Why the European textile system produces the result it does, what closes the gap, and why retail is the channel best positioned to do it.
The European textile collection rate rose from approximately 10.7% in 2016 to 15.0% in 2022, a gain of about 0.72 percentage points per year. At that rate of improvement, the EU's 2030 base case ambition of 50% gets reached around 2071. The 2030 upside case of 80% gets reached around 2113. The trajectory is moving in the right direction; the rate at which it is moving is not on track to close the gap on the timeline policy needs.
The 85% of post-consumer textile waste that ends up in mixed household waste is not the cohort being lazy or apathetic. It is the cohort being mis-served by a system that asks for the wrong behaviour pattern. The existing collection model asks consumers to batch: accumulate worn textiles at home, then take them somewhere. Two failure modes follow predictably. The wardrobe stash, where the bag never fills enough and the textiles end up in the household bin during a clear-out. The guilt-purge, where the bag eventually gets dropped at a charity shop on the way to something else, mixed and unverifiable. Both produce bursty, contaminated, unidentified volumes, the input shape that produces the 36% sorting yield.
A different model is operationally available. Retail take-back at the point of replacement purchase converts the ask from an accumulation behaviour to an in-flight behaviour. 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. The cadence of recovery matches the cadence of purchase. Across nine months of operational deployment, identified customers return roughly 1.5 items each, with email captured for around 9 in 10 scans, and over a quarter of scans coming from repeat customers. This is the empirical signature of the drip model.
The structural argument that follows from these observations is that retail is the channel best positioned to close the gap, on volume as well as on quality. If the European fibre-to-fibre recycling infrastructure plan does not include the retail counter as a collection node, the plan has a blind spot. This artefact sets out the evidence, the arithmetic, the operational data, and the policy implications that follow.
The European Environment Agency's most recently published figures show the EU separate-collection capture rate for post-consumer textiles rising from approximately 10.7% in 2016 to 15.0% in 2022 [3]. That is a gain of 4.3 percentage points over six years, or about 0.72 percentage points per year. The number is empirical, sourced, and independent of any commercial party's interpretation.
Project that rate forward. At 0.72 percentage points per year, the EU's 2030 base case ambition of 50% gets reached around 2071. The 2030 upside case of 80% gets reached around 2113. The arithmetic is mathematically straightforward; the implication is not. The system is moving in the right direction. The pace at which it is moving is not on track to close the gap on the timeline policy needs.
The trajectory is moving in the right direction. The rate at which it is moving cannot close the gap on the timeline policy needs.
One framing point on the figures used here. The same European textile system supports two different headline rates, depending on the denominator. The EEA capture rate of approximately 15% in 2022 measures the share of all household textile waste that is separately collected, with mixed waste fractions in the denominator. Industry analysis from Boston Consulting Group and ReHubs reports a collection rate around 33% on a more restrictive denominator that excludes some mixed-waste fractions [13]. Both figures are correct against their own definitions. This artefact uses 15% throughout, because it is the stricter and more conservative figure, the one most cited in EEA reporting, and the one from which the trajectory arithmetic above is derived. Where the 33% figure appears later in the artefact, it is named and the denominator difference is explained on the spot.
The remainder of the artefact addresses three questions in turn. Why does the existing system produce this trajectory? What kind of mechanism would close the gap? And what does the operational evidence say about whether such a mechanism is available, and at what cost?
What is happening between the 16 kg per person of textile waste that EU residents generate each year and the less than 1% that re-enters textile production through textile-to-textile recycling [13]? The simplest way to see it is to walk through the system in four stages. Each stage tightens the constraint on the next. The aggregate fate at the end is the result of those constraints stacking.
The European population, in respect of end-of-life behaviour for textiles, is not uniform. It does not divide cleanly into "those who recycle" and "those who do not." It divides into roughly three cohorts whose dispositions are different in kind, not just degree.
