For consumer orders, Amazon’s logistics are significantly ahead of Walmart’s.
Walmart built systems optimized for a traditional retail environment:
Regional DC’s to support the flow of imported inventory
More DC’s to support stores
Focus more on cross-dock to get inventory to the stores (vs storage at DC’s)
Truck deliveries to stores
Manufacturer held inventory (to save DC costs) that can be part of pickup and delivery routes
These types of systems involve movement of inventory in relatively large quantities like cases and pallets. When I says systems, think conveyor and other types of expensive automation. All built for a specific model of operation.
Amazon, on the other hand, optimized for a consumer/online environment:
DC’s that store inventory (vs cross-dock)
Inventory storage that allows for piece picking (vs cases and pallets)
Piece picking and packing systems
Walmart’s (and anyone else’s) hardware and software in the DC to support distribution to the retail stores is about 75% useless when trying to efficiently pick and pack large volumes of consumer orders.
To support large volumes of consumer orders, especially when you have a large sku count (e.g. millions) with possibly low units per sku per day (e.g. only one of sku X shipped today), you have to solve some tricky problems.
1 - Picking
Naively picking order by order would be incredibly inefficient and would cause the picker to have to walk all over the DC many times.
The typical solution is some sort of batched picking (group parts of orders together) with a sort operation (marry up the pieces for each order into it’s own container). There are multiple options, but, for high volume, they involve expensive equipment and software geared towards that specific problem.
2 - Packing
To ship to consumer, each individual order needs to be placed in an individual shipping carton with a shipping label (e.g. UPS, FedEx).
This operation is an order of magnitude more labor intensive than just taking a case from receiving or from reserve and shipping it back out to the store.
There is some automation that can be used here, like auto-baggers, but that doesn’t work for everything.
Amazon has multiple levels of traditional optimization related to the picking problem (batched picking, sorters, putwalls, etc.) in use at their DC’s which the robots are slowly replacing. The robots basically bring the inventory to the packers.
From a consumer distribution standpoint, Walmart is retooling to solve this new type of problem. Last I saw they were projected to have about 8 e-commerce fulfillment centers by end of 2017. Compare that to over 100 for Amazon in the US (plus maybe 50 in rest of world).