The fragmentation of your returns data is not a technology problem. It is not a resourcing problem. It is the direct and predictable output of a supply chain that was built to play a different game. The old game was cost efficiency. The new game is value recovery. The architecture cannot play both at once. The data it produces reflects which game it was designed for.
How the architecture was built
The reverse logistics chain was not designed. It was assembled. A returns platform was added when volume demanded it. A 3PL took on grading when in-house processing became impractical. A carrier network was contracted to manage the physical movement. A liquidation partner was brought in for what did not sell.
That assembly did not happen in a straight line. Growth into new markets does not follow a single logical sequence. It happens at different speeds in different geographies, driven by customer demand, commercial opportunity, and regulatory timing rather than any centralised infrastructure plan. Each new country brought its own carrier relationships, its own grading and processing specialists, its own local liquidation or resale partners. The returns operation in Germany was not built by the same team, at the same time, or against the same priorities as the returns operation in the Netherlands or the UK. The experts in each lane — carrier, 3PL, returns platform, downstream channel — are different organisations in different markets, each with their own systems, their own data standards, and their own definition of what a completed transaction looks like.
Each of those decisions was rational at the time it was made. Each partner built the system they needed to execute their function. The carrier needed to track a parcel. The 3PL needed to record a grade and assign a route. The liquidation partner needed to know how many units they were receiving.
None of them were building systems to answer a question no one had yet asked: what is this item worth right now, in its current condition, in the channel that will take it?
That question was not part of the original brief. The chain was built for the old game: cost efficiency and logistics throughput. Value recovery was never the design objective, so it was never what these systems were built to capture.
What each partner sees, and what they do not
Each partner in a typical reverse logistics chain holds a portion of the picture. The carrier holds movement data: when the parcel was collected, where it travelled, when it arrived. The 3PL holds grading data: what condition the item was in when it was assessed, which tote or channel it was assigned to. The returns platform holds reason-code data: what the consumer said was wrong with the item.
None of these systems talk to each other by design. They record what is needed to run their operation and invoice their client. The data stays where it was created, formatted for the system that created it, structured around the operational question the partner needed to answer rather than any commercial question a brand might want to ask.
The non-uniformity of that data is not a failure. It is entirely expected. Independent businesses build their own systems for their own purposes. They develop their own taxonomies, their own condition hierarchies, their own category structures, their own field names. There is no reason for a carrier’s data model to align with a 3PL’s grading schema, or for a returns platform’s reason codes to map onto a liquidator’s lot classification. Each was built to make sense internally, not to communicate with the others. The only data points that reliably cross organisational boundaries are those where a common structure has been imposed from outside: a product code shared by the brand, or a classification forced by regulation, such as an HS code at customs. Everything else diverges at the boundary between organisations.
So a brand trying to measure its own recovery performance is left collecting data from multiple partners in multiple formats and reconciling it by hand — data that arrives late, rarely agrees, and shares no common item-level structure.
The question the old game never asked
The question that drives value recovery decisions is specific. It is not “what did this cost to process”. It is not “what reason code did the consumer select”. It is: at the moment a routing decision is made, what is the best available channel for this item, given its condition, given current demand in each channel, and given the compliance requirements that now apply to that disposition?
That question requires real-time condition data, channel-level demand signals, compliance status by item type, and a recovery rate benchmark to measure decisions against. It requires those inputs to be connected and available at the point where the routing decision is made.
None of the systems described above were built for this. The grading system records a grade. It does not score that grade against current channel demand. The returns platform records a reason code. It does not link that reason code to the recovery history of items in that condition from that category. The carrier system records a movement. It does not track where the item eventually ended up or what it recovered.
This is not a failure of the individual systems. Each does what it was designed to do. The failure is structural: the information environment was built for the old game. The old game asked what did this cost. The new game asks what is this worth. Those are different questions, and they require different data.
Why this matters now
For most of the last twenty years, the old game was the only game. Returns were a cost line. The primary objective was to process them efficiently and move on. Data fragmentation was a friction rather than a material business problem, because the data the old game needed was the data the chain already produced. If recovery performance was unmeasured, that was unfortunate but it did not change the operating model.
The game has changed. The architecture has not. That gap is now the problem.
First, recovery performance has become a commercial priority rather than a secondary consideration. Returns volumes are rising, margin pressure from primary sales is increasing, and the recoverable value sitting in reverse logistics is large enough to be material on the P&L. Brands need to understand what they are actually recovering and what better routing is worth.
Second, compliance has created a documentation requirement. Under Article 25 of ESPR (Regulation (EU) 2024/1781), the destruction of unsold apparel, clothing accessories and footwear is prohibited from 19 July 2026 — first for large enterprises, with medium-sized enterprises following from 2030. Destruction is permitted only where a specific listed derogation applies, and large enterprises must publicly disclose what they discard and the reasons for it. To rely on a derogation, a brand must be able to show item-level evidence that the derogation conditions are met for that item. That evidence cannot be produced from a fragmented, partner-by-partner data environment.
Third, the information environment has changed. Item-level value prediction at scale is now technically feasible in a way that it was not five years ago. The data to run routing intelligence exists in the chain. It is trapped in the wrong systems and the wrong formats, but it is there. The question is whether it can be connected and applied in time to be useful.
The old architecture cannot play the new game
The most common response to data fragmentation in reverse logistics is to attempt to fix it from within the existing architecture: ask partners to share more data, build a data lake, commission a systems integration project. DS-003 covers why this approach does not resolve the underlying problem. An architecture built for the old game cannot be incrementally adjusted to answer the new game’s questions.
The fragmentation is not a gap that can be filled with better data-sharing agreements. It is the output of systems built around a different question. Fixing it requires connecting to a layer that sits above the partner systems and aggregates the signals those systems produce into a form that can answer the question they were never designed to answer.
That is not a technology transformation programme. It does not require replacing the carrier system, the 3PL WMS, or the returns platform. It requires introducing the connective layer that the chain has never had.
The data exists. It is fragmented because the old game never needed it connected. The new game does. That is where the work starts.
Data Stack series, RMX Recommerce. Related: RI-009 The Old Game is Over · DS-002 The six questions your fragmented data stack cannot answer · DS-004 You cannot fix your returns data stack in time to use it.