"Do things that don't scale"[1] references Stripe specifically for actively recruiting users, but your description of the Atlas program makes it sound like an extension of that strategy. Whereas launching in a single tiny overseas market probably didn't look like it had favorable numbers in absolute terms, they probably got better at opening new markets after each success, and an ever-increasing portfolio would made it increasingly difficult for competitors to imagine ever catching up.

[1] http://paulgraham.com/ds.html