A research artefact bringing peer-reviewed operations literature, EU policy data, and Netherlands pilot evidence into one frame.
Producer responsibility organisations and policy-shapers have, until now, treated retail take-back as a marginal contributor to the European circular economy. The reasoning is intuitive: communal collection handles the bulk, retail handles the long tail, and policy is therefore designed primarily around the bulk channel. This artefact argues that the volume assumption is wrong, or rather that it is right only under the conditions where retail take-back is currently designed to fail. Properly designed, the in-store moment delivers both higher quality and higher volume than the existing system recognises. It is a customer-acquisition moment, a transparency moment, a CRM moment, and a feedstock-qualification moment, taking place simultaneously and through the same physical interaction. Most retailers, regulators, and producer responsibility organisations are treating it as a fraction of the first thing only, and missing the rest.
This paper makes the case for the second reading. It synthesises three streams of evidence. EU residents generate 16 kg of textile waste per person per year, with 82% of it post-consumer [3]; industry analysis from Boston Consulting Group and ReHubs projects volume rising 36% by 2035 [13]. Peer-reviewed field experiments find that transparency about a retailer's internal responsibility practices produces measurable lift on purchase behaviour [1], and that participation in clothing take-back programmes is shaped more by the design of the moment than by the size of any discount on offer [2] [14]. The Netherlands pilot conducted by Utilitarian with Runnersworld and INTERSPORT during 2025 demonstrates that the channel can be operated as a customer-acquisition mechanism, producing structured data at a unit cost approximately an order of magnitude below paid social [8] [12].
The pilot envelope is small (five stores, 750+ pairs scanned, 600+ customers). The structural argument it supports applies to physical retail at any scale.
The starting point is a number that is not in dispute. The European Environment Agency estimates that EU-27 residents generate 16 kg of textile waste per person per year [3], on the most recent EEA data covering 2022 [4]. 82% of that waste is post-consumer [3]. These are clothing and household textiles that left a retail shelf, entered a home, were used, and then exited that home as waste.
The trajectory is rising, not stable. Industry analysis published by Boston Consulting Group and ReHubs in 2026, drawing on a harmonised baseline of European waste flows, finds that European post-consumer textile waste will rise from 13.3 Mt per year in 2025 to 18.1 Mt by 2035, an increase of 36% over the decade [13]. The same analysis finds that of the 13.3 Mt generated in 2025, only 1.5 Mt is collected and sorted into recycling-grade streams. That is a collection rate of 33% and a sorting rate of 36% on what is collected. The structure of the European system is one in which most post-consumer textile waste exits the operating boundary before any retailer or recycler sees it.
The downstream consequence is that less than 1% of post-consumer textile waste in Europe is currently recycled back into new textiles [13]. The constraint is not technological. The constraint is feedstock: collection produces volume, but not the data and the material quality that recyclers need to operate at scale. The argument this artefact develops is that the right collection model, hosted in retail rather than at communal or municipal points, materially changes both the data and the quality of what enters the recycling stream. Section 6 returns to this point in detail.
The breakdown of the 16 kg per person matters more than the headline. Of that volume, approximately 4.4 kg is collected separately for reuse and recycling. The other 11.6 kg ends up in mixed household waste. That 11.6 kg figure is the operative one for any retailer thinking about end-of-life. It is the volume that already escaped the retail relationship the moment it entered the home, and that, by the time it reaches the bin, is unrecoverable for any commercial purpose.
Every kg of textile waste that ends up in mixed household waste is a customer interaction that did not happen.
Separate collection of textiles has been mandatory across EU member states since 1 January 2025 [5]. That mandate moves volume out of the bin, but it does not specify where the volume moves to. The EEA notes that most existing collection in Europe happens at street containers and bring points [3]. These are locations with no relationship to the retailer who originally sold the product. From the retailer's perspective, the mandate redirects a stream that has already exited the customer relationship; it does not bring the customer back.
This is the structural gap the rest of the artefact addresses. The product was bought from a retailer. The customer kept it. The customer wore it out. The product is on its way back into the world, into a stream that, on current trajectory, will be a third larger by 2035 than it is today. The question is whether the retailer is in the room when that next decision is made.
