The Branch That Already Has It
The purchase order goes out on Tuesday. The same item is sitting in a branch four kilometres away, above what that branch needs, and has been for a month. What the balancing calculation looks at, the one number that makes or breaks it, and why a suggestion is not an instruction.
A branch manager notices a shelf getting thin, raises a purchase request, and it is approved because it is obviously reasonable — the item sells, the shelf is low, the supplier delivers on Thursday. Nobody in that chain did anything wrong.
Four kilometres away, the same item has been sitting at four times what that branch turns over, since a delivery in March that was allocated on a rule nobody now remembers. It is not visible to the person raising the request, it is not visible to the person approving it, and it will not become visible until someone runs a report that nobody has a reason to run.
So the business buys stock it already owns, pays for it twice in working capital, and the overstocked branch stays overstocked until the item is discounted or written off. This is one of the most expensive quiet failures in multi-location retail, and it is entirely a visibility problem.
Why nobody sees it
The information required to spot this is not hard to compute. It is hard to be looking at at the moment the decision is made, which is a different problem.
- The person who feels the shortage cannot see the surplus. A branch manager watches their own shelves. The stock four kilometres away is somebody else's number on somebody else's screen.
- The person who can see everything is not in the room. Head office has the network view and is not present when a branch decides it needs more of something.
- Surplus does not raise its hand. A shortage announces itself — an empty shelf, a customer asking. Overstock sits quietly and costs money the entire time it does.
- Nobody is rewarded for finding it. A branch that gives up stock has less stock. Whatever the network gains, the branch experiences a loss, and people notice their own numbers first.
A shortage announces itself. Surplus sits quietly and costs money the whole time it does. That asymmetry is why one is always noticed and the other never is.
What the balancing calculation actually does
The mechanism is deliberately simple, and being clear about how simple it is matters more than making it sound clever — because you will be deciding whether to act on its output.
-
Group stock by item, warehouse and location
Quantities are summed across batches, so an item held in three batches at one location is one figure rather than three.
-
Mark every location short or surplus
A location holding less than the item's reorder point is short. A location holding more is surplus. That single comparison drives everything that follows.
-
Match shortages against surpluses of the same item
For each short location, look for locations of the same item with stock to spare. A location is only a candidate donor for the amount it holds above its own reorder point — the calculation will not drain a source below its own floor to fill somebody else.
-
Size the move
Move the smaller of what the destination needs and what the source can spare. If one source cannot fill the gap, the next one is asked for the remainder.
-
Do not promise the same stock twice
As each move is allocated, the source's available surplus is reduced in the working figures, so a single overstocked branch is not offered simultaneously to three different destinations.
-
Flag the urgent ones
A destination sitting at zero is marked high priority; everything else is medium. A shelf that is empty now is a different problem from a shelf that will be empty on Friday.
The output is a list: this item, from this location, to this location, this quantity, this urgency. It is available on the transfer screens and through the API, so it can be looked at as part of a weekly rhythm or pulled into whatever the branch team already opens each morning.
The one number everything depends on
Every part of that calculation rests on the reorder point, and there is a characteristic of the reorder point worth understanding clearly before you rely on any of it: there is one per item, not one per location.
A regional hub turning over two hundred units a week and a kiosk turning over eight are both compared against the same figure. If that figure is set for the hub, the kiosk looks permanently and dramatically short. If it is set for the kiosk, the hub looks permanently in surplus and will be suggested as a donor of stock it actually needs.
That is not a subtle bias to be corrected for. In a network with genuinely different branch sizes it is the dominant effect in the output, and it is why these suggestions should be read as a prompt to look rather than as an instruction to act.
The zero case, which behaves in a way people find surprising
An item whose reorder point is zero produces no suggestion at all, in either direction. A location can never hold less than zero, so it never registers as short — and with nothing short, nothing is ever matched against it. This matters because counter-only items compute a reorder point of zero unless you set safety stock, which means the very items a shop sells fastest can be silently absent from the balancing list. Setting safety stock fixes it, and it is worth doing before you conclude the network is well balanced.
What AWRA OpsHub does today
- Every location is compared against the item's reorder point, with quantities summed across batches at each location.
- A donor is never drained below its own floor — only stock above the reorder point at the source is offered.
- Allocations are tracked as they are made, so one surplus location is not promised to several destinations at once.
- Destinations at zero are marked high priority, separating "empty now" from "thin".
- Available on the transfer screens and through the API, so the list can sit inside an existing branch routine.
What it does not do
- Reorder points are one number per item, not per location. Every branch is judged against the same figure, which in a network with genuinely different branch sizes is the largest single influence on what the list contains.
- An item with no reorder point never appears. Both the short and the surplus tests require one to be set, so items you have never configured are invisible to balancing entirely.
- An item whose reorder point is zero produces nothing — nothing can hold less than zero, so it never registers as short. Counter-only items land here unless safety stock is set.
- No cost, distance or lead-time awareness. The calculation compares quantities. Whether moving stock is cheaper than buying it, given the distance and the handling, is a judgement it does not attempt.
Not ours, by choice
- These are suggestions and they stay suggestions. Nothing moves stock automatically — a person creates the transfer, and it goes through the same dispatch, transit, receipt and approval path as any other, because stock that moves without a decision is stock nobody can account for.
- We will not set your reorder points for you. How much cover a branch holds is a working-capital decision with your cost of money and your supplier reliability in it, and a default shipped by a vendor is a guess with a number on it.
