It's a misunderstanding over the ambiguous term 'up to'
It is not being used in the sense of: You will be able to drive your car at speeds of up to 200mph.
It is being used in the sense of: Your car is capable of being driven at up to 200mph.
The former statement is unlikely for the vast majority of us (good luck finding an opportunity to do that on your daily commute) and any car salesman claiming that is on thin ice. But the latter statement is entirely reasonable. Even if you don't have a driving license and leave the car parked in your garage the statement is still true.
The problem stems from marketing mis-use of the term and customer optimism. The same techniques that snake-oil sellers have used for ages. The trick is simply never to buy anything purely on the basis of marketing material. There's nothing particularly difficult to understand about why DSL is line dependant and all ISPs have to offer an estimate at sign-up.
The 10% solution suggested by the ASA just further confuses the issue in my opinion. The fact that 10% of an ISP's customers get the headline speed means nothing because it is still not considering my particular circumstances. All it's saying is that 10% of an ISP's customers happen to have a good quality line. Whether or not I can get those speeds is purely down to the quality of my line, not anyone else'.
The only time it can matter is if an ISP is cherry picking the lines it accepts but they don't. Most DSL based ISPs have no control over the line quality (KCOM is the only exception) so connection speed will be the same no matter which ISP I go with. Now the ASA or Ofcom could choose to impose rules based on network capacity by forcing ISPs to publish figures for peak time slow down. That at least is under ISP control and is a useful differentiator. Sadly they don't do that (but Thinkbroadband does.
So customers are left with a pointless ASA rule and no practical way to influence the performance of their telephone line.