Digital sales growth has become the headline metric that retail investors and leadership teams track most closely. Quarter after quarter, e-commerce revenue figures appear in earnings calls as evidence of strategic progress. What appears less frequently, and receives considerably less attention, is what that growth is actually costing. Albertsons Companies reported a gross margin rate of 27.4% in Q3 fiscal 2025, down 50 basis points from the prior year. The company's own reporting attributed the compression to two primary factors: increased delivery and handling costs associated with digital sales growth, and strong growth in pharmacy sales, which carries a structurally lower gross margin rate. Revenue grew. Margins fell. The faster the digital channel expanded, the more it cost per unit to fulfil. This is one earnings report from one retailer, but the dynamic it describes is widespread across the sector.
The fulfilment cost that compounds
When a product sells in a physical store, the fulfilment economics are relatively simple. The retailer carries inventory, the shopper selects and takes it home. When the same product sells online, it moves through a different cost structure: pick and pack operations, last-mile delivery, and in a significant proportion of cases, a returns process that can cost as much as the original fulfilment. In categories with high return rates, particularly fashion and footwear, the economics of digital fulfilment can turn a margin-positive product into a margin-negative transaction once the full cost cycle is accounted for. The product may carry a 40% gross margin on paper. Once returns, reprocessing, and secondary sale discounts are included, that margin compresses significantly. The underlying issue is that most assortment decisions are still made against traditional margin calculations that reflect physical store economics. The cost assumptions built into a category plan, the pricing logic, the depth of buy, are often calibrated for a world where most units sell from a shelf rather than a fulfilment centre.
ASOS and the infrastructure reckoning
ASOS provides a sharper illustration of how this plays out at scale. In January 2025, as part of its ongoing push to recover profitability, the company announced the closure of its Atlanta distribution centre — a decision that came with a largely non-cash adjusting item of approximately £180 million in its H1 FY25 results. Going forward, US orders will be fulfilled primarily from ASOS's automated UK facility, supplemented by a smaller local US site and the company's expanding "Partner Fulfils" programme. The closure reflects a fundamental reassessment of how much fulfilment infrastructure a digital-first fashion retailer actually needs when the priority is margin recovery rather than volume growth. The strategic shift is clear: carry less inventory in fewer locations, fulfil faster from centralised infrastructure, reduce the cost base. But the underlying challenge that created the need for that restructuring is one of assortment and buying discipline. Fulfilment centres become oversized when retailers buy too broadly, carry too many SKUs, and hold inventory that moves slowly or generates high return rates. The infrastructure problem is a downstream consequence of assortment decisions made upstream.
Where the problem actually starts
The margin drag from digital growth is real, but it is largely avoidable when assortment decisions account for the full economics of how products sell across channels. A product bought in excess depth because historical sell-through rates looked strong generates a different outcome when that inventory sits in a fulfilment centre at carrying cost for six months before clearance. The retailers that are managing digital growth without the margin compression that Albertsons and others are reporting have made one common adjustment: they have connected assortment planning to channel economics. Which products sell well online with low return rates? Which categories are structurally better suited to physical retail? Where is the depth of buy creating inventory risk rather than service level confidence? These questions require a different kind of analysis than traditional category management processes were built to answer. They require data that connects product-level performance across channels, return rates, fulfilment costs, and competitive pricing in one view, so that buying decisions reflect the full margin impact of how products actually behave in a multichannel environment.
The assortment lever that most digital strategies overlook
The conversation in retail tends to frame digital margin pressure as a logistics and fulfilment problem. Better warehouse automation, smarter routing, lower last-mile costs. These improvements matter. But they address the cost of moving products that have already been bought and listed. The more powerful intervention happens earlier, at the point where the assortment is being built. Deciding which products to carry, at what depth, in which channels, is where the margin profile of the digital business gets set. Improving fulfilment efficiency on a poorly constructed assortment recovers some margin. Building the right assortment in the first place avoids the cost entirely. Albertsons grew its digital sales and compressed its gross margin simultaneously. That combination is a signal that the cost structure of the digital channel — both in fulfilment and in the mix of categories driving growth — has not yet been fully absorbed into assortment and buying logic. Closing that gap does not require a logistics transformation. It requires better decisions about which products to put into the system in the first place.
The retailers posting strong digital margins are not necessarily the ones with the most sophisticated fulfilment operations. They are the ones whose category teams understand, with precision, which products belong in their digital assortment and at what quantities - and if you feel like this is something your team could do better: try the new world of AI Agents.