I don't doubt their execution, but its domain(diapers.com) might have helped big time in their success.
The domain grab really was like the land grab of the west in the US, first come first serve and if you defended your property then you're now a millionaire!
"We built software with computational algorithms to determine what the optimal number of boxes to have in the warehouse is and what the sizes of those boxes should be. Should we stock five different kinds of boxes to ship product in? Twenty kinds? Fifty kinds? And what size should those boxes be? Right now, it's 23 box sizes, given what we sell, in order to minimize the cost of dunnage (those little plastic air-filled bags or peanuts), the cost of corrugated boxes, and the cost of shipping. We rerun the simulation every quarter. Using the right box probably adds close to 1 margin point."
1% is incredible - that works out to being 890K+/year (though that might depend on what their "null" case is/was).
Anyhow, my point is that rather than stunning innovation, that sounds rather like Packaging 101. However, I'm sure that's because that's the summary for the journalists and I wouldn't be surprised that they are doing actually innovative stuff beyond a journalist's grasp... because as we should all know by now, getting past a journalist's level of knowledge is not a challenge.