Also, it will have been either a modest variation in caffeine intake, or else a variation in modest intake (e.g. adjusting intake say from 6->5, 2->1 or 1->0). These are people who’ve already stabilized their caffeine intake to not disrupt their lifestyle, and were just adjusting that sometimes to remove the first coffee of the day.
He does seem to, where possible, taste blind and take serious efforts to exclude confounding factors.
If it weren’t for Hoffman doing this sort of thing with grinders I’d have been intensely skeptical that there was anything more than a placebo difference between a fairly decent grinder and a very good one. At least if his videos are to be believed (and I am inclined to), he consistently distinguishes grinders even at a fairly similar price point.