Buell and Kalkanci, writing in Management Science in 2021 [1], ran field experiments testing how disclosure of internal versus external responsibility initiatives affected consumer choice. The internal version is the company describing what it does itself: paying living wages, lowering its own emissions, taking responsibility for its own end-of-life. The external version is philanthropy: donating to causes outside the operating boundary. The two are different rhetorical acts.
The finding is that transparency about internal responsibility practices lifts purchase probability by 6–46%, an effect at least as large as transparency about external causes. Purchase probability here means the share of consumers who proceeded to buy after seeing the disclosure, measured in actual purchasing behaviour rather than in stated preferences. The range is wide because conditions varied across treatments. The lower bound is real, the upper bound is real, and the lower bound is already large enough to matter to a marketing director with a customer-acquisition-cost target.
The relevance to take-back is that the in-store collection point is, by definition, an internal-responsibility moment. The customer is in the retailer's store, with a product the retailer sold, choosing to bring it back. The retailer designed the moment, designed what happens to the product after, and designed how the customer experiences it. There is no philanthropy mediating the interaction. There is no external cause being supported. The interaction is the transparency, and the transparency is the interaction.
The mechanism does not require physical staff handling of the item. In a typical deployment, the take-back point is a collection box with a QR-code poster. The customer scans with their own phone, completes a short flow that captures product information and contact details, and deposits the item. Staff are present in the store and available to advise or support the customer if asked, but the design does not require them to handle the product or perform the disclosure themselves. That distinction matters operationally. The retailer who is considering a take-back programme is not being asked to add labour to the store; the retailer is being asked to provide an option, with staff support as a feature available to be used rather than a workload that comes attached.
The transparency premium does not require additional staff time. The retailer's choice to provide the take-back option, and to communicate what happens next, is itself the disclosure.
This matters because the marketing literature has, for a generation, treated environmental and ethical claims as a soft asset: nice to have, brand-building, hard to attribute to revenue. Buell and Kalkanci's experiments push back. They find that internal-responsibility disclosure produces measurable, attributable lift in purchase probability under conditions designed to test exactly that. The finding is robust across two field experiments and complementary online experiments. It is not a survey of stated preferences. It is a measurement of behaviour.
One framing point bears flagging here, because it sets up the section that follows. The take-back moment is also a commercial exchange. The customer who brings their old running shoes back to the store on the same visit they buy a replacement pair is, in commercial terms, completing a trade-in. The transparency premium runs alongside that exchange; it does not replace it. The next section addresses the commercial structure of the moment directly.
The behavioural-economics literature on incentives and prosocial behaviour is rich. Most of it is reasoning about a different problem from the one a retailer faces. The studies that frame the public conversation about rewards (cash payment for blood donation, deposit refunds in long-established beverage-can recycling schemes) are studies of what happens when payment is introduced into pre-existing prosocial habits with cultural tradition behind them. Retail take-back is not in that category. Most consumers are not currently participating in any take-back stream. Their old products sit in cupboards, or they go in the household bin. The question for a retailer is not how to preserve a prosocial habit. It is how to create an exchange where one does not currently exist.
The frame that makes this section coherent is simpler than the literature it draws on. From the consumer's side, the take-back moment is a cost-benefit calculation. The cost is the effort required to participate: locating the in-store collection point, scanning, photographing the item, supplying contact details. The benefit is what the consumer gets in return: a discount on a replacement they were going to buy anyway, plus, if the design is right, a sense that the small additional effort contributed to something. The literature on take-back participation, read carefully, is a literature about that cost-benefit equation and where it bends. The trade-in framing below is what defines the benefit; the prosocial overlay is what changes the perception of the cost; the three sub-questions at the end are three separate angles on the same equation.
A discount on a replacement purchase is not a reward for prosocial behaviour. It is the commercial structure of a trade-in. The same form has operated in automotive, mobile phones, and white goods for decades, and the analytical literature on those exchanges does not reach for crowding-out theory to explain them. A customer who brings an old phone into a store on the day they buy a new one is recognised as completing a commercial exchange: the residual value of the old item is applied to the cost of the replacement. Nobody calls that a reward for being prosocial. They call it a trade-in.
