Operating several stock locations creates a problem that single-site manufacturers never face: you can have plenty of an item and still be short of it. Total network stock looks healthy, and the location that needs it has none. Meanwhile another site is holding six months of cover on the same part.
How Imbalance Develops
Independent replenishment is the usual mechanism. Each location orders to its own reorder point without visibility of what others hold. Nobody is doing anything wrong, and the aggregate result is stock sitting where it isn’t needed.
Local incentives reinforce it. A site manager measured on their own service level will hold generously and resist transferring stock out, because the risk of the transfer is theirs and the benefit is somebody else’s. This is rational behaviour producing a poor network outcome, and it can’t be fixed by asking people to be less rational.
Historical drift adds to it. A large order was placed for a site that has since lost the customer. Nobody reallocated the stock because nobody was looking across locations.
Deciding What to Transfer
A transfer is worth making when the cost of moving stock is less than the expected cost of the imbalance persisting. That means comparing the shortage risk at the receiving site against the transfer cost and the risk created at the sending site.
Some practical rules keep this manageable. Set a minimum imbalance threshold so you’re not moving small quantities at disproportionate cost. Consolidate transfers between the same pair of locations rather than shipping individually. And check whether a normal replenishment would arrive sooner or cheaper than the transfer — sometimes the right answer is to wait.
Allocation When Supply is Short
When incoming stock can’t cover all locations, the allocation rule matters and should be decided in advance rather than negotiated each time.
Proportional allocation by demand is the common default and is reasonable. Allocating by days of cover, so every site ends with similar protection, is usually better — it prevents a site that happens to be low from being starved further. Allocating by customer priority is appropriate where some accounts genuinely matter more, provided that’s an explicit decision rather than whoever shouts loudest.
Making it Work Organisationally
The technical side is the easy part. The hard part is that network-level optimisation requires locations to give up some autonomy, and that needs to be addressed directly.
Measure sites on network service rather than local service alone. Make transfers cost-neutral to the sending location so there’s no penalty for cooperating. And be transparent about allocation decisions — most resistance comes from a suspicion that decisions are arbitrary, and publishing the rule removes it.
Start by simply making the network position visible to everyone. A shared view of who holds what changes behaviour before any policy does, because most imbalance persists through ignorance rather than resistance — sites can’t offer stock they don’t know is needed elsewhere.
ticktick.ai maintains a network view of stock and demand, recommends transfers against cost thresholds, and applies consistent allocation rules during shortage.
