Why retailers lose money between stores without knowing it

In most retail executive committees, network performance is discussed store by store. You look at revenue at Paris-Haussmann, comment on the drop at Lyon-Part-Dieu, worry about margin at Marseille-Bonneveine. Each location is analyzed as an autonomous entity, with its KPIs, P&L, targets. It's reassuring, it's readable, and it's, in large part, missing the real story.

Because in a modern retail network, most of the lost value isn't inside the stores. It's between them. In those tens of thousands of situations where a product selling poorly in one place would have moved at full price elsewhere. Where stock sleeping in a stockroom would have prevented a stockout at the store next door. Where a transfer that didn't happen forced a markdown that shouldn't have existed.

These inter-store losses are massive. They typically represent several points of net margin at the chain level. And most retailers absorb them without even measuring them — because the tools, organizations, and management rituals of the sector are almost all calibrated to steer each store in isolation, never the network as a system.

This article looks straight at this blind zone: where inter-store losses hide, why they stay invisible, and how some retailers are starting to address them.

The network as a system: a change of perspective

To understand what's at stake, you have to start by changing the focal point. A retailer with 200 stores isn't the sum of 200 independent entities. It's a coupled system where every decision taken in one store has, in theory, consequences on the others.

When a product is shipped to Bordeaux instead of Toulouse, it's not just a logistics choice: it's a choice that determines where this product will sell, how fast, at what margin, and what happens when it ends up overstocked or out of stock. When a store keeps a product that doesn't rotate, it's not a local problem: it's a product potentially missing from another store where it would have been a best-seller.

This interdependence is known to every retail leadership team. Yet it's rarely steered as such, for a simple structural reason: tools, processes, and performance indicators were historically built on a per-store logic. The store manager is accountable for their P&L. The merchandiser is graded on category rotation. Supply chain is measured on average service rates. Nobody, in this architecture, is explicitly responsible for inter-store performance — meaning the quality with which the network, as a system, allocates and reallocates its resources over time.

Result: inter-store losses don't appear in any clear reporting. They dilute into global numbers. They get absorbed into "normal markdown," "inevitable stockout," "unfavorable mix." And they keep, month after month, silently eroding net margin.

The four major sources of inter-store losses

When you do the exercise of precisely mapping where these losses sit, you always find the same four categories. Each deserves a detailed look.

Source #1: structural stock-demand imbalance

By far the biggest. On any product in a retail assortment, there are always, at any given moment, stores in under-cover (the product sells faster than available stock) and stores in over-cover (the product sleeps in the stockroom or on the shelf). It's a statistical certainty: consumption behaviors are local, initial allocations are by construction approximate, and gaps widen as the season progresses.

The problem isn't that these imbalances exist — they're inevitable. The problem is the absence of a continuous rebalancing mechanism. In most chains, a product overstocked in Marseille and stocked out in Nice will stay that way for weeks. Nobody sees it, because nobody looks at the network at the right level of granularity. And by the time someone does see it, it's usually too late: the product has already been marked down in Marseille, and the lost sales in Nice are gone for good to the competitor.

At the scale of a 200-store network with 20,000 active SKUs, that's tens of thousands of SKU/store pairs out of balance at any moment. No human team can handle that volume by hand. And until the situation is treated systemically, it keeps costing money every day.

Source #2: inter-store transfers that never happen

Logically, the solution to the previous imbalance would be transfer: ship excess stock from Marseille to Nice before it's too late. In theory, almost every chain has the operational capacity to do it — the logistics flows exist, transfer tools are in the ERP, stores know how to ship.

In practice, these transfers almost never happen. Several reasons.

First, the logistics cost of a transfer is immediately visible and charged to a specific budget, while the cost of inaction (future markdown, lost sale) is diffuse and charged to no one. In any rational organization, this accounting imbalance discourages action.

Next, identifying transfer opportunities requires a cross-view of stock/demand at the SKU/store level, in near real-time, across the entire network. That view doesn't exist in classic tools — it requires cross-referencing ERP, POS, and forecasts, which means aggregation work nobody has time to do every week.

Finally, the store manager who "gives up" stock to another store has no direct incentive to do so. Their P&L will be cut by the potential revenue, with no clear counterpart. Without formalized arbitration rules and compensation mechanisms, transfer remains a voluntary act that runs against local optimization logic.

