When a circuit length is being discussed I think it is useful to think of both the physical length and also the electrical length. Depending upon the condition of the pair, the joints, the metal of which the wire is constructed, etc, the latter can be significantly greater than the former.
The physical length can be measured, either when the cable is installed or by subsequent processes which include estimations of the extra length in footway joint boxes, up/down the height of any pole, etc.
Bald_Eagle1 has first hand, practical experience of such quantification. The end result is a figure in metres, yards, kilometres or miles which can be easily compared with another circuit.
The electrical length is far more difficult to quantify. We have been doing so, somewhat crudely, for many years . . . we look at the attenuation, in decibels, for both downstream and upstream. The problem is that there is no hard and fast rule how those to figures should be obtained. With a circuit that carries a VDSL2 service we can (with suitable hardware and software manipulation) look at (say) a Hlog plot. We all, probably, are aware of what a Hlog plot looks like, for both circuits that are generally considered to be good and those that are deemed to be bad. The visual representation that we see when considering a Hlog plot could be regarded as yet another "handle" on the electrical length of a circuit. Consider the area under the curve. The smaller the area under the curve, the greater the electrical length of the circuit. The greater the area under the curve, the smaller the electrical length of the circuit. I.e. The area under the curve of a circuit's Hlog plot is inversely proportional to the circuit's electrical length.
My apologies for the above nebulous caterwauling. I have had numerous ideas on the subject going back over a number of years. The problem is having a sufficiently large set of data to use as experimental input. Some of you may remember that I asked for specific data from VDSL2 circuits and I was gifted with fifteen sets of data. Manipulation and munging of those data sets did not produce any usable result. My thought then was that (1) the data set was too small (2) the manipulation of the data was incorrect. Since that time I have occasionally thought about what information could be deduced with a sufficiently large data set that was manipulated in other ways. I currently have scribblings on a few sheets of A4 paper as thoughts come (& go) . . .