Supermarkets as Supply Depots

The real prize in online grocery is not building new depots, it is turning existing grocery supply into software routed demand. A supermarket already has inventory, refrigeration, labor, and inbound trucks, but store picking and one off delivery runs usually leave too little margin after labor and last mile costs. The better model is to aggregate digital demand across many stores or neighborhoods, then route it through fewer, higher volume fulfillment nodes that buy better, waste less, and deliver denser routes.

• Marketplace grocery proved that stores can act like local inventory nodes, but not that every store should fulfill every order. Instacart scaled by sending shoppers into existing supermarkets, yet the model originally lost about $15 per delivery, which shows how in store picking can overwhelm the convenience of using store shelves as ready made supply.
  2 sacra 3 sacra
• Vertically integrated operators chase the opposite advantage. They pull supply into warehouses or dark stores, cut distributor layers, and use forecasting to reduce spoilage. In this framework, each middle layer can add roughly 20% markup, and higher basket sizes matter because delivery and labor stay mostly fixed per order.
  1 sacra 4 sacra
• The important shift is from store by store demand to pooled demand. Farmstead argued it could get over $80 baskets with about 1,500 SKUs, versus roughly $100 baskets at supermarkets carrying 30,000 products, which suggests online grocery can concentrate around the items customers actually buy most often instead of mirroring the full supermarket aisle by aisle.
  1 sacra

This points toward a hybrid market where grocers keep stores as local demand sensors and pickup points, while more of the actual online volume moves into dedicated fulfillment infrastructure. The winners will be the operators that can concentrate orders, predict demand, and turn scattered store inventory into repeatable, profitable delivery density.