The prosocial baseline, roughly 18% of the population, already participates in separate collection. They take used clothing to charity bins, drop bags at municipal collection points, donate to the Salvation Army, post bags through Vinted. The system as currently designed is built for them. They are, by their behaviour, the cohort the existing collection infrastructure was built to serve.
The middle cohort, roughly half the population, is willing in principle but not currently activated by the design of the existing channels. They are the customers who say yes in surveys to "do you support textile recycling?" and then put the worn t-shirt in the household bin three weeks later because the trip to the charity bin never happened.
The indifferent cohort, roughly 32%, has no prosocial baseline to displace. The system as it currently exists asks them to act against type and rewards them with nothing they value.
Population shares above are indicative. The analytical point is structural: any policy or commercial mechanism designed for the prosocial cohort will leave the other two unaddressed by design, and any mechanism designed to reach the middle and indifferent cohorts will need a different shape from the one the existing system has.
Move from population shares to waste volumes. The cohort distinction matters here because each cohort sends its volume to different destinations, and the aggregate result is a function of how the volumes flow.
The aggregate result, ~15% collected, ~85% straight to mixed household waste, is what the EEA publishes. The schematic shows how that aggregate is composed. It is not the prosocial cohort failing. It is the middle and indifferent cohorts being absent from the separate-collection stream by design.
Of the volume that does enter the separate-collection system, the European average sorting yield is 36%. That figure is empirical and well-cited [13]. The remaining 64% is downcycled, exported, or routed to lower-value end uses. The breakdown of the 15% by channel matters here: most of the collected volume is communal, and most of that volume runs through the 36% yield.
A note on the illustrative figure. The 36% communal sort yield is empirical and well-cited; the 85% retail figure used here is not. We use 85% as an illustrative anchor because the structural reasons sort yield is currently low do not apply in the retail-collection case. Communal collection arrives in batches of mixed composition, contaminated by months of household conditions, in unverifiable condition, with no information about what specific products are inside the bag. Sorting that material is the constraint the recycling industry has been describing for years. Retail collection arrives one item at a time, identified at the counter, photographed, classified, and tied to a specific customer. The four failure modes that drive the 36% figure (contamination, mixed composition, unverifiable condition, unknown provenance) are not present at retail. A meaningfully higher sort yield is what the input shape allows. The specific figure is illustrative; the direction of the difference is not.
Each of the four failure modes deserves a sentence on its own, because each is doing work in producing the 36% figure, and each is structurally absent at the retail counter.
The retail counter, in other words, is not a marginally better collection node. It is structurally a different kind of input. The same physical material reaches the sorter in a fundamentally different state.
Combine the previous three stages and the aggregate fate of post-consumer textile waste in Europe falls out of the arithmetic. Approximately 85% never enters separate collection at all and goes directly to mixed household waste. Of the 15% that does enter separate collection, 64% is lost in sorting (about 9.6 percentage points of the original total). The remainder, around 5 percentage points of the original total, reaches reuse or open-loop recycling. Less than 1 percent reaches textile-to-textile recycling [13].
~94% incineration or landfill is a striking number. It is also the arithmetic that follows from the underlying figures, taken together. The 85% direct mixed-waste figure is straight EEA. The ~9% sort loss is derived from the 36% sort yield applied to the 15% collected: 0.64 × 0.15 = 0.096, rounded to 9 percentage points of the original total. The combined ~94% can be defended in any policy or industry meeting because the underlying inputs are.
The four-stage walkthrough is what the existing system produces. The remainder of the artefact addresses why, and what changes if a different mechanism is added.
The 85% of post-consumer textile waste that ends up in mixed household waste is often discussed as a system performance gap, as if the cohort had been served by the existing system and had failed to participate. That framing misses the structure. The 85% is not the cohort being lazy or apathetic. It is the cohort being mis-served by a system that asks for the wrong behaviour pattern.
