The Returns Crisis Is a Data Problem - And Your Platform Created It
Every year, around this time, the returns headlines arrive. Billions of pounds of goods flowing back through UK carrier networks. Warehouses overwhelmed. Margins eroding. And somewhere in the boardroom, the accepted explanation: customers have become too comfortable returning things. It's a behaviour problem. It's a policy problem. It's a cultural problem. It isn't.
The UK's returns crisis is, at its root, a data problem. And for most mid-market retailers, the platform and systems choices they have made are the infrastructure that created it.
The Scale That Makes This Urgent
UK non-food eCommerce returns represent tens of billions of pounds of goods moving backwards through supply chains that were only ever designed to move forwards. The cost of processing a single returned item - accounting for return shipping, warehouse handling, grading, reworking, and either restocking or writing off - can consume between a fifth and two thirds of the item's original value. For low-to-mid ticket fashion and lifestyle goods, that frequently means processing a return costs more than the margin the sale generated in the first place.
Fashion is the category where the problem is sharpest. UK online fashion return rates run at roughly 30% on average, with some pure-play D2C brands at the upper end of a range that extends to 40%.
That is not a rounding error. It is a structural leak in the P&L that deserves the same analytical rigour as cost of goods, marketing efficiency, or platform pricing.
What Is Actually Causing the Returns
The industry narrative frames returns as a customer behaviour issue - bracket buying, wardrobing, the expectation of free returns, the dissolution of the commitment to purchase. There is some truth in each of those. But the data on why customers return tells a different story.
- Fit and sizing account for somewhere between 70% and 93% of fashion returns, depending on the study and category. "Not as described" - meaning the product differed meaningfully from how it was presented online - is the second major cause. Together, these two categories represent the overwhelming majority of returns. And both of them are directly caused by failures in product data, not failures in customer behaviour.
- A customer who orders a size 12 and receives a garment that fits like a size 10 did not make a frivolous decision. They made a reasonable decision based on inadequate information. A customer who buys a jacket described as "midnight navy" and receives something closer to charcoal is not abusing the returns policy. They are correcting a mismatch between the product page and the physical product.
- Bracket buying - ordering multiple sizes with the intention of returning all but one - is routinely characterised as the problem. It is the symptom. The problem is that a majority of UK shoppers have concluded, through experience, that they cannot trust sizing information to be consistent or accurate enough to commit to a single size. That conclusion is correct. UK clothing has no enforced sizing standard. A size 12 at one retailer will not correspond to a size 12 at another, and within many multi-brand or multi-category catalogues, it will not even correspond to itself across different collections or seasons. When the data cannot be trusted, customers adapt their behaviour accordingly. The cost of that adaptation lands with the retailer.
The Platform That Created the Problem
If the primary causes of returns are data problems - missing, inaccurate, or inconsistent product information - then the infrastructure that manages product data is directly implicated.
Most mid-market eCommerce platforms were not built with returns reduction in mind. They were built to get products live quickly, support high-volume catalogue management, and serve a storefront. The product data quality controls that prevent a garment from going live with an empty size guide, a generic description, or an inconsistent fit classification - those are not default features. They require deliberate architectural choices.
A Product Information Management (PIM) system built into the eCommerce platform - rather than bolted on as a third-party tool - changes the quality guarantee. PIM systems that enforce attribute completeness before publication prevent the "empty size guide" failure mode entirely. A product cannot go live without the required fields populated, validated against schema rules, and approved through a content workflow. The returns that stem from missing sizing data cannot happen if missing sizing data cannot get past the publishing gate.
The evidence from retailers implementing structured PIM solutions is consistent: return rates fall by 15% to 40% as a direct consequence of better product data. Not from changes to return policy. Not from charging for returns. From giving customers accurate, complete, standardised information before they buy.
But even excellent product data cannot fully solve the returns problem if the operational system sitting behind the storefront is disconnected from it.
The Disconnection That Makes Everything Worse
The second major structural cause of avoidable returns is the gap between the eCommerce platform and the warehouse management system. For many mid-market retailers, these are separate systems - managed by different teams, synced through manual exports, CSV files, or batch API updates that run on a delay. That gap is where operational failures live.
