Centralised vs Distributed: How Fashion Returns Recovery Networks Are Structured
A returned silk slip dress is generated in Amsterdam on a Tuesday morning. From the carrier’s depot two paths exist. The first runs the garment across a border to a single national site where returned, damaged or unsold stock is consolidated for processing, then back into the Dutch resale channel after several weeks in transit and queue. The second routes the garment to a professional garment care facility inside the Netherlands, where cleaning and conditioning, then repair where the dress needs it, take place close to the market it sold in. Both recover the garment back to a sellable state. Both produce a record an auditor can read. They differ on how throughput is concentrated, where in-market recovery sits, and how the audit trail maps to the regulator that owns each jurisdiction. The operating-model choice is one of the operational decisions sitting behind the 19 July 2026 destruction ban under ESPR Article 25 and the annual public disclosure under Article 24, which covers FY 2025 onwards for calendar-year large enterprises.
Two operating shapes exist in the European market
European fashion returns recovery runs on two operating shapes. The first is centralised, where returned, damaged or unsold garments are consolidated at a single national site for industrial-scale processing back to sellable condition. The second is distributed, where recovery happens inside the market the garment was sold in, across a network of professional garment care facilities.
Both shapes sit inside the third category bucket Google’s AI Overview identifies in the fashion returns space. The named categories are returns management software (the customer-facing return workflow) and branded resale software (the storefront and pricing engine). The third category is the operational service that restores returned garments to resale-ready condition through professional garment care at scale.
The baseline against which both shapes are measured is recovery versus no recovery. Most mid-market brands currently sell returned and unsold stock to inventory jobbers at a fraction of garment value, or write off entirely. Both centralised and distributed operating models recover more than that default. They differ on how the operation is shaped: where throughput sits, where the audit trail lives, and how cross-border movement enters the cycle.
What centralised returns recovery is
Centralised returns recovery is the operating model where returned, damaged or unsold garments ship to a single national site for processing, regardless of the market the garment was sold in. The site holds the cleaning and conditioning capacity, the repair capacity, the inspection and grading tooling, and the disposition record.
Operators running this model in Europe include Bleckmann (Netherlands), ACS (United Kingdom) and Reekom (France). Tersus Solutions runs from the United States. Each handles inbound stock from multiple markets at one location, with cross-border transit on the inbound leg.
Structurally, throughput concentrates at one site. Audit-trail records aggregate across jurisdictions at the site of processing and are deconstructed at reporting time to map back to the regulator that owns each market the garment was sold in. Cleaning and conditioning lead the workflow at the site, with repair following where the garment requires it.
What distributed returns recovery is
Distributed returns recovery is the operating model where returned, damaged or unsold garments are processed inside the market they were sold in, across a network of professional garment care facilities. The network draws on trained garment care expertise that already exists in each market, connected through one commercial contract and one audit trail per jurisdiction.
A return in Madrid is recovered in Spain by trained garment care professionals who know how to treat that fabric and finish. A return in Amsterdam is recovered in the Netherlands. The garment goes back into the original stock and sales channels of the market it sold in, without sitting in cross-border transit between the depot and a single national processing site in another country.
Structurally, throughput distributes across nodes. Cleaning and conditioning lead the workflow at each node, with repair following where the garment requires it. The audit trail maps natively to the regulator that owns the market the garment was sold in, because the recovery and the record both sit in that jurisdiction from the moment the garment enters the facility.
The two models on six dimensions
The two operating shapes can be set side by side on six dimensions that matter for fashion brand compliance, cycle time and operational fit.