The existing collection system asks consumers to batch. Accumulate worn textiles at home until enough to justify a trip, then take them somewhere: a charity shop, a street container, a municipal collection point. The batch model is intuitive from the operator's point of view because it concentrates volume per visit, and intuitive from the prosocial-baseline cohort's point of view because they have already adopted a routine that makes the trip happen.
For the middle cohort, the batch model produces two failure modes that follow predictably from its design.
The wardrobe stash. The bag never fills enough to justify the trip, or the trip never gets prioritised against the rest of the week. The textiles accumulate in a cupboard, a spare-room corner, a pile next to the washing machine. Eventually a clear-out happens, often during a move, a redecoration, or a seasonal change, and the textiles enter the household bin alongside other items being decluttered. The volume never enters the separate-collection stream at all.
The guilt-purge. The bag does eventually get taken somewhere, often a charity shop on the way to something else. By the time it arrives, it is a mixed batch of unsorted items in unverifiable condition. The bag has spent months absorbing household conditions, mixing items of different categories, without any information about composition, age, or condition attached. The volume enters the separate-collection stream, but in a form that produces the input shape the recycling industry has been describing as the constraint for years.
Both failure modes produce bursty, contaminated, unidentified volumes. Both are predictable consequences of asking the cohort to perform an accumulation behaviour that has no naturally arising trigger. The wardrobe stash is the failure mode that loses the volume to mixed waste; the guilt-purge is the failure mode that delivers the volume to sorting in a state that produces the 36% yield. Either way, the design of the moment is what produces the outcome, not the disposition of the consumer.
Retail take-back at the point of replacement purchase converts the ask from an accumulation behaviour to an in-flight behaviour. 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.
This is not a refinement of the batch model. It is a different mechanism class. The infrastructure that supports batch collection (street containers, charity bring-points, civic amenity sites) cannot adopt drip; the moment drip requires is the in-store moment. The moment a customer is in the retailer's store, with a product the retailer sold, choosing to replace it, is not available to communal infrastructure by definition. The drip model is hosted at retail or it is not hosted at all.
The cadence of recovery matches the cadence of purchase. One item at a time, in known condition, identified at the counter, exactly the input shape the recycling industry has been waiting for.
The pull quote captures the synthesis. The way retail returns happen mirrors the way retail purchases happen, which is what produces the feedstock characteristics the recycling industry needs. The customer's replacement rhythm, which the retailer already understands at the level of their CRM and inventory planning, becomes the cadence at which used material flows back into the recovery system. The two cadences are the same cadence.
The next section sets out why the retailer is the operator best positioned to host this mechanism, and what it takes to operate it well.
Set aside the volume question for the moment. The reasoning that has historically positioned retail take-back as a marginal contributor to European textile collection deserves direct engagement, and is addressed in the policy implications later. This section addresses a different question: assuming the volume question can be closed, what makes retail the right operator? What can a retailer do that a charity bin or municipal collection point cannot?
Six things, all of which compound.
The customer is in the store because they need a new pair. The retailer is the only operator who sees them at that exact moment of replacement intent. Communal collection has no access to this moment; it sees the customer at a moment of disposal intent that is disconnected from purchase. The replacement moment is the in-flight behaviour the drip model converts the ask into. It is what makes one-and-a-half items per customer per nine months a realistic operating signature, against a communal-collection signature of one bursty bag per year per participating consumer.
A retailer can offer a discount on the new pair, store credit, loyalty points, free service (gait analysis, fitting, tuning), an exchange offer. The unit economics work because the incentive cost is offset by the value of the resulting transaction. Communal collection has no transaction to attach an incentive to. This means the commercial machinery of retail (marketing budgets, CRM systems, loyalty programmes, transaction infrastructure) can be applied to the take-back problem at zero marginal cost. Communal collection has no equivalent machinery. The marginal euro of investment in retail-hosted take-back leverages systems that already exist; the marginal euro in communal infrastructure has to build the systems alongside.
