100 microvolts of DC offset is spectacularly impressive particularly for a channel that I assume was operating at half the rated maximum power (700W) when this measurement was taken! I'd hate to be the engineer that is responsible for mainting this offset during production. Tom Cumberland is even more of a "slave driver" than I am
I could be dead wrong but I think your estimate of 10 to 20 times for most manufacturers is low. I would think that 15mV (150 times more) is a respectable figure that should not cause any audible speaker non-linearity.
This should certainly help to make any speaker more efficient and full sounding. Nice job.