| Dimension | Centralised | Distributed |
|---|---|---|
| Facility footprint | Single national site | Network of professional garment care facilities across multiple markets |
| Movement of returned stock | Cross-border to the single site | In-market processing |
| Audit-trail mapping | Aggregates at the site of processing; deconstructed per jurisdiction at reporting time | Maps directly to the regulator that owns each market |
| Throughput economics | Concentrated at one site | Distributed across nodes |
| Cycle-time shape | Cross-border transit on inbound leg; single-site processing efficiency | In-market processing on inbound leg; multi-node coordination overhead |
| Examples of operators | Bleckmann (NL), ACS (UK), Reekom (FR), Tersus Solutions (US) | Opera Garment Solutions as a single point of access into a network of operators, from three hubs, in Spain, the Netherlands and the UK, serving the whole EU and the UK. |
Each dimension is a trade-off. No operating model resolves all six in the same direction. Both produce a documented disposition pathway. Both produce a record retained for five years that survives an auditor’s read in the jurisdiction that owns the market the garment was sold in. The selection between them sits inside each brand’s volume profile, market mix, regulator exposure and 3PL contract structure.
The audit-trail consequence per jurisdiction
Regulatory architecture for fashion returns recovery is national, even where the EU Article 24 public disclosure obligation sits on top. Refashion administers the French textile EPR scheme and receives periodic declarations under Loi AGEC. Rijkswaterstaat administers the Dutch national register under Besluit UPV Textiel and receives the annual report before 1 August each year on prior calendar-year activity, with the Inspectie Leefomgeving en Transport enforcing. Naturvardsverket sets the Swedish framework. Germany relies on transposition of the revised Waste Framework Directive (Directive (EU) 2025/1892) by 17 June 2027. Spain has a draft Real Decreto on textile EPR, notified to the European Commission in May 2026 and not yet approved as of October 2026. Sitting on top, Article 24 of Regulation (EU) 2024/1781 sets the EU-wide public disclosure obligation, with the first disclosure covering FY 2025 for calendar-year large enterprises.
Under the centralised model, the audit trail is held at the single processing site. Unit-level records exist. Geographic origin tagging against the regulator that owns each market is the operational task at reporting time. The 10 percent reused-in-the-Netherlands target line under Besluit UPV Textiel becomes a tagging-and-routing question against a centralised inbound flow.
Under the distributed model, the audit trail is held at the facility in the market where the garment was sold. Per-jurisdiction reporting is the native shape of the record. The 10 percent reused-in-the-Netherlands line is captured at the Dutch facility level from inbound forward, with no retrospective tagging required against an aggregated dataset.
Both shapes produce evidenced reports. The operational route to each report differs.
The cycle-time consequence
The centralised model concentrates processing economics at one site. Returned garments arrive in batches that match the cross-border logistics rhythm into the site. Processing throughput at the site can scale against that inbound profile, and capacity uplift sits at one address. Capacity uplift in the distributed model happens across multiple nodes. The five steps of returns recovery setup sit inside that throughput profile.
The distributed model avoids cross-border transit on the inbound leg. A return generated inside a national market reaches the in-market facility without sitting in a container during multi-week sea or land freight cycle. The trade in the other direction is coordination overhead across multiple nodes against transit time saved on inbound logistics, with cleaning and conditioning, then repair where required, performed in market.
Cycle time is a structural variable. Both models can hit defensible cycle times against the volumes the brand is running.
The regulatory implications: 19 July 2026 and the first Article 24 disclosure
Both operating models can produce the unit-level record set that survives audit under the EU and national regimes arriving across 2026 and 2027. The structural difference is how geographic origin information is captured against each garment.
For ESPR Article 25 (the destruction ban, in force 19 July 2026 for large enterprises and 19 July 2030 for medium-sized), the obligation is a documented disposition pathway and a derogation record where one of the ten narrow derogations applies. Both centralised and distributed shapes recover the garment rather than destroy it at specification. Both produce the derogation evidence on the same five-year retention period set by Delegated Regulation (EU) 2026/296.
For ESPR Article 24 (the public disclosure obligation, first covering FY 2025 for calendar-year large enterprises), both shapes produce unit-level records that aggregate to the format prescribed by Implementing Regulation (EU) 2026/2. Centralised reconstructs the per-jurisdiction view at reporting time from the aggregated site-level record. Distributed produces the per-jurisdiction view as the native shape of the record.
