Unlike the May numbers from Channel Advisor, Wulfraat just looked at the 20 largest U.S. cities—which are arguably the best candidates for same-day delivery, since they have the highest customer density. Then he measured the distance from each city's epicenter to the closest Amazon warehouse, figuring a generous 100 miles for a DC to be able to do same-day delivery. The results weren't generous to Amazon's same-day plans: Only eight of the 20 cities currently have an Amazon DC within 100 miles.
"That means Amazon today has roughly 45 to 50 million people that they can service on a same-day basis, which represents somewhere in the neighborhood of 14 to 15 percent of the U.S. population," Wulfraat told Supply Chain Digest. "So still a long way to go."
It would take another 12 DCs—some of which are currently under construction—just to get same-day delivery for whatever limited selection of Amazon merchandise could be put in one of Amazon's giant warehouses for the 20 largest U.S. cities. And that would still just cover 37 percent of the U.S. population. Only Amazon knows just how much of its e-commerce customer population is in those 20 cities.
But wait, there's more Amazon will have to do, Wulfraat said. He concludes, based on what Amazon CEO Jeff Bezos has said, that Amazon won't believe same-day makes economic sense until AmazonFresh is part of the service.
That actually makes lots of sense, based on Amazon's experience offering same-day in a limited number of cities as part of Amazon Prime. In practice, while customers seem to like the idea of same-day, they're not using it much for general merchandise. AmazonFresh's grocery delivery, on the other hand, isn't just time-sensitive because people really want that chicken today. The chicken will go bad if it sits on a truck at room temperature for an extra day or two.
The problem, as Wulfraat points out, is that AmazonFresh uses a completely separate distribution network from the one for general merchandise. It has to, because Amazon's regular DCs have their operations set up to get shipments out the door to UPS, FedEx and other carriers on well-understood schedules to do delivery the next day or two days out. Same-day grocery delivery has to send trucks out on schedules that change constantly, while still getting the highest density of deliveries per truck.
(This is what happens when you get these traditional supply-chain guys analyzing e-commerce—always with the facts about actual warehouses, trucks and drivers. It all looks a lot easier on the back of a napkin.)
If Amazon wants AmazonFresh for those 20 biggest U.S. cities, it may not need just another 12 DCS, but an extra 20 DCs beyond that for AmazonFresh. Some existing DCs might be retrofitted for groceries, and the ones not yet built might be designed for both general merchandise and groceries from the beginning. But that might not make sense either, since a regular Amazon DC should be as close as possible to the airport, while an AmazonFresh (or same-day) DC should be as close as possible to town.
All this is possible, especially given Amazon's willingness to build and its indifference to Wall Street's howls about a lack of on-the-books profit. ("You don't like the fact that we plow it all back into warehouses? Fine—go invest in some other e-tailers that are growing the way we are. You know, the ones that don't exist.)
But in practice, just the warehouse requirements mean Amazon won't be ready for a top-20-cities rollout by 2014. It may expand on a city-by-city basis—AmazonFresh should be ready for New York City next year, for example—but exactly how to make the deliveries profitable is still not clear.
As usual, the facts of e-commerce—and retail in general—make wonderful ideas a lot harder to implement in practice. That may mean the Amazon threat is a bit farther off than it appeared.
But it's still there. Logistical realities have bought Amazon's competitors (that is, every major retail chain) a little time.
But not forever.