The customer bringing a worn pair of running shoes back to the store on the day they buy replacements is doing the same thing. The act is a commercial exchange in which the old item, surrendered, contributes to the purchase of the new one. The discount is not a reward for an environmentally virtuous act. It is the commercial expression of the surrender. The email-for-discount mechanic that runs alongside the deposit is itself a commercial pattern customers already understand from a hundred other retail interactions: newsletter signup for first-purchase discount, loyalty enrolment for members-only pricing, the email gate at checkout. The take-back scan sits inside this same accepted retail pattern, not adjacent to it.
Layered on top of the commercial exchange, the customer is asked to do a small additional action. In a typical Utilitarian deployment, the QR-driven flow asks the customer to photograph the product. The photograph captures product information (material composition, condition, brand, model) that improves the recycling process by qualifying the material before it leaves the store. The framing presented to the customer at that step is that the photograph contributes to the recycling process, and that this small additional effort matters.
This is a prosocial overlay on a commercial exchange, not a prosocial act being paid for. The framing is a deliberate design choice. Its job is to convert what would otherwise read as an administrative request (please photograph your item before you deposit it) into a contribution (your photograph helps improve the process). McKie et al.'s finding that donation framing outperformed recycling framing in their experiments [2] is best understood through this lens: meaning-making in the framing affects compliance with the small additional ask, not the underlying motivation to participate in the exchange itself.
Public debate tends to collapse the question of incentives into "should we pay people more?" That is the wrong question, or at best a weak version of three different questions, each of which has a different answer.
The convenience question. Is retail take-back more convenient than the alternatives? For a customer already going to the store to buy replacement shoes, yes. The friction cost approaches zero because the trip is happening anyway. Retail take-back competes first with the cupboard and with the household bin, not with charity bins or municipal collection. McKie et al. find that convenience is the largest single lever in their experiments and that lower-friction return at the moment of disposal partially substitutes for monetary incentive [2]. The implication is that the design of the moment (where, how, when) matters more than the design of the reward.
The existence-of-incentive question. Does there need to be a reward at all? The crowding-out literature, which is sometimes cited to argue against incentives in take-back, warns that introducing payment into a strongly prosocial act can damage participation. But that warning assumes a prosocial baseline. The retail-take-back baseline is the cupboard. The cohort the system needs to reach (the non-participating cohort whose volume is currently lost to mixed waste) has, by definition, no prosocial habit to displace. For this cohort, a small incentive does not crowd out an existing motivation. It provides a reason to act in a context where, until now, no reason has been provided. The selection bias is sharp: the consumers who are already participating in charity bins and municipal collection are the ones with established prosocial motivation. Reaching the rest, who account for the majority of volume, requires a different mechanism.
The size-of-incentive question. Given a reward exists, does its size matter? The peer-reviewed evidence is that it does not, in the linear way most discount-mechanic discussions assume. Wollbrant and Knutsson, writing in Nature Human Behaviour in 2022 [14], used the gradual replacement of Sweden's lowest-denomination coin to study how participation in beverage-can recycling responded as the deposit value increased. The relationship was S-shaped, not linear. At low reward levels, participation rose with the reward; in a middle band, it fell; only at higher reward levels did it rise again. The interpretation: framing dominates magnitude, and bigger does not reliably mean better. McKie et al. find the same pattern specifically in clothing take-back: larger discounts produce diminishing returns, and the binding lever is the design of the moment [2]. The marginal euro is more productive spent on the design of the moment than on the size of the discount.
Retail take-back at the in-store moment is a commercial exchange with a prosocial overlay, not a prosocial act with a commercial reward. The exchange (item surrendered, discount applied to a replacement purchase, email and product data captured) sits inside accepted commercial-transactional patterns customers already understand. The overlay (the framing of the photograph as a contribution to the recycling process) is a deliberate design choice that converts a small administrative ask into a meaningful one. The behavioural design of the moment dominates the financial design.
