Why are you comparing 12 weeks with the entire 2017 revenue? You'd have a better argument that this is peanuts for them if you'd make a fair and clear comparison.
They aren't. They're comparing to revenue per hour. It doesn't matter what period of time you calculate revenue per hour with, you always get the same number.
The overall goal of these lazy estimates is to understand how significant a cost this is for a company like Amazon. Could activists have hoped for more? Are other companies, for whom the costs would be felt quite differently, likely to offer something similar?
In any case, it's not clear that revenue hours is a meaningful comparison anyway because it ignores how much of their cost is labor vs anything else.