With one depot, stock discrepancies are at least easy to investigate: the physical stock and the people who manage it are in one place, and when a number looks wrong someone walks twenty metres and counts. The day a second site opens — a depot across town, a branch in the next county — that walk stops working, and a set of problems arrives that has surprisingly little to do with the second site's competence and everything to do with structure: stock now exists in more places than any one person can see, including the invisible place in between.
Distributors rarely plan for this because the second site is opened for good reasons — growth, geography, a big contract — and stock control feels like a solved problem being copied. It is not being copied; it is being multiplied. This article maps what specifically breaks in multi-depot stock visibility, and the small set of structural fixes that keep two sites from being twice the guesswork.
What specifically breaks
- “Stock on hand” stops being one number. Every question now needs a location attached. A system — or a spreadsheet habit — that answers “how many do we have?” with a single total will happily promise a customer stock that is forty miles from their route.
- Transfers create the invisible middle. Goods leave Site A on Tuesday and arrive at Site B on Wednesday. Where were they on Tuesday night? In many operations: nowhere — deducted from one site, not yet added to the other, or worse, still showing at both. In-transit stock that has no state of its own becomes the standing discrepancy every stock take trips over.
- Replenishment becomes a tug-of-war. Two sites drawing on shared purchasing, each defending its own service level, produce hoarding at one and stockouts at the other — while the combined holding quietly rises, because each site buffers against the other's unpredictability.
- Orders pick from the wrong site. A customer near the boundary gets served from whichever site took the order rather than whichever holds the stock or runs the sensible route — and nobody notices, because each site's picture looks locally fine.
- Counts drift apart. Stock takes happen per site, on different days, with different rigour. The consolidated figure is then a sum of snapshots taken at different moments — a total that can be wrong even when every count was locally right.
The structural fixes
1. Make location part of every stock fact
Site-level stock is the first-class record; the company total is derived, never entered. Availability questions — from telesales, from a portal, from a rep — get answered per fulfilling site, so promises match geography.
2. Give transfers a lifecycle
A transfer is dispatched, in transit, then received — three states, each visible, with in-transit stock held as its own auditable inventory state between a recorded dispatch and a counted receipt. Receiving is a real event with a count, not an assumption; discrepancies between dispatched and received surface at the dock, while the trail is a day old, instead of at the quarterly count, when it is folklore.
3. Replenish from combined visibility
Purchasing and inter-site replenishment decisions read both sites' stock, both sites' demand, and everything in transit — which is what dissolves the hoarding equilibrium. The sites stop buffering against each other the day neither has to guess what the other holds.
4. Share the masters, once
One product file, one customer file, one price architecture, maintained centrally and consumed by every site. The alternative — each depot's own spreadsheet dialect — is how the same SKU develops two codes and the same customer two credit positions.
5. Count on a common discipline
The same control standard and reconciliation method at every site — blind counts included — with cadence allowed to vary by SKU velocity, risk and location, so consolidated figures are sums of comparable snapshots, and shrinkage patterns can be compared across sites instead of argued about between them.
Where RouteMagic fits
RouteMagic treats multiple locations as the normal case rather than an add-on to a single-site design. Stock is tracked per warehouse, with inter-site transfers as recorded movements and receiving handled as a real, counted event in the Warehouse App — the transfer lifecycle from the fix list, running on scanners rather than goodwill. Customers, products and pricing live once on the shared spine every site consumes (branch-level cash positions exist too, for the adjacent control). Blind stock takes run per site on the same discipline, and reporting reads both the site-level truth and the consolidated view — so the second depot adds reach without adding a second version of reality.
Conclusion
The second depot does not break stock control because anyone at either site is careless; it breaks it structurally, by creating more places for stock to be than any single view covers — including the in-transit middle that belongs to nobody. The remedies are structural to match: location on every stock fact, transfers with a real lifecycle and a counted receipt, replenishment decided from combined visibility, cash and counts owned per branch on one shared discipline, and master data maintained exactly once. None of this requires the second site to be open yet — the cheapest time to build the structure is before the removal vans arrive. If it is already open, start where the discrepancies concentrate: give transfers their three states this month, and receive against a count. Many recurring discrepancies trace back to transfer handling, and giving that middle stage a name is a large part of the cure.
