The trajectory says 2071 for the EU's 2030 textile ambition. Our research sets out the case for retail take-back as the channel that closes the gap.

The cohort the existing system was designed for

EU residents generate 16 kg of textile waste per person per year. Approximately 4.4 kg is collected separately. The remaining 11.6 kg ends up in mixed household waste, which is operationally unrecoverable for reuse or recycling. The current European capture rate, on the EEA's own measurement, is just under 15%. The cohort whose volume ends up in the bin is, by definition, the cohort that is not driven by prosocial motivation to seek out a charity bin or a municipal collection point. The architecture of the existing system reaches the consumers who already want to be reached.

The volume gap (the 11.6 kg per resident currently ending up in mixed waste) is not an extension of the cohort the existing system serves. It is a structurally different cohort, and closing the gap with more communal infrastructure of the same kind treats the symptom and not the structure.

The 85% is not disengaged. It is mis-served.

The 11.6 kg per resident that ends up in mixed waste is not disposed of out of indifference. It is disposed of after sitting somewhere else first, often for months. The wardrobe stash. The shelf in the garage. The bag by the door that was meant to go to the charity shop. The latent prosocial intent is real; the mechanism the existing system offers for discharging it is not.

What the existing system asks of consumers is a batch behaviour. Accumulate worn textiles at home until you have enough to justify a trip, then take them somewhere. Two failure modes follow predictably from that ask. The first is the wardrobe stash, where the bag never fills enough or the trip never gets prioritised, and the textiles eventually end up in the household bin during a clear-out. The second is the guilt-purge, where the bag does get taken somewhere on the way to something else, often charity, and arrives as a mixed batch of unsorted items in unverifiable condition. Both modes produce the same downstream outcome from a recycler's perspective: bursty volumes of unidentified material, contaminated by the conditions of accumulation, arriving in a rhythm that has no relationship to anything the recycler can plan against.

Retail take-back, designed properly, replaces the batch model with a drip model. The customer brings one worn product back to the store on the day they buy a replacement. The product is identified at the counter. The condition is fresh because the textile has just left active use; it has not spent six months in a bag absorbing household contamination. The customer does this once every few months, in proportion to their actual replacement-purchase rhythm, rather than once a year in a guilt-driven bag-purge. The mechanism converts the ask from an accumulation behaviour to an in-flight behaviour.

This is what we see empirically. Across nine months of operating the platform in our deployment partners, identified customers have returned, on average, around one and a half items each. Not five. Not ten at year-end. One and a half items per identified customer, brought back on the trips they were already making to replace those items. The identification rate at the counter sits at around 90%. Each returned item is photographed, classified, and tied to a specific customer at the point of deposit. That is what the drip model produces in the field, and it is the input shape the recycling industry has been describing as the missing ingredient in feedstock supply.

The four stages below trace what actually happens to a kilogram of textile waste in Europe, from the moment a consumer is finished with it to the moment it reaches its final fate. The aggregate destinations are anchored in published EEA and Refashion data; the cohort split and the channel attribution are conceptual readings of how the system distributes consumers across behaviours. Each stage is labelled accordingly. Drip-fed, identified, single-item feedstock at predictable cadence is the input shape that resolves the 36% sort yield constraint the recycling industry has been describing for years. The infrastructure to support it cannot be retrofitted onto communal collection because the moment it requires is the in-store moment.

The first stage is the consumer base. Three rough cohorts of consumers (not cohorts of waste): a small prosocial group that already seeks out collection points, a large middle group that would participate if the friction were low enough, and a smaller cohort that is genuinely disengaged from the question. The middle is the largest single cohort and the one the existing system has not figured out how to recruit. Bear in mind these are population shares. The waste volume each cohort produces, and where that volume actually goes, is the next stage.

The second stage shows where each cohort's waste volume actually goes. The prosocial cohort, only 18% of consumers, supplies most of what the existing collection system captures (around 14.5 of the 15 units that get collected). The middle and indifferent cohorts together produce roughly 80 units of waste, almost all of it going to mixed household waste. The 15% capture rate the EEA reports is essentially the prosocial cohort working as intended. Everything else is the gap.

A note on the next chart. The 36% European average sort yield is empirical and well-cited. The 85% used to illustrate retail collection is not measured at scale; it is an illustrative anchor reflecting that the four failure modes driving the 36% figure (contamination, mixed composition, unverifiable condition, unknown provenance) are structurally absent from retail collection. The specific figure is illustrative; the direction of the difference is not. Full methodological treatment in the white paper.

The third stage is what happens to the small share that actually enters a collection channel. The European average sorting yield is 36% on collected volumes, which means even of the 15% that does get captured, more than half is lost during processing and ends up in incineration or landfill anyway. The recycling industry has been quite open about this constraint. The reason the constraint exists is that most collected material arrives in conditions that cannot be qualified after the fact. Retail collection produces material that arrives qualified at the source, and the sort yield rises accordingly.

The fourth stage is the aggregate. Combining the direct mixed-waste with the post-collection sort losses, roughly 94% of European textile waste reaches its final fate as incineration or landfill. Less than half a percent reaches textile-to-textile recycling. The system, taken as a whole, is producing the outcome the system was inadvertently designed to produce. The argument for retail take-back is not that it produces a better channel within an otherwise-functional system. It is that the system, on the data, is not producing recycling outcomes at all, and the channel that has the best chance of changing that is the one running through stores where consumers are already shopping.

The reason that channel is retail, specifically, is that retail is the only operator who already has the moment, the incentive, and the relationship. The customer is in the store on the day of replacement intent, not disposal intent. The retailer can attach a discount, a credit, or a service to the return that no communal collector can match commercially, because the retailer is the one selling the next product. The relationship and the data event are continuous with an existing CRM, not adjacent to one. Retail take-back is not a sustainability programme grafted onto the store. It is the store doing what only the store can do.

The trajectory says the same thing, empirically

The structural argument the previous sections framed has an empirical signature. The chart below plots the actual EEA capture rate against the policy ambition.

This is the empirical anchor for the structural argument. The trajectory exists, but the slope of the trajectory is the slope of the existing system serving its existing cohort. To bend the curve, the channel mix has to change.

What the participation evidence says about levers

The underinvestment in retail take-back is not solved by larger discounts or higher-cost incentive structures. The peer-reviewed evidence on participation (McKie et al. in Manufacturing & Service Operations Management; Wollbrant and Knutsson in Nature Human Behaviour) is that the design of the moment matters more than the size of the discount, and that bigger does not reliably mean better. Convenience is the largest single lever; information framing, when meaningful, outperforms generic appeals. The investment that produces results is the investment in the design of the moment itself: the friction, the framing, the meaning, the information layer. This is a smaller and more replicable investment than the inherited frame allows for, and one only retail can make.

What we keep coming back to

For nine months we have been operating a retail take-back platform across the Netherlands, with INTERSPORT, Runnersworld, EK Sport, and FastFeetGrinded. The pattern we kept seeing was that the framing the industry was working from was producing the wrong answer about what retail collection can do. So we pulled the published evidence into one frame: peer-reviewed field experiments, EU policy data, recycling industry analysis, and our own pilot findings. Four streams. None originally about retail take-back specifically. All point the same way. The full case is in the white paper; this article is the wedge into it.

Retail take-back has been treated as the long tail. The evidence is now consistent enough to say it is not. The investment case has changed. The framing has not yet caught up.

So here is what we keep asking ourselves, and what we think the producer responsibility organisations, the brands, the retailers, and the policy-shapers reading this should be asking themselves too. If the evidence is this consistent, and the operational mechanism is this available, what is the reason, exactly, that the investment has not yet followed?