When the eCommerce platform and the WMS are disconnected:
Inventory data is always slightly out of date.
Orders are placed against stock levels that may no longer reflect physical reality. Items that have been picked for another order, returned from a customer and sitting unprocessed in a returns bay, or moved in the warehouse without a system update can appear as available when they are not. Overselling - and the cancellations, substitutions, and wrong-item dispatches that follow - becomes a statistical certainty at volume.
Returns processing creates a commercial black hole.
A returned item that cannot be automatically scanned back into available inventory sits in a holding area, physically present but invisible to the storefront. During peak periods - post-Christmas Returnuary, sale clearance windows, Black Friday aftermath - this reconciliation backlog grows faster than manual teams can process it. Stock that customers want to buy cannot be sold because the system does not know it is there. The average returned item spends approximately ten days in limbo from return initiation to resale readiness. Every one of those days is working capital and margin tied up in the reverse supply chain.
Returns data never reaches the people who need it.
Perhaps the most damaging consequence of a disconnected system is what does not happen: the information from returns is captured in a warehouse system that the product and merchandising teams cannot access. The fact that a specific denim jacket is being returned at 45% - and that 80% of those returns cite "runs very small" - never reaches the content team, never informs a description update, never triggers a size guide revision. The same return reason repeats at scale, indefinitely, because there is no feedback loop between what comes back and what is published.
A connected system - where the WMS and eCommerce platform share a single data layer - eliminates these failure modes. Inventory accuracy in integrated warehouse environments with barcode verification regularly exceeds 99.5%, compared to 85–90% in disconnected operations. Returns processed through the WMS immediately update storefront availability. And returns reason data flows back to the teams responsible for the product information that caused those returns in the first place.
Why Remarkable's Architecture Is Different
The distinction between an embedded WMS and a third-party integration is not a technical footnote. It is the structural difference between a system where the eCommerce platform and the warehouse operation share a single source of truth, and one where two separate systems are perpetually trying - and sometimes failing - to stay in sync.
Remarkable Commerce's Warehouse Management System is not a third-party bolt-on connected through middleware. It is built into the same platform that powers the storefront, the order management system, and the customer account. The returns management module processes returns directly within this unified environment - meaning returned stock is immediately available for inventory decisions, returns data is immediately accessible alongside trading data, and there is no lag, no reconciliation queue, and no commercial black hole between the customer return and the resale opportunity.
The inventory management system provides real-time visibility across all locations and channels. Overselling is structurally prevented because the platform knows, in real time, what is physically available. The order management system processes orders as they arrive, reserves stock immediately, and maintains live status that the storefront reflects without delay.
On the product data side, the Product Information Management module provides a centralised, configurable hub for all product attributes - the single source of truth from which the storefront, feeds, and integrations all draw. Size guides, fit descriptors, material composition, colour descriptors, and any other attribute that influences a customer's purchase decision can be governed centrally, with completeness rules enforced before any product goes live.
For retailers whose return rates are being driven by the two dominant causes - sizing and description mismatch - this is the architecture that addresses both simultaneously. Better product data prevents returns from happening. A connected WMS ensures those that do happen are processed fast, fed back into trading intelligence, and converted from a cost centre into a learning loop.
The Question Worth Asking
If your brand is experiencing return rates above 25% - which places you squarely in the mainstream of UK fashion eCommerce, not the outlier category - the question worth asking is not how to change your returns policy. It is where those returns are coming from.
If the majority are attributed to fit, sizing, or "not as described," you have a product data problem. If you are seeing wrong-item or substitution returns at meaningful rates, you have a systems integration problem. If your returns data is not routinely informing your content and buying decisions, you have a feedback loop problem.
All three are solvable. None of them require customers to behave differently. They require the platform and systems that manage product information and warehouse operations to be built for the job.
Make the right platform and system choices
Talk to the Remarkable team about how an embedded WMS and structured product data architecture addresses returns at the source
