[DufusM]

I hope it works too - it seems so cool.

But I am not sure this is the big problem in this space.

Current delivery time = 1 day for last mile + >1 day for warehouse -> local area

This method reduces the last mile delivery time down to 30 mins but it is not clear how the second part would be optimized down. At least, without losing the value proposition of Amazon - giant centralized warehouses instead of local retail outlets.

[Turing_Machine]

It's like the levels of caching in a computer.

(just using orders of magnitude for an example, the actual numbers for best efficiency probably aren't quite so nice)

The 1000 most frequently ordered items in a warehouse near your neighborhood. The 10000 most frequently ordered items in a city-wide warehouse. The 100,000 most frequently ordered items in a regional warehouse.

Amazon already has the regional warehouse part.

I'd guess that this might be worth doing even if only the 1,000 most frequently ordered items can get there in 30 minutes, and should definitely be worth doing if you can get the 10,000 most frequently ordered items there in 30 minutes.

[DufusM]

Agree, it could totally work for Amazon Fresh items, say, that are bound to be ordered regularly.

[pilsetnieks]

Plus they don't have to roll out worldwide simultaneously, they can just do a few select cities at first.

[roel_v]

By having more local warehouses that stock the top x most sold things. 100% of the things I can readily remember that I ordered the last 2 or 3 months are so common that they were all in stock in the local shop that I bought most of them at (a large online retailer here in the Netherlands has 6 or 8 or so shops across the country, for local pick up or to try stuff in-store). And they were not the most common things either - a steam wall paper remover, a router (of the woodworking type, not the networking type), a specific type of headphone. If they can have those levels of stock in high street bricks and mortar shops, they could have even larger inventories when they could store them in fully automated warehouses in cheap industrial areas. The accuracy in stock prediction ('shop stock', not 'financial markets stock', obviously) is jaw-dropping.

Man this is going to destroy high street shopping (even further, I might add) - imagine being able to order 10 outfits for trying on from your own home, try them on and send the 9 you don't want back right away, in less time than it would have taken you to just drive to wherever you can buy it off line. If clothing (and other main street) shops are having a hard time now, wait until they have to compete with this...

[sitkack]

I assume they would start to have distributed micro ware houses (shipping containers, dry vans) that are launching points for drones or dudes on scooters. I feel sorry for minimart.

Also, if serious, they should buy 7-11. I still don't understand why Blockbuster wasn't scooped up like hot cakes. All those juicy retail locations.