Obviously, it's impossible to capture all the variables, but in quantifying the good we do, we value all lives the same, and strive to maximize the good we do.
If the EA movement forces big charities to focus a bit more on bang for buck, then that's a win. Perfect efficiency will never be possible anyways.
I have some experience in the nonprofit sector. While you obviously don't want to see people being cavalier with donor dollars, nonprofits that are obsessed with cost cutting are more effective at scolding people for wanting a chair that isn't falling apart or making a few too many photocopies than they are at their missions.
Looking at givewells analysis, they don't seem too focused on office supplies, but rather whether the charity did any follow-up studies, how far it scales and what it costs.
But the follow-up studies seem far more important, than the per unit cost of mosquito nets.
If the EA movement forces big charities to focus a bit more on bang for buck, then that's a win. Perfect efficiency will never be possible anyways.