The result is a massive paradox: most retailers have the logistical capacity to rebalance their network in a few days, but almost never use it. The associated loss — avoidable markdowns on one side, avoidable stockouts on the other — typically weighs between 0.5 and 1.5 points of net margin across the network.

Source #3: uniform markdowns that ignore store dynamics

Third major source of inter-store losses: the still very widespread practice of applying a uniform markdown across the entire network, or across large store clusters. "We mark down the autumn collection at -30% starting week 42" — decision taken centrally, applied everywhere.

The problem is the same decision doesn't have the same consequences across stores. In a store where the product still rotates well, the markdown destroys margin needlessly: it would have sold at full price. In a store where the product is already in big trouble, a -30% markdown is too weak to truly restart sales: you should have gone to -45%.

The uniform markdown is therefore almost always doubly losing. Too strong where it wasn't needed, too weak where you should have hit harder. At the scale of a season, across a network of several hundred stores, this hidden cost of non-localized markdowns regularly amounts to millions of euros — without any standard reporting making it visible.

Why does this practice persist? Not from incompetence, but because personalizing the markdown store by store across tens of thousands of SKUs is simply impossible by hand. Merchandising teams prefer a uniform, imperfect decision over differentiation work they don't have time to do properly.

Source #4: stockouts that could have been avoided

Fourth source: local stockouts on products that are, at the same moment, abundant elsewhere in the network. It's the most infuriating situation for a chain executive — knowing a customer walked away without their product when the stock existed, just not in the right place.

These avoidable stockouts have a direct cost (the lost sale) and an indirect cost (the customer who doesn't come back, or who starts going to the competitor). Industry benchmarks put stockout impact at several points of annual revenue lost — a significant share, maybe the majority, of which could have been prevented by better inter-store orchestration.

Here again, the issue isn't technical. It's a problem of continuous orchestration — the ability to see imbalances as they form and trigger corrective actions before they become irreversible.

Why these losses stay invisible in classic reporting

If the amounts at stake are this large, why don't retail leadership teams see them? The answer lies largely in the very structure of standard reporting in the sector.

Classic dashboards aggregate. They give the network's average service rate, average markdown rate, average sell-through by category. These averages are useful for strategic steering, but they mask by construction inter-store losses. A network showing a 95% average service rate can absolutely have, simultaneously, a product in total stockout in 30 stores and massive overstock in 50 others. The average is fine, the operational performance is catastrophic.

To see inter-store losses, you have to step out of averages and drop down to the SKU/store level. That's where, and only where, the imbalances become visible. But this granularity is rarely exploited in standard steering — not from bad will, but because it represents tens of thousands of lines no one has time to look at one by one.

On top of that, a methodological difficulty: these losses are opportunity costs, not actual costs. An avoidable markdown shows up in the P&L as lost margin, but it's diluted in the mass of normal markdowns. An avoidable stockout translates into an unrealized sale — which, by definition, appears in no number, since it didn't happen. Measuring what didn't occur requires analytical effort that standard tools don't carry.

Result: retail leadership knows vaguely that these losses exist, but can't quantify them, attribute them, or steer them. They're managed by ad-hoc actions — a transfer decided in a meeting, an urgently adjusted markdown — that are never proportional to the real volume of the problem.

The wrong fix: letting stores coordinate among themselves

Faced with this difficulty, some chains have tried a pragmatic approach: let stores coordinate directly. Internal messaging tools, peer-to-peer transfer apps, regional manager communities. The idea sounds appealing — who better than the floor to know what's missing where?

Experience shows this approach quickly hits its limits. Four reasons.

First, it relies on individual willingness. A store manager might flag a stockout, but they're not going to systematically check whether one of the other 199 stores in the network has the product overstocked. Peer-to-peer coordination works for obvious cases, but it misses the essential — the latent imbalances no one perceives as priorities.

Next, it's asymmetrically incentivized. The requesting store has every interest in receiving stock. The giving store has no direct incentive to give. Without formalized rules, this imbalance makes spontaneous transfers rare and conflictual.