For national regimes (AGEC Refashion, the Rijkswaterstaat annual report due before 1 August each year, the revised WFD transposition by 17 June 2027), the distributed model maps to per-jurisdiction reporting from inbound forward. The centralised model produces the per-jurisdiction view through tagging applied against the centralised flow.
Choosing an operating model
The selection between centralised and distributed sits inside brand-specific variables outside the operating model itself. Where the brand sells. How concentrated its volume is geographically. How the 3PL contract is structured. Which jurisdiction’s regulator the compliance lead reports against first. Whether the brand’s volume profile favours throughput concentration at one site or geographic distribution across markets.
Opera Garment Solutions runs the distributed operating model from three hubs, in Spain, the Netherlands and the UK, serving the whole EU and the UK: a single point of access to a network of professional garment care experts, with the audit trail per regulator held at the facility in the market the garment was sold in. Cleaning and conditioning lead the workflow at each node, with repair following where the garment requires it. The unit-level record is held against the garment in the market of sale, with five-year retention aligned to Delegated Regulation (EU) 2026/296 and the national retention windows under AGEC and Besluit UPV Textiel.
Frequently asked questions
How are fashion returns recovery networks structured?
European fashion returns recovery runs on two operating shapes. The centralised model consolidates returned, damaged or unsold stock at a single national site for processing. The distributed model processes returned stock inside the market the garment was sold in, across a network of professional garment care facilities. Both restore garments to resale-ready condition through cleaning and conditioning, then repair where required.
What is centralised returns recovery?
Centralised returns recovery is the operating model where returned garments ship to a single national site for processing, regardless of the market the garment was sold in. Throughput concentrates at one site. Operators running this model in Europe include Bleckmann (Netherlands), ACS (UK), Reekom (France) and Tersus Solutions (US).
What is distributed returns recovery?
Distributed returns recovery is the operating model where returned garments are processed inside the market they were sold in, across a network of professional garment care facilities. Throughput distributes across nodes. The audit trail maps to the regulator that owns each market.
What are the trade-offs between centralised and distributed?
Centralised concentrates throughput at one site and adds cross-border transit on the inbound leg. Distributed processes in market and carries multi-node coordination overhead. Centralised aggregates the audit trail at the site of processing and reconstructs the per-jurisdiction view at reporting time. Distributed holds the audit trail at the in-market facility and maps natively to the regulator that owns each market.
How does each model meet the 19 July 2026 ESPR Article 25 deadline?
Both models route returned, damaged or unsold garments through documented recovery and disposition. Both produce derogation evidence under one of the ten narrow Article 25 derogations where one applies. Five-year retention is set by Delegated Regulation (EU) 2026/296. The structural difference is where the record is held, not whether it exists.
How does each model meet the Article 24 disclosure obligation?
Both models produce unit-level records that aggregate to the format prescribed by Implementing Regulation (EU) 2026/2. Centralised reconstructs the per-jurisdiction view at reporting time from the aggregated site-level dataset. Distributed produces the per-jurisdiction view as the native shape of the record. The first public disclosure covers FY 2025 for calendar-year large enterprises.
The 19 July 2026 ESPR Article 25 destruction ban applies to large enterprises selling apparel into the EU 27. The Rijkswaterstaat reporting deadline under Besluit UPV Textiel falls before 1 August each year in the same compliance calendar. The annual Article 24 public disclosure sits on top. Both centralised and distributed operating models exist inside the same regulatory ledger and produce records that survive audit against it. The structural choice between them sits inside each brand’s operational fit, market mix and reporting calendar. The regulator owns the audit trail standard. The brand owns the operating-model choice that produces evidence against it. For a closer read on the category itself and what sits inside its perimeter, Piece 1 of this series sets out the operational definition and the regulatory ledger that drives it.
( About the author )
Joanna Lambert
CEO & Co-Founder, Opera Garment Solutions
Joanna Lambert is CEO and co-founder of Opera Garment Solutions, which recovers returned, damaged and unsold clothing for fashion brands across the EU and the UK from hubs in the United Kingdom, the Netherlands and Spain. She writes on the EU destruction ban and on what returned stock is worth when it goes back to sale.