The full behavioural-science treatment, with the supporting literature on motivation crowding-out, Self-Determination Theory, and reward sequencing, is developed in Purpose Before Payoff, the canonical Utilitarian piece on this question.
The peer-reviewed studies cited so far are the strongest available evidence on, respectively, the consumer-purchase effect of internal-responsibility transparency [1], the participation-design effect of information and convenience in clothing take-back [2], and the relationship between reward size and participation in deposit-style recycling schemes [14]. They do not, however, address one another directly. Buell and Kalkanci's experiments were not set in a take-back context. McKie et al.'s experiments were not set up to measure subsequent purchase behaviour by participating consumers. Wollbrant and Knutsson's natural experiment was set in a national deposit-refund scheme rather than in retail take-back at the store.
The bridge between the three findings is the empirical question of what happens when an internal-responsibility transparency moment is also a take-back trade-in moment. A customer brings a worn product back to the store where they bought it. The customer is replacing the product on the same visit. The retailer captures structured data alongside the exchange. What does the customer do next, in the weeks and months that follow?
The literature does not answer this directly. It is, however, the question every retailer running a take-back programme is implicitly answering with their own data, whether they realise it or not.
What follows is the partial answer the Netherlands pilot provides.
The Netherlands is, on the published European data, the country where the system already works hardest. BCG and ReHubs identify the Netherlands as the European reference case where binding EPR and enforcement translate into measurable performance: collection at 47% of post-consumer textile waste against a European average of 33%, sorting at 97% of collected volumes against a European average of 36% [13]. The country also has the highest fibre-to-fibre recycling targets in Europe, with binding goals of 33% by 2030. The pilot took place inside this regulatory and operational environment. The conditions under which it was tested are the conditions other European markets are moving towards, not away from.
The pilot ran across five stores in the Netherlands during 2025, in partnership with Runnersworld and INTERSPORT [8]. The published envelope is 750+ pairs of running shoes scanned and 600+ customers participating. The pilot was designed to test the operational reliability of an in-store collection programme deployed as a customer-acquisition channel rather than as a logistics channel. Three things were measured deliberately:
The answer to all three is yes. The mechanism: a poster at the take-back point with a QR code; the customer scans, photographs the worn product, identifies it (with AI assistance for brand and model), and provides their email address; the customer receives a discount voucher; the retailer receives a structured record (brand, model, store, timestamp, customer email, consent) in their CRM.
The directional commercial finding from the pilot is one already reflected on utilitarian.world and in the ROI calculator: cost per verified customer email captured at the take-back counter sits in the €1–3 range [12], against industry benchmarks of €12–25 from paid social and €18–31 from website popups [9] [10] [11]. The 4–25x cost differential is not the most interesting finding. The most interesting finding is what the email is attached to.
An email from the take-back counter arrives with a product, a store, a date, and a customer who is, by definition, in a replacement-consideration window.
An email from a paid social campaign arrives with a click. An email from a website popup arrives with an opt-in. An email from the take-back counter arrives with a product, a store, a date, and a customer who has just demonstrated, behaviourally, that the product they brought back has reached the end of its useful life. The marketing team that receives that record into their CRM is not receiving a lead. They are receiving a structured replacement-consideration trigger.
The envelope is small. Five stores is a pilot, not a deployment. 600+ customers is a sample, not a population. The pilot does not, by itself, establish what the long-run conversion rate of take-back-counter emails to subsequent purchases will be at scale. It establishes the operational viability of the channel and the directional cost economics. The conversion rate at scale is a question for further measurement as the platform deploys across more stores and longer time horizons.
What the pilot does establish is that the channel exists, runs without dedicated staff training or new hardware, and produces an email plus product-level data record at a unit cost in the €1–3 range. Those are the conditions necessary for the structural argument in the rest of this artefact to be operationally credible.
The point made briefly in section 1 needs to be argued in full here. Less than 1% of post-consumer textile waste in Europe is currently recycled back into new textiles, against a system that generates 13.3 Mt per year and is heading to 18.1 Mt by 2035 [13]. The constraint on closing this gap is not the absence of recycling technology. Multiple textile-to-textile pathways are technically demonstrated. The constraint is feedstock: the material that recyclers receive, today, does not arrive at a quality and with a level of attribution that allows recycling to operate at industrial scale.