Per-location reorder points are scope, not a ceiling — the nightly recalculation already runs on a schedule and already reads the movement history it would need, so scoping it per location is a written specification and a price. Cost- and distance-aware balancing is a larger piece of work and would need your actual transfer costs to be worth anything.
Read the list as a prompt to look rather than a queue to work through, at least until per-location points exist. Its value is that it puts a question in front of a person who was not going to ask it — and that value survives the caveats intact.
Transfer or buy: the decision the list does not make
A suggestion tells you a move is possible. Whether it is the right move is a commercial judgement that depends on things the calculation has no access to, and it is worth being deliberate about it rather than treating every suggestion as an obligation.
Transfer is usually right when
- The stock is genuinely surplus at the source — slow-moving there, fast-moving at the destination.
- The branches are close, or you already run a vehicle between them for other reasons.
- The item is seasonal or short-dated and the source will not sell it in time.
- Working capital is tight and the alternative is buying stock you already own.
- The supplier has a long lead time and the shelf is empty now.
Buying is usually right when
- The source only looks surplus because the shared reorder point is set for a smaller branch.
- The move costs more in handling and transport than the margin on the units justifies.
- The item is fragile, bulky, or awkward enough that moving it twice damages some of it.
- The supplier delivers to both branches anyway, on terms that make the extra order nearly free.
- The source branch will need it back within the month, and you will be moving it twice.
The right-hand column is not an argument against balancing. It is an argument for reading the list with a person's judgement attached, which is exactly why it is a suggestion rather than an automatic transfer.
Making it a habit rather than a report
The failure mode for this feature is not that it produces bad suggestions. It is that nobody opens it, because it is one more screen in a week that already has too many.
What makes it stick
- Give it to one named person with a day of the week attached, rather than to everybody with no day attached.
- Look at it before purchase requests are approved, not after — the whole value is in the sequence.
- Set safety stock on your fast movers first, or the items that matter most will be missing from the list.
- Track how many suggestions were acted on. A list that produces no transfers is either wrong or unread, and those need different fixes.
- When you decline a suggestion, note why. The same one will come back next week, and a reason recorded once saves the deliberation every time.
- Watch which branches are always the source. A branch permanently in surplus has an allocation problem upstream that balancing is treating the symptom of.
That last one is the most valuable thing the list produces over time. A single transfer suggestion is worth a few thousand shillings of working capital. A pattern of the same branch always being the donor tells you that your buying or allocation rule is wrong, and fixing that is worth considerably more than any individual move.
The transfer mechanics themselves are in running multi-branch retail, the reorder point behaviour that drives all of this in the till sells, the reorder point does not hear it, and the underlying maths in reorder point and safety stock.
Our take
Look at the balancing list before you approve purchase requests, not after — the sequence is where the money is. Set safety stock on your fast movers first, because an item with a reorder point of zero produces no suggestion in either direction and those are exactly the items you care about. And read the output knowing that one reorder point is shared across every location, so a large branch will look permanently in surplus and a small one permanently short. The list is a prompt to ask a question nobody was going to ask, which is worth a great deal even before it is precise.
See what your network could move instead of buying
Every location compared against the item's reorder point, donors never drained below their own floor, empty destinations flagged as urgent, and a transfer path with dispatch, transit, receipt and approval behind every move.
Explore stock transfersFrequently asked questions
How does the system decide a branch has surplus?
It compares that location's quantity against the item's reorder point — above it is surplus, below it is short. Only the amount held above the reorder point is offered as available, so a donor is never drained below its own floor to fill somebody else. Quantities are summed across batches at each location, so an item held in three batches counts once. It is a deliberately simple comparison, and knowing exactly how simple matters when you decide whether to act on the output.
Why does a large branch always show as having surplus?
Because there is one reorder point per item rather than one per location, so every branch is compared against the same figure. If that number is set for a smaller outlet, a regional hub will look permanently in surplus and be suggested as a donor of stock it actually needs. In a network with genuinely different branch sizes this is the largest single influence on what the list contains — which is why the suggestions are worth reading as a prompt to look rather than a queue to work through.
Some of our best-selling items never appear in the suggestions. Why?
Almost certainly because their reorder point is zero. A location cannot hold less than zero, so it never registers as short, and with nothing short there is nothing to match against — the item produces no suggestion in either direction. Counter-only items compute a reorder point of zero unless safety stock is set, because measured usage is counted from stock issued rather than from sales. Set safety stock on those items and they will start appearing.
Does it move stock automatically?
No, and it is not going to. The output is a list of possible moves; a person creates the transfer, and it then goes through the same dispatch, in-transit, receipt and approval path as any other transfer. That is deliberate — stock that moves without somebody deciding is stock that nobody can account for later, and the sequence of who dispatched and who received is what makes a transfer auditable at all.
Does it know whether transferring is cheaper than buying?
No. The calculation compares quantities; it has no view of distance, handling cost, vehicle availability or supplier terms. That judgement stays with you, and there are real cases where buying is right — when the branches are far apart and the margin is thin, when the item is fragile enough that moving it twice damages some of it, or when the supplier delivers to both branches on terms that make the extra order nearly free. Treat the list as raising the question, not as answering it.
What should we do if the same branch is always the source?
Treat that as the more valuable finding. An individual transfer saves the working capital on one order; a branch that is permanently in surplus is telling you that your buying or allocation rule for that branch is wrong, and correcting that upstream is worth considerably more than any single move. Balancing is treating the symptom well; the pattern in the list is where the cause shows up.
