.svg)
.svg)
.svg)
.svg)
.svg)
The simplest way to plan a multi-warehouse network is to dedicate vehicles to locations, this vehicle only ever picks up from warehouse A, that one only from warehouse B. It feels organized, and it avoids the appearance of criss-crossing the map. It's also a rule someone added, not a limit the geography imposes.
The vehicle nearest both warehouses still only gets sent to one of them.
Splitting a fleet by location is a policy, not a routing constraint
A vehicle capable of visiting two warehouses in the same morning, picking up a small load from each, gets blocked from doing so under a location-locked assignment rule, and a second vehicle has to run instead, often mostly empty, to cover the pickup the first one could have absorbed on the way past. Whether combining pickups is the better call depends on real distance and density that day, not a fixed policy either direction. Sometimes keeping vehicles separate genuinely is more efficient. The value is letting the plan decide per instance, not enforcing one rule across every route.
One vehicle, two warehouses, the same morning
Picture a network running under a standing one-vehicle-one-depot policy.
- The rule, vehicle A is assigned to warehouse 1, vehicle B to warehouse 2, as a standing policy to keep planning simple.
- The reality, vehicle A finishes its pickups early and passes within a few minutes of warehouse 2, which has a small remaining pickup of its own.
- The result, vehicle B still has to run, mostly empty, to cover a pickup vehicle A could have absorbed on the way past.
Nothing about vehicle B's run was a mistake given the rule it was following. The rule was the part that didn't reflect where the vehicles were that morning.
One pickup weaved into a route, not one route per pickup
One flower delivery network coordinates deliveries through hundreds of independent local florists, each historically covering its own small slice of a city as a separate, low-density route. Planning pickups from multiple florists into a single connected route looks like this instead of one vehicle per florist:
One route instead of several turned dozens of thin routes into a smaller number of dense ones. That consolidation is where the cost savings came from.
The same consolidation works in either direction, multiple pickups feeding one route, or multiple origins feeding one delivery.
One delivery, two warehouses, no second trip
One big and bulky retailer fulfills orders that often span more than one distribution center, a sofa from one location, a lamp or a rug from another. Planned as one connected pickup-and-delivery route instead of two separate fulfillment trips, the customer receives everything in a single visit. That matters especially in big and bulky delivery, since a second trip doesn't just mean a second parcel, it means a second full crew.
This isn't only about warehouses
The same planning applies beyond warehouse-to-warehouse pickups. A customer return picked up and carried back to a distribution center, or a delivery routed between two points that are neither a depot nor a typical drop-off, fit the same model, pickup and delivery locations planned together, not treated as a special case requiring its own separate logic.
This shows up wherever more than one pickup point serves the same customer base
Grocery networks sourcing from multiple distribution centers feel it in consolidation density. Retailers fulfilling large items from more than one warehouse feel it in delivery experience, one visit instead of two. Networks with any independent pickup points, franchise locations, local partners, regional hubs, feel it in how many separate, thin routes get planned where one connected route would do. The constraint isn't the map. It's whether the plan is allowed to treat every pickup point as part of one problem instead of several.
The takeaway
Check whether vehicles in your network are assigned to a single pickup location by policy, or by what's closest and most efficient that day. If it's policy, the network is likely running more vehicles, and planning more separately, than the geography requires.