The reason is the structure of the existing collection system. Most post-consumer textile waste in Europe is collected through street containers and bring points [3], infrastructure designed for volume rather than for material quality. Collected textiles arrive at sorters as mixed bags, typically without information about brand, composition, age, or condition. Sorting is itself constrained: the European average is 36% of collected volumes converted into recycling-grade streams [13]. The remainder is downcycled, exported, or routed to lower-value end uses. The economics of this system are bounded by what the sorters can produce; the economics of recycling, in turn, are bounded by what the sorters supply.
In-store collection at the point of sale changes both inputs to that equation. The collection moment captures product-level data at the source: brand, model, category, store, date, and (with consumer-supplied photography) material composition and condition. The data is attached to the item at the moment it enters the recovery system, not inferred or reconstructed downstream. This is the difference between a sorter receiving a mixed bag of unidentified garments and a sorter receiving a stream of garments each of which already carries a structured record of what it is. The first stream is feedstock with significant downstream qualification cost. The second stream is feedstock that arrives qualified.
The same mechanism also raises the average quality of what enters the system. A customer who is bringing a worn pair of running shoes to the store on the day they buy a replacement is participating in the moment because the moment is designed for them. The collection environment is sheltered, the items are not exposed to weather, contamination from food and household waste is minimised, and the products entering the stream are concentrated in categories the retailer actively sells (in this case, sportswear and footwear with a strong polyester and cotton bias, the two fibre families that account for 79% of recycling-grade post-consumer feedstock) [13]. Communal and municipal collection produces a broader and noisier mix; retail collection produces a narrower and cleaner one.
The implication for the recycling industry is direct. The volume problem (collection rates stuck around 33%) and the quality problem (sorting rates around 36%, with much of the output not recycling-grade) reinforce each other in the existing system. The right collection model, hosted in retail rather than at communal points, addresses both at once. It does not replace municipal collection; it complements it, by adding a stream that is structurally higher in quality and information density. For the brands whose products are entering the stream, this is also the route by which their own materials become identifiable as their own when they arrive at recyclers, which matters for circular-content claims, for closed-loop programmes, and for the verifiable evidence such claims now require.
The same data that arrives in the marketing team's CRM has a parallel destination on the disclosure side. Under the EU's ESRS E5 standard on resource use and circular economy [6], large in-scope companies are required to disclose resource outflows, the destination of waste, and the extent of circular economy practices in their operations. The standard is part of CSRD-driven reporting and applies under the post-Omnibus-I scope of more than 1,000 employees and more than €450 million in net turnover.
The disclosure requirements specify that information must be verifiable. ESRS E5-5 on resource outflows, together with related sub-requirements, asks for product-level data, methodologies, and traceability. Aggregate weight reports produced by recyclers do not satisfy the requirement directly. Not because they are wrong, but because they are not attributable to specific products at specific points in the operating boundary.
Product-level take-back records of the kind the pilot produces (brand, model, category, store, timestamp, circular outcome) sit at the right granularity for E5 disclosure. This is not a coincidence; it is the logical consequence of what is being measured. If a retailer wants to disclose the destination of products it sold, the most defensible record is one captured at the point those products re-enter the retailer's operating boundary. That point is the take-back interaction.
The compliance use of the data and the marketing use of the data are, structurally, the same data exported into two different reporting destinations. The customer's email address is redacted in one destination and present in the other; the product, the store, the timestamp, and the circular outcome are present in both. The feedstock-qualification use of the same data is exported into a third destination, the recycler, with its own redaction profile. One capture, three uses, three destinations.
For most retailers, the marketing team and the sustainability team have, until now, been working on adjacent problems with adjacent budgets and almost no shared data. The take-back interaction does not require them to merge. It produces data that goes to both, and to the recycler, from a single moment.