Beyond that, it ignores the global optimization dimension. A transfer that seems locally relevant can be sub-optimal at the network level: maybe a third store needs the product even more, or a transfer from the central warehouse would be more economical. Without a systemic view, peer-to-peer coordination produces local optima that don't sum to a global optimum.

Finally, it doesn't scale. On a 200-store network, the number of potentially relevant transfers each week numbers in the hundreds. No human coordination, however well-intentioned, can handle that volume.

The conclusion is clear: inter-store coordination can't rest on the goodwill of the floor. It has to be centrally orchestrated, with the right tools, at the right frequency, on the right criteria.

What it takes to take back control

Stepping out of this blind zone requires a change that's both technical and organizational. Concretely, three conditions must be met.

First condition: a unified, continuous view of the network at the SKU/store level. As long as stock, sales, and forecast data live in unreconciled silos, no serious inter-store optimization is possible. The first brick is therefore a data unification layer, updated in near real-time, usable at the granularity where losses sit.

Second condition: an engine capable of continuously identifying action opportunities. At a network's scale, you need to monitor tens of thousands of SKU/store pairs permanently. No human team can do this by hand. You need an engine that automatically scans the network, identifies imbalances as they form, and formulates for each the appropriate action recommendation — transfer, targeted markdown, priority replenishment, or nothing.

Third condition: execution without breakage. Identifying a transfer opportunity is useless if the decision takes five days to move from dashboard to actual order. The engine must be connected downstream to execution systems — ERP, WMS, pricing — to propagate validated actions without re-entry, with a delay compatible with real commercial dynamics.

Beyond these three technical conditions, an often underestimated organizational one: identify an explicit owner of inter-store performance. As long as this topic belongs to nobody — it's neither in the store manager's scope, nor the merchandiser's, nor supply chain's — it will keep being the blind spot of steering. Chains making progress on this front almost always start by naming this owner, often attached to retail operations leadership, with a clear mandate and dedicated KPIs.

The Solya approach: orchestrating the network continuously

That's precisely the mission Solya carries for retailers who want to take back control of inter-store performance. Not by replacing your existing systems — your ERP, WMS, BI stay in place — but by adding the orchestration layer those systems were never built to carry.

Concretely, Solya connects to your data sources — POS, ERP, e-commerce, supply chain, internal tools — and rebuilds a unified view of the network at the SKU/store level, continuously updated. The decision engine permanently scans that view to identify action opportunities: rebalancing transfers, store-cluster targeted markdowns, priority replenishments, supplier returns. For each opportunity, it formulates a reasoned recommendation, applies your business rules (minimum logistics costs, store constraints, margin floors), and propagates validated decisions into your execution systems without re-entry.

The teams keep their hand on structural trade-offs. They define orchestration rules, validate massive transfers, adjust strategy based on commercial context. Solya takes care of the mechanics: the thousands of small inter-store decisions which, added up, make the difference between a network steered as a system and a network that endures its imbalances.

The result isn't marginal. On chains that have industrialized this orchestration, observed gains typically sit between 1 and 3 points of net margin recovered across the combination of reduced avoidable markdowns, lower stockouts, and better network stock rotation. Not because you sell more products, but because you stop paying the hidden bill of un-steered inter-store performance.

The real question to ask

How many inter-store transfers does your network execute every week? How many markdowns are differentiated by store cluster rather than applied uniformly? How many stockouts could have been prevented by timely reallocation?

If your answers count in the dozens rather than the thousands, you're probably letting most inter-store optimization opportunities slip through. Not out of incompetence — but because no human organization can, alone, handle that volume of decisions on a modern network.

It's that reservoir, immense and structurally invisible in standard reporting, that leading retailers have started addressing. Not by working harder. Not by asking store managers to talk to each other more. But by building the orchestration layer that turns a set of stores into a network actually steered as one.

---

Map your inter-store losses

At Solya, we offer retail leadership teams a personalized 30-minute diagnostic to identify, on your own network, where the main inter-store losses sit — avoidable markdowns, local stockouts, missed transfers — and quantify the margin potential recoverable.

👉 [Book your Solya diagnostic] — 30 minutes, by video, with one of our retail experts.

You'll walk away with:

• A map of the main inter-store loss sources on your network
• A quantified estimate of recoverable margin potential
• The first high-ROI use cases to move from per-store steering to network orchestration