One of the fun things about a rate adaptive ADSL service is that, unlike fixed speed services (500K, 1M, 2M) an adaptive service can, err, adapt!
This means that when there is a fault, instead of just breaking, a line can often adapt to a lower speed. Sometimes much lower.
Of course there are also cases where speeds change anyway due to changes in line characteristics. This can be short term and long term variations. Other lines in the same bundle using ADSL, etc. So some variation is part of what is normal for an adaptive service.
So, if you are sensible, you work out what a new line can do, a maximum stable rate (MSR), and then you set a fault threshold (FTR) below which a sync rate is considered to be a fault based on a straight margin below the MSR.
This gives a nice clear defined rule on the allowable variation of a line sync rate and divides normal from faulty nicely with no argument. A line has been drawn and everyone knows where it is.
Of course, if you are a major telco, you may find this rather inconvenient, and think that maybe it would be a lot easier to just move the goal posts rather than fix the fault. Doing that manually was bad enough, but wouldn't it be even easier to move it automatically when a fault is reported.
Then, of course, you make sure that the agents on the fault desk think that is normal and have never seen the contract or handbook which clearly explains the FTR is set for the life of the service and that sync below the FTR shall be investigated. No, far better to have long arguments with ISPs until the give up.
Of course, for added fun, make it so that if you report a below threshold fault you send engineers that do not understand and just check there is sync (at any rate) and report as right when tested so you can charge £160 for a special engineer (even if a special engineer was not ordered). Just to rub salt in the wound you can have the engineer report what sync speed he got, being below the threshold, and still say right when tested.