The reasoning that has historically positioned retail take-back as a marginal contributor to European textile collection deserves direct engagement here. The argument runs roughly as follows: communal collection (street containers, charity bins, municipal points) handles the majority of separated textile flows; retail take-back handles a long tail of brand-loyal participants; therefore, policy should be designed primarily around the bulk channel and retail involvement is a useful but secondary contribution. This reasoning has held for the period in which most retail take-back programmes have been narrow, brand-specific, voluntary, and operationally minimal. Under those conditions, it is not unreasonable.
The reasoning fails when retail take-back is designed differently. The European collection rate sits at 33% because the existing channels reach the prosocial-baseline cohort and do not recruit beyond it [13]. The non-participating cohort, whose volume is the 11.6 kg per person currently lost to mixed household waste, is by definition the cohort that is not driven by prosocial motivation to seek out a charity bin or a municipal collection point. Reaching that cohort requires a different mechanism. Retail take-back at the point of replacement purchase is one of the few mechanisms that can reach it, because the trip is happening anyway and the marginal effort approaches zero. The cost-benefit equation laid out in section 3 explains why this works at the level of the individual consumer; the volume implication at the system level is that retail collection has the potential to convert a meaningful share of currently-uncollected volume into qualified feedstock, not because retailers will outcompete municipal collection but because retailers will reach a cohort municipal collection has not reached.
The point bears underlining. The argument here is not that retail collection should replace communal infrastructure. Communal collection serves the prosocial-baseline cohort and serves it well. The argument is that the volume gap (the gap between current 33% collection and the policy ambition of 50% and beyond) cannot be closed by adding more communal infrastructure of the same kind, because the binding constraint is not the absence of more bins; it is the absence of a mechanism that recruits the non-participating majority. Retail is the channel best positioned to close the gap, on volume as well as on quality. Producer responsibility frameworks designed around the bulk-versus-tail mental model will systematically underweight retail's contribution to the volume question, not because retail's contribution is small but because the model is asking the wrong question.
The data captured at the in-store scan is not the full extent of what the moment of identification can do. Computer vision and product classification, applied at the point the customer is preparing to deposit an item, also create a screening function. The system can determine, before the item is committed to the bin, whether the item belongs in this collection stream, whether it has been prepared correctly, and what the right response is if either of those checks fails. This is a step that today happens at the sorting facility, downstream, after the item has already entered the system. Moving it forward into the moment of identification changes the economics of the entire chain.
What the screen does in practice is straightforward. Consider a collection point set up to receive denim for a specific recycling pathway that requires cotton-rich feedstock. A customer arrives with a cotton t-shirt, scans it, and the app recognises that the item does not match the collection point's target material. Rather than rejecting the customer or accepting an item that will contaminate the stream, the app redirects: this is a cotton garment, but not denim; here is where it should go instead, here is why this collection point is structured the way it is, here is what happens to the items that are deposited correctly. The customer leaves either with the t-shirt and a clearer understanding of where it belongs, or with the t-shirt routed to an alternative collection arrangement at the same store if one is available. The same mechanism handles preparation requirements: where items need to be prepared in a specific way for safe handling or to preserve recyclability, the instructions appear on the customer's phone at the moment they are needed, not on a poster the customer did not read.
This is structurally different from sorting. Sorting is cost recovery: it is the work of dealing with contamination after the fact, and its cost falls on the operator of the sorting facility. The yield of the sorting process is bounded by what the sorter can fix on the back end, and items that arrive in a state the sorter cannot fix are landfilled or downcycled at a loss. Screening is contamination prevention: it stops bad inputs from entering the system at all. The marginal cost of screening at the scan is essentially zero (the customer's phone is doing the work, the application is doing the classification, the customer is doing the redirection if required); the marginal cost of sorting at the back end is significant, and is rising as labour costs in European sorting facilities continue to climb. The BCG and ReHubs analysis sizes the European sorting build-out at €300–450M in CAPEX and €500–850M in annual OPEX through 2035 [13]. Screening at the source does not eliminate the need for sorting, but it reduces both the volume of contamination that sorters have to process and the share of incoming material that fails sorting and is lost.
