In Retail, Freight Goes to the Wrong Stores. Every Period.

When freight cost allocation is wrong at the store level, every commercial decision built on store P&L is built on a wrong number.

Retail store P&L is the fundamental unit of commercial decision-making in large retail organizations. Which stores are performing? Which regional formats work? Where should capital be deployed? Which stores should be closed? Every one of these decisions is made on the store P&L. If the store P&L is wrong — not by a rounding error but systematically, because freight costs are being assigned to the wrong locations — the decisions built on it are wrong at the same rate.

In large retail networks, freight cost misallocation is structural and persistent. The problem is not that freight costs are not being tracked. It is that the tracking system — the GL mapping that assigns freight costs to specific stores and cost centers — was configured at some point in the past, does not update dynamically as the store network evolves, and produces systematic misallocation that compounds over time as the gap between the current network and the configuration that was set when the network was smaller grows wider.

How the misallocation works at grocery scale

A major European grocery retailer processing more than 20,000 shipments per week from a network of distribution centers to more than 1,100 stores has a freight allocation problem that is invisible at the individual shipment level and material at the aggregate. A delivery run from the DC in Rotterdam to 12 stores in the Noord-Holland distribution zone generates a single freight invoice from the carrier. The carrier charges for the full delivery run. The TMS assigns the cost to a single store's cost center — whichever store happens to be the primary destination in the routing logic — rather than distributing it proportionally across the 12 stores served.

Across 20,000 shipments per week, the cumulative misallocation is €2 million annually. Some stores appear to have freight costs 15 to 20% higher than they actually do. Others appear cheaper to serve than they are. The regional managers making merchandising and staffing decisions for these stores are working from P&Ls that reflect neither the actual cost to serve their stores nor the correct freight cost basis for their commercial decisions.

Why the GL mapping never gets fixed

The GL mapping that produces this misallocation was typically correct when it was configured — the store network was smaller, the delivery routes were simpler, and the primary destination logic was a reasonable approximation of the actual cost distribution. As the network grew, as new stores were added, as distribution routes were restructured, the underlying delivery reality diverged from the GL mapping. But updating the GL mapping requires someone to identify the discrepancy, document the correct allocation logic, configure the update in the TMS, and validate that it applies correctly across all active routes.

This is not a project that a logistics operations team undertakes proactively. It is the kind of infrastructure maintenance that happens when someone notices a problem — which typically requires a specific store's P&L to look implausibly wrong for long enough that someone investigates. In the meantime, the misallocation continues.

“The store P&L is the source of truth for commercial decisions in retail. When the freight allocation in that P&L is wrong, the source of truth is wrong. Every decision built on it carries the error forward.”

What dynamic GL allocation requires

Accurate freight cost allocation at the store level requires GL coding that is applied at the shipment level — not at the route level or the invoice level — based on the actual stops served, the actual volume or weight delivered to each stop, and the specific cost center that each stop maps to. This coding needs to update dynamically as the store network evolves, as routes are restructured, and as new distribution relationships are added.

The 81% reduction in analyst workload that one major European grocery retailer achieved through automated GL coding was not primarily from processing invoices faster. It was from eliminating the manual correction work that systematic misallocation had made necessary every period. When the allocation is correct from the moment of tender — because the shipment data drives the GL coding, not a static mapping from years ago — the correction work disappears. The analyst hours that were spent reconciling allocation errors become available for the analysis that actually requires human judgment.