Yes IMD would be very nice. One problem is to get high snr and consistency without an anechoic chamber.
Erin presents the IM of raw drivers already, The results are shockingly bad: 1..2..3 mm of excursion and one ends up with a contamination by IM of 3..10%.
But that is only for the motor's contribution alone. Many drivers, even costly and highly regarded ones may show on some cone/surround resonances ten-folds of that value.
Once an enthusiast for 'quality' would hear the IM in isolation, he would never anymore accept the possibility of it to occur; "female voices" ...
The lately introduced "multi-tone" test-signal is of no use, because it is too unspecific. One has to seek for the IM, but which OEM would do so as to promote his product with happily found problems in this field? So, the prob is defined as either a non-issue, or a mythical thing.
Sorry, I forgot the original question?
Ah - why doesn't Erin use the technique deployed for single drivers for speaker-boxes also? Yes, it would, see above, need to seek out for design-specific problems. Hence it wouldn't be the same for all, and is thus considered "unfair"?
I just guess the 420 will show less IM than its competitor from Genelec, judged by the THD graphs.