Take-back programmes: what data do fashion brands need to collect and report?

A take-back programme needs data on four things: what is collected (volume, weight, product type), how it is sorted (grades and fractions), where each fraction goes (reuse, resale, recycling, disposal) and what evidence the downstream partners provide. With textile extended producer responsibility (EPR) coming across the EU by 2028, these data points move from sustainability reporting into compliance and cost. The checklist below covers the full chain from store counter to recycler.

Why does garment take-back need better data now?

The revised Waste Framework Directive, approved by the European Parliament in September 2025 and in force since October 2025, requires Member States to establish EPR schemes for textiles and footwear within 30 months, by April 2028. Producers, including online sellers and companies based outside the EU, will finance the collection, sorting and recycling of the textiles they place on the market. The scope covers clothing, footwear, accessories and household textiles such as bed linen and curtains, and Member States may extend it to mattresses. Micro enterprises get an additional year.

The law firm Beveridge & Diamond summarises further points that matter for take-back: Member States had to set up separate collection of used textiles and textile waste by 1 January 2025; operators should prioritise local sorting and reuse before shipping abroad; items unfit for reuse should be sorted for remanufacturing and recycling; and producers must appoint an authorised representative to fulfil their EPR obligations. The directive also requires eco-modulated fees, which are intended to charge more for less sustainable products and to reward design for durability and recycling.

Separately, the ESPR ban on destroying unsold apparel and footwear applies to large enterprises from 19 July 2026, with annual disclosure of the number, weight, reasons and treatment of discarded unsold products. Take-back and returns flows increasingly meet in the same sorting facilities, so data definitions should be aligned.

What collection data should a take-back programme capture?

1. Collection point identifier (store, partner, mail-in) and date.
2. Weight per collection bag or container at hand-over, measured, not estimated.
3. Count of items where feasible, at least for items accepted for trade-in credit.
4. Product category and, where readable, brand and product identifier (label, QR or NFC tag).
5. Incentive issued (voucher, credit, points) and the customer record, with consent.
6. Chain of custody: who transported the goods, when, and to which sorting site.

Customer data needs particular care. Vouchers and trade-in credits link a take-back event to a person, so the processing basis, retention period and consent for marketing use should be defined under GDPR (and the Swiss nFADP for Swiss stores) before the programme starts. Store staff also need a simple, consistent way to record hand-overs; a weighing scale and a scanned bag label at the counter are usually more reliable than manual estimates entered at the end of the day.

Stores are often the main collection point. H&M Group, which began collecting garments in 2013, set up Looper Textile Co. with Remondis in February 2023 as a 50/50 joint venture to collect, sort and sell used and unwanted garments and textiles for reuse and recycling, with plans to use automated sorting technologies such as near-infrared sorting.

Read also

Is there a DPP business case beyond compliance? Resale, service and customer data

What sorting data matters for reuse and recycling?

This evidence rarely arrives automatically. Contracts with collectors and sorters should specify the fractions to be reported, the unit (weight, item count), the reporting frequency, the file format and the right to audit downstream partners. Without such terms, brands typically receive a single total weight per period, which is not enough for EPR reporting or for public claims about reuse.

Composition data is the bottleneck for recycling. Automated sorting by near-infrared can identify many fibre types, but blends, coatings and trims complicate it. Product data that travels with the garment, for example via a label code or a future Digital Product Passport, can make sorting faster and more accurate.

Where can AI help in take-back operations?

• Image based pre-grading. Classifying items by category, condition and visible damage to support human sorters.
• Label and tag reading. Extracting brand, size and composition from care labels to enrich item records.
• Routing suggestions. Recommending reuse, resale or recycling based on grade, brand and demand data.
• Report drafting. Aggregating partner data into EPR and sustainability reports, with humans checking the figures.

What should the reporting checklist include?

1. Total weight collected per country and collection channel, per reporting period.
2. Breakdown by fraction (reuse, resale, repair, recycling, downcycling, disposal) with evidence for each.
3. Share sorted and reused locally versus exported, with destination countries.
4. Incentives issued and redeemed, and their cost.
5. Products placed on the market per country, by weight and category, as EPR fees are based on such data.
6. Product attributes relevant to eco-modulation, such as durability, repairability and recycled content.
7. Alignment with unsold goods disclosures: quantity, weight, reason and treatment, using the same definitions.

Read also

ESPR explained: what the EU ecodesign regulation means for textiles

What are common gaps in take-back programmes?

The most common gap is the hand-over: once collected goods leave the store, many brands receive only a total weight from their partner. Without contractual data requirements, fractions and destinations remain unknown. A second gap is the link between products placed on the market and products collected, which EPR schemes will need for fee calculations. Brands should write data fields, formats and audit rights into contracts with collectors and sorters now, test a quarterly data exchange, and treat the first EPR reporting period as the deadline for a working process rather than a starting point.