From the customer's perspective, the screen is the prosocial overlay made operational. The framing argument in section 3 was that the photograph contributes to the recycling process and that the small additional effort matters. The screening interaction is the same argument carried into a moment where it could otherwise feel like rejection. The customer who is told "this isn't right for this collection point, but here is where it should go" is being treated as a contributor, not a violator. The information is offered, not enforced, and the customer leaves the interaction with more knowledge than they arrived with. This is the difference between a take-back programme that operates on bins-and-trust (where contamination is accepted as a downstream cost or where rules drive volume away because customers cannot verify they are using the system correctly) and a take-back programme where the system tells the customer how to comply. The volume penalty of strict rules approaches zero when the customer is being redirected rather than rejected.
For producer responsibility organisations and policy-shapers, the implication is that the choice between strict rules (which drive volume away) and loose rules (which produce contaminated streams) is a false dilemma when the screening function exists. Strict rules with screening produce both clean streams and high participation, because compliance is supported in real time. This is a property of the in-store moment that no other collection mode in the European system can match.
Pull the threads together. EU residents generate 16 kg of textile waste per person per year, and 11.6 kg of that volume ends up in mixed household waste, outside any retailer's operating relationship with the customer. The trajectory is rising: post-consumer textile waste in Europe is projected to grow 36% by 2035 [13]. Field experiments find that transparency about internal responsibility practices produces measurable lift in purchase behaviour. Independent peer-reviewed research finds that take-back participation responds to the design of the moment more reliably than to monetary incentive size. Less than 1% of the post-consumer textile waste generated in Europe is currently recycled back into new textiles, with feedstock data and quality identified as the primary constraint. The volume problem and the quality problem reinforce each other in the existing system, and the channel best positioned to close both gaps simultaneously is the one that has been treated as marginal: in-store collection at the point of replacement purchase. The Netherlands pilot demonstrates that a take-back programme designed for this purpose can operate as a customer-acquisition channel, producing structured data at a unit cost approximately an order of magnitude below paid social, and producing the same data in a form that satisfies the granularity required by emerging EU disclosure standards and by recyclers seeking qualified feedstock.
The structural argument follows. The take-back interaction is a moment that has, until recently, been treated by physical retailers as a logistics interaction: an obligation under EPR, a cost line, a moment to manage. The peer-reviewed evidence and the operational data converge on a different reading. It is a customer-acquisition moment, a transparency moment, a compliance-evidence moment, and a feedstock-qualification moment, taking place simultaneously and using the same physical interaction.
A retailer who treats it only as a logistics interaction is paying the cost of the moment three times: once in the EPR fee, once in the marketing budget that buys leads at €12–25 each from paid social, and once in the sustainability team's cost of producing audit-grade evidence from a recycler's aggregate weight report.
One poster. One twenty-second interaction. Four business cases solved with the same data.
This is the case for treating the take-back counter as infrastructure, not as compliance. The platform that does this work is named in the byline of this artefact and at utilitarian.world, which is where readers who want to follow up commercially should go. The structural observation, however, does not depend on which platform performs it. The observation is that the data is at the counter, available to be captured, and that retailers with thousands of stores already have the customer relationship and the physical footprint required to capture it.
The remaining question is operational, not strategic. It is which retailers move first.
This artefact synthesises three streams of evidence: peer-reviewed operations and behavioural research [1] [2] [14], EU policy and statistics together with industry analysis on European textile-waste flows [3] [4] [5] [6] [7] [13], and primary operational data from the Netherlands pilot conducted by Utilitarian with Runnersworld and INTERSPORT during 2025 [8] [12]. Industry cost-per-email benchmarks [9] [10] [11] are cited as commercial comparators, not as findings.
The pilot data is reported within the published envelope: five stores, 750+ pairs of shoes scanned, 600+ customers participating. Specific store-level statistics, individual partner outcomes, and per-store breakdowns are commercial-in-confidence and are not disclosed here. The directional cost-per-email range (€1–3) is consistent with the figures already published on utilitarian.world and reflected in the ROI calculator; it is presented as a directional figure, not as a claim about long-run unit economics at scale.