The retailer already has a CRM relationship with most customers (loyalty programmes, email lists, app installs). The take-back interaction adds a high-quality data event to that existing relationship: a new email, a verified product, a confirmed customer profile, a behavioural signal of replacement-consideration. Communal collection produces an anonymous bag. The same physical item, surrendered through retail, becomes a record attached to an identifiable customer with an existing relationship surface to use.
This is also the route by which the retailer operationalises the transparency premium. Field experiments published in Management Science by Buell and Kalkanci [1] find that transparency about a retailer's internal responsibility practices lifts purchase probability by 6–46% under field-experiment conditions. The take-back interaction 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. There is no philanthropy mediating the interaction; the interaction is the transparency. The 6–46% lift is the empirical anchor for the commercial-uplift case the existing relationship surface allows the retailer to capture.
Customers trust the retailer with their replacement decision. Extending that trust to the disposal of the worn product is a small step, not a leap. Communal collection requires trust in a third-party operator with no existing customer relationship. The trust transfer means the retailer can establish a take-back programme with a fraction of the consumer-education cost a third-party operator would face; the relationship is already there, and the take-back interaction is an extension of it.
A retail employee can have a conversation about the worn product, the new product, the customer's use case. That interaction is what makes the moment behavioural rather than purely transactional. Bins do not have conversations. Behavioural reinforcement at the moment of return is a lever only retail has. This is also where the take-back moment becomes a sales conversation: the staff member asking "do you want to replace it?" is doing the work of converting a recovery interaction into a transaction interaction, and the trust transfer above is what makes the question land.
Identified take-backs feed back into product design, sustainability reporting, EPR compliance, and brand storytelling. The retailer that operates take-back becomes a node in their own brand's circular value chain. Brand-attributable circular outcomes are visible in a way they are not when the same items reach the recycler through a communal stream. For ESRS E5 disclosure, for closed-loop programmes, for the verifiable evidence that emerging EU disclosure standards now require [6], the data captured at retail is the data that satisfies the requirement. Communal collection produces aggregate weight reports that, however accurate, are not attributable to specific products at specific points in the operating boundary.
The six capabilities compound. The replacement moment supplies the access. The incentive asymmetry pays for the participation. The relationship surface captures the value. The trust transfer reduces the activation cost. The staff interaction reinforces the behaviour. The data feedback loop converts the interaction into a usable record across marketing, sustainability, and the recycling chain. None of these six is operationally available to a communal collection point. All six are operationally available to a retailer who runs a take-back programme designed for the purpose.
The drip-model argument above is structural. It would remain structural in the absence of operational evidence. We have nine months of operational evidence (22 August 2025 to 9 May 2026) from the active platform deployment across our retail partners in the Netherlands.
Across this period, the deployment data shows a consistent operational signature. The figures below are reported as aggregate patterns from across the active partner footprint; specific store-level statistics, individual partner outcomes, and per-store breakdowns are commercial-in-confidence and are not disclosed here.
The 1.5-items-per-customer figure is the empirical signature of the drip model. It is the operational shape that follows from the cadence-of-recovery argument. If the model worked the way the batch model works, we would see one large interaction per customer concentrated at the end of an accumulation period. We do not. We see a steady distribution of single-item returns, spread across the deployment window, scaling with the underlying replacement-purchase rhythm of the partners' customer bases.
The 90% identification rate matters because it is the value-capture figure for the marketing team. An interaction in which the customer surrenders a product but no email is captured produces compliance value (the recovery happened, the audit trail exists) but no commercial value to the retailer. An interaction in which the email is captured is one that produces commercial value alongside compliance value. The 90% figure says, in operational practice, that the channel produces both at once, not one or the other.
