At nine on a Thursday evening, a shopper orders a wool coat in a medium. The order system looks for the nearest store with one in stock. It finds one in the flagship, three kilometres from her flat. On Friday morning someone in that store lifts the last medium off the rail and packs it. Across town, a smaller store has four of the same coat in the same size. They have hung there since October, and in January they will be marked down.
The coat arrives on Saturday, a day ahead of the promise. The routing rule has done exactly what it was written to do. The flagship then spends its busiest day of the week without a medium in its best-selling coat. Meanwhile the quieter store keeps four coats it will eventually sell for less.
None of that appears on a delivery dashboard, so nobody in the delivery review asks about it.
In short: Routing online orders to the nearest store with stock shortens the last few kilometres. That is the part of the order that is easiest to measure. The rule ignores what the unit is worth where it sits and what the store's team is doing at that hour. It also ignores whether the rest of the basket is there. A better rule sends each order from wherever the unit is least valuable to keep. That holds as long as the basket stays whole and the store has the hands to pick it. Target's stores fulfil the large majority of its online orders. Even so, it has started concentrating shipping in fewer stores and easing it where foot traffic is heaviest.
Mostly because it is easy to explain. When a retailer first opens its stores to online orders, the hard problem is knowing what is in them. Once the stock figures are good enough to trust, several stores may show the same item. Distance is the obvious way to choose between them. It also lines up neatly with the delivery promise. Carriers charge by zone and couriers plan by postcode. On every report, a shorter trip looks faster and less costly.
The rule also rests on an assumption nobody writes down: that every unit of the same product is the same. To the shopper it is, since a medium coat is a medium coat. To the retailer it is not. The same coat in two stores is worth a different amount of money. That depends on where it would otherwise have sold, at what price and how soon.
The map shows distance. It does not show the four things that decide whether an order was worth taking from that store.
What the unit is worth where it sits. Picture a flagship that sells its full size range at full price. A coat on its rail is worth more than the same coat in a store where it has hung for nine weeks. Taking the flagship's last medium saves a few kilometres and gives away a full-price sale. Taking an ageing unit from the quieter store does the opposite. It clears stock that was heading for markdown.
What the store's team is doing at that hour. Online picks land on the same people who serve the shop floor. The busiest stores are often busy at exactly the hours online orders peak. Send orders to the flagship on a Saturday afternoon and the store has a choice to make. It can serve the customer in front of it or the one on the screen.
Whether the rest of the basket is there. Say the nearest store holds two of the three items in an order. The third comes from somewhere else, and the shopper receives two parcels. Each extra parcel is another label, another delivery and another doorstep. Target told investors in November 2024 that it had made changes to reduce split shipments. Those changes raised the number of units in each package. In its words, they meant "reducing last mile delivery costs as well".
How near is measured. Kilometres are not hours. Suppose the nearest store's courier collection has already gone. For today's orders, it is further away than a store an hour's drive off whose van is still waiting. Borders stretch the gap too: a store in Geneva may be nearer a shopper in Annecy than the French warehouse, yet slower once the parcel clears customs.
Few large retailers are as committed to fulfilling from stores as Target. The company's own fact sheet put the share of digital orders fulfilled by its stores at 95% in April 2024. Its latest annual report says stores still fulfil most of the sales that start online.
Even so, it no longer treats every store as an equal candidate. In September 2025 Gretchen McCarthy, Target's chief supply chain and logistics officer, described a test in Chicago. The company concentrated a higher share of shipping volume into six stores. It moved shipping out of 18 others and sent more volume through its fulfilment centres. Stores with heavy foot traffic, she said, "can reduce shipping operations to protect the walk-in guest experience".
Orders are now assigned on real-time signals, such as where the inventory sits and how each store is staffed. They go to wherever there is "the right availability and cost structure". In the test, Target reported shipping almost a full day faster. Local fulfilment cost per unit, it said, was significantly lower than before the test. She put it in one line: "Stores will continue to fulfill the majority of online orders, but which stores may change."
Target's scale and sortation centres are its own, but the choice it made depends on neither.
There are four common answers, and each one is right for somebody.
None of these has to be permanent. Starting with the nearest store is a sensible way to get the service live. The trouble starts when the launch rule is still running two years later. By then the estate, the range and the order volumes have all moved on.
The rules that hold up look something like this, with thresholds set by each retailer.
Each rule also needs a reason a store manager would accept. A rule that reads as the website raiding the stores gets quietly worked around, however well it was designed. Credit is usually where it starts. When a quieter store packs a coat for a shopper near the flagship, someone has to decide whose sales the order counts towards. If the answer is neither store, both managers have a reason to resist. The stock floor raises the same question of ownership. It is a merchandising decision, not a system setting. Someone has to decide what each store must keep to sell at full price, and own that number. Retail supply chain leaders at a roundtable we hosted said much the same about delivery: one size fits nobody.
Speed will not tell you, since the nearest-store rule is good at speed, but these measures will:
A first test needs none of this. Take last month's online orders from your three busiest stores. For each one, check whether a quieter store held the same item in the same size on the same day. The share that could have gone elsewhere tells you how much this conversation is worth.
Our Fulfilment platform sits on the retailer's existing systems and decides which store, dark store or warehouse fulfils each order. The rules stay the retailer's to set, and the platform makes every order follow them.
Distance should be one input, not the rule. Weigh what the unit is worth where it sits and how busy the store is at that hour. Check whether the whole basket can ship together and which location makes today's courier cut-off. Each store also keeps a floor of stock for its own shoppers.
Neither should be the fixed default. The nearest store is fast but can empty the best stores and split baskets. The warehouse protects the stores but is slow away from the big cities. Rules that weigh value, capacity and whole baskets choose between them order by order.
A split basket means more parcels to pack and deliver, and the shopper receives the order in pieces. Target told investors in November 2024 that reducing split shipments raised the units in each package. It said this also reduced its last-mile delivery costs.
Not necessarily. In a 2025 test in Chicago, Target concentrated shipping volume into six stores and moved it out of 18 others. It also sent more volume through its fulfilment centres. Target reported shipping almost a full day faster, with lower local fulfilment cost per unit.