The question before the conclusion.
- Lens
- commerce
- Journey
- 6 decision points
- Reading
- 6 min
The promotion launches on Monday. Posters are printed, prices loaded and communications ready.
On Thursday, a customer compares two stores from the same chain, thirty kilometres apart. The price is not the same. One applies the promotion; the other still has the old price at the till.
No one disobeyed. The decision was made, communicated, and somewhere between the two it was not applied. More importantly, no one at headquarters knew before the customer did.
Deciding is not applying
This is the blind spot of most networks. They measure decisions made—the number of promotions launched, ranges updated and prices changed. They almost never measure the actual rate of implementation.
Yet between the central decision and the till receipt lies an entire chain: distribution of information to the store, its receipt by the local system, the shelf display and the till. Every link can fail silently.
The most common symptom is that everyone at headquarters is convinced the decision applies everywhere because it was sent. Sending is confused with implementation.
One indicator is enough to change the conversation: the proportion of stores where the price charged matches the price decided, measured on launch day. This figure generally comes as a surprise.
Three models, three different authorities
Before any discussion of tools, the network must decide what it really is.
An integrated network. The stores belong to the company and their teams are company employees. Headquarters can impose. The question is technical: does the information arrive and is it applied?
A franchise network. Stores belong to independent entrepreneurs. Headquarters can impose only what the contract provides for. A recommended price is not an imposed price—in many contexts, imposing it would even be unlawful. The question becomes contractual before it is an IT matter.
A mixed network. The most common and most complex: two regimes of authority coexist, often with two levels of information available. Headquarters sees detailed sales from company-owned stores, while franchisee data arrives later and at a lower level of detail.
The classic mistake is to treat a franchisee like a subsidiary. It is not one: it is a commercial partner with its own interests. Anything asked of it must be negotiated, written down and useful to it—otherwise it will provide the contractual minimum, late.
What must be shared, even in a franchise
A useful distinction: what belongs to the brand, and what belongs to store management.
Almost always shared, regardless of the regime:
the item master—codes, descriptions and units. Without a shared foundation, consolidation is impossible;
category vocabulary—to compare like with like;
indicator definitions—a ‘stockout rate’ calculated differently from one store to another means nothing;
the format and frequency for reporting sales and inventory.
Negotiable or local depending on the regime: prices, ranges, opening hours, replenishment policy and team management.
This boundary must be written down. It rarely is, which is what makes discussions with franchisees difficult: each case is negotiated instead of applying a framework.
Store inventory: the figure no one believes
The causes are always the same:
unrecorded unknown shrinkage;
breakages and expired goods discarded without entry;
unchecked receipts—the store receives and shelves goods, then validates everything later, sometimes never;
transfers between stores arranged directly by managers without going through the system;
till sales entered against a generic item when the barcode does not scan.
The last point deserves attention: it is often the largest and least visible contributor. A cashier unable to scan an item who rings it up as ‘miscellaneous’ at an approximate price sends an item out through the till that remains in inventory forever.
No system corrects that alone. What corrects it is making the right entry easier than the wrong one, and making everything entered generically visible.
Replenishment: between two poor answers
Two models conflict, and neither always wins.
The store orders. It knows its customers, local contingencies and calendar. But it orders based on what it sees on the shelves, without visibility of central availability, and often with a safety bias that inflates inventory.
Headquarters pushes. It optimises globally, arbitrates between stores and considers incoming supplies. But it decides using store inventory data that, as we have seen, is approximate.
In our experience, the best-performing model lies between the two: headquarters proposes and the store adjusts within a range. This requires two things—the proposal must be credible, so sales data must return quickly, and the adjustment must be recorded so the network can learn from the deviations.
What we recommend measuring first
Three figures, before any project:
Sales reporting delay—how long between checkout and visibility at headquarters. The entire management capability depends on it.
The implementation rate for central decisions—prices and ranges, measured in the field, not in the central system.
The inventory discrepancy by store—to identify not poor stores, but practices that diverge.
Measuring these three indicators for the first time requires no investment. It requires someone to go and look.
