Oricon manga reports tend to fall well, well short of the “actual” values reported by publishers. I’ve written before about one of the most extreme cases (Yuruyuri’s one million copies with 0 weeks in the charts), but there are plenty of others out there. This isn’t a comprehensive look at it (and these series may or may not represent “average” cases of underreporting); it’s just me picking a couple confirmations and counting up the total (i.e. total on weekly charts from last week a volume appeared) Oricon reported sales up to the point when the publisher reported a given number.
Some of these may simply be unsold copies, another part may represent copies sold at a 10k/week rate in the long-tail shadow of the charts, and a few may come from the overseas sales Oricon chooses not to count. How to separate those three is anybody’s guess. I’m not trying to do that, just give some idea of how big these gaps can be.
Animetics 