I've seen from them (I bought one awhile back...pretty expensive, but insightful).