The peer-reviewed studies cited here measure effects under their own experimental conditions. Buell and Kalkanci's transparency-premium finding is from field and online experiments in retail contexts that are not specifically take-back contexts. McKie et al.'s clothing take-back finding is measured under information and convenience treatments designed to isolate those variables, not in a complete operational deployment of the kind the Netherlands pilot represents. Wollbrant and Knutsson's S-curve finding is from a national deposit-refund scheme for beverage cans, an instrument with cash refunds rather than store-credit discounts; its application to retail take-back is interpretive, not direct. The bridge between these three literatures and the pilot's operational evidence is an interpretive one, not a statistical one.
The BCG and ReHubs 2026 industry analysis is cited for its harmonised baseline of European post-consumer textile waste flows and country-level performance, not for its strategic conclusions on system financing. The figures used here (volumes, collection and sorting rates, country snapshots) are descriptive baselines on which independent organisations and authorities now broadly converge.
The argument in section 7 is, accordingly, structural. It is the strongest case the available evidence supports and is offered as a frame for further measurement, not as a closed empirical claim.
For a retailer reading this artefact, the operational implications are three.
First, the take-back counter is data infrastructure. It produces a structured record of brand, model, store, timestamp, customer email, and circular outcome at a unit cost in the €1–3 range when designed as a customer-acquisition channel rather than as a logistics channel. The data exists; the question is whether it is captured.
Second, the marketing team and the sustainability team are working on the same problem from opposite ends. The marketing team needs cheaper email and product-level intelligence about replacement-consideration windows. The sustainability team needs audit-grade product-level evidence for E5 disclosure. The take-back counter produces both from the same interaction.
Third, the structural argument does not depend on conviction. It depends on whether the data is captured. The peer-reviewed evidence on the transparency premium is independent of the platform that operationalises it. The EEA statistics are independent of the retailer who acts on them. The ESRS E5 disclosure requirement is independent of the operational design that satisfies it. The retailer who does nothing pays the cost of the take-back moment three times. The retailer who treats the moment as infrastructure pays it once.
The artefact is written primarily for marketing, retail-leadership and sustainability readers at retailers. It is also addressed, secondarily, to the producer responsibility organisations and EPR-implementing bodies designing the rules under which retail take-back will operate.
Most existing EPR frameworks for textiles were built around an assumption that collection happens at communal infrastructure (street containers, charity bins, municipal points), with retail involvement treated as optional or supplementary. The evidence in this artefact suggests that assumption is now structurally incomplete. Retail collection produces material that arrives at sorters with information density an order of magnitude higher than communal streams, and produces a record of brand-attributable circular outcomes that no other collection mode can match. For a producer responsibility organisation seeking to raise both collection rates and feedstock quality, retail is not a supplementary channel; it is the channel where the marginal euro of EPR funding produces the most usable output.
The policy implication, offered for the reader's consideration rather than as advocacy: producer responsibility frameworks should consider whether their current rules adequately recognise the value retail collection produces, whether eco-modulated fees should reward higher-quality collection routes differently from volume-based collection, and whether the scope of what retailers are permitted to do (and required to do) under EPR adequately reflects retail's structural position. These are policy questions, not commercial ones, and they will be answered differently in different jurisdictions. The artefact's contribution is to surface the evidence that makes the questions worth asking.
The case study at /case-study/ describes the Netherlands pilot in operational detail. The cost-per-email comparison and ROI calculator at /roi/ works through the unit economics for a specific retailer's footprint. The author is reachable via /contact/ for a commercial conversation. The structural argument, however, stands on its own evidence and is offered for the reader's own use, whether that reader is a retailer, a recycler, a producer responsibility organisation, or a policy-shaper.
Tim Lee is Co-Founder and CEO of Utilitarian, a circular economy data platform that turns in-store product collection programmes into customer acquisition channels for retailers. The platform is currently deployed with INTERSPORT, Runnersworld, EK Sport, and FastFeetGrinded across the Netherlands, with rollout in development internationally. Tim writes weekly on the commercial case for take-back, the regulation and compliance reality of EU circular economy policy, and the behavioural design of the take-back moment. He is Technical Editor at TexSPACE Today. Reach him at linkedin.com/in/tplee or t.lee@utilitarian.world.
One conversation. We'll walk you through the data, the deployment, and the unit economics for your specific footprint.
Book a call →