Actually, once upon a time when Amazon actually had coherent strategic focus, inventory obsolescence was a clear competitive advantage. With physical books (before digital offset printing), the setup costs for printing runs were themselves a significant fraction of the TCO for books. So, printers often produced WAY too much, creating excess inventory issues (and a deadweight loss for the book industry). Traditional inventory optimization should make it obvious that there was a clear advantage on the "tail" (the 80% of books that comprise less than 20% of sales), Amazon gained huge advantage by centralizing inventory of physical books.
Because of the shift in products/categories and in manufacturing economics over time, the value creation from more centralized inventory has diminished. That's part of the reason that Amazon has built so many fulfillment centers. But, it still doesn't change the last mile economics, particularly for lower value density items (see today's Mad Money for an explanation from a successful mall developer/operator's perspective).
The math is only going to get worse, not better... Don't get greedy and make sure you can see this through. Within a few months, Bezos is going to have some serious explaining to do because it will be time for the "story" to deliver.
I still maintain that Sell Side analysts should be required to have 100% of their bonus pool invested in the companies they cover (allocated to each) and be unable to reallocate without publicly changing ratings and price targets first. You'd see completely different analysis... (assuming they actually are as good as they claim and get paid to be...).