The 27% repeat-customer share is what the drip model produces over time. Retail take-back, designed properly, is not a one-time interaction but a recurring relationship. The implication for CRM value, lifetime value calculations, and the unit economics of the customer-acquisition argument is that the retail-take-back customer is not equivalent to a paid-social-acquired email captured once; it is a customer in a recurring relationship surface. Over a multi-year horizon, the repeat-customer pattern compounds the value of each initial capture.
What the deployment data does not establish: the long-run conversion rate of take-back-counter emails to subsequent purchases at scale, the long-run retention rate of repeat take-back behaviour, or the steady-state items-per-customer figure once the deployment matures past nine months. These are questions for further measurement as the platform scales. What the deployment data does establish is that the drip-model signature, predicted by the structural argument, is what the operational data shows.
The retailer reading this section is asking a different question from the policy reader. The policy reader is asking whether the system can be moved. The retailer is asking whether their store, specifically, captures value that justifies operating the channel. The answer is the unit economics, anchored by the transparency-premium evidence and worked through against the deployment cost structure.
Buell and Kalkanci's Management Science field experiments find that transparency about a retailer's internal responsibility practices lifts purchase probability by 6–46% under conditions designed to test exactly that [1]. The effect is robust across two field experiments and complementary online experiments. It is not a survey of stated preferences. It is a measurement of behaviour. 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 direct. 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. The interaction is the transparency, and the transparency is the interaction. The 6–46% range is the band within which a properly designed in-store take-back programme can expect to lift purchase probability for participating customers.
One framing point. The mechanism does not require additional staff time to operationalise. 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 if asked, but the design does not require them to handle the product or perform the disclosure themselves. The retailer 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 cost-per-verified-customer-email captured at the take-back counter sits in the €1–3 range under the deployment conditions reflected in the active partner footprint, against industry benchmarks of €12–25 from paid social and €18–31 from website popups [9] [10] [11] [12]. 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.
Consider a retailer running a take-back programme across 30 stores. At an email-capture rate of approximately 90% (consistent with current deployment experience), and at a programme volume of 30 take-back interactions per store per month, the channel produces 810 verified customer emails per month, 9,720 per year, at a unit cost in the €1–3 range. The same 9,720 emails through paid social would cost €116,640 to €243,000 per year on the lower-bound benchmark. The take-back-counter cost is in the range of €9,720 to €29,160 per year for the same volume. The cost saving alone, before any consideration of the transparency premium or the higher conversion quality of replacement-window emails, is in the high six figures annually for a deployment of this scale.
The same data simultaneously satisfies the sustainability team's compliance requirement under ESRS E5, the recycler's feedstock-qualification requirement, and the marketing team's customer-acquisition pipeline. One poster. One twenty-second interaction. Three teams' deliverables produced from the same data, captured once.
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.
The European fibre-to-fibre recycling investment story, told from the recycler's side, is a story about plants. Renewcell raised €200M and went bankrupt because feedstock never arrived. Multiple textile-to-textile pathways are technically demonstrated. The constraint on closing the less-than-1% gap is not the absence of recycling technology, and it is not the absence of investment in recycling capacity. 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.
This is the same constraint the four-stage walkthrough above produces. The 36% sort yield is the manifestation of the four failure modes (contamination, mixed composition, unverifiable condition, unknown provenance) on the input side of the recycling industry. The recycler is the constrained party at the receiving end of an upstream input shape that the existing collection model produces. Fixing the recycler will not fix the upstream input. Fixing the upstream input is what produces feedstock the recycler can operate on at industrial scale.
If your fibre-to-fibre recycling infrastructure plan does not include the retail counter as a collection node, the plan has a blind spot.
This is not a rhetorical line. It is an observation about what infrastructure is. A fibre-to-fibre recycling system is not the recycler alone; it is the chain that produces the input the recycler can operate on. The retail counter is the node in that chain where the four failure modes are not present, where the material arrives identified, attributed, and in known condition. A plan that funds plants without funding the upstream node that produces the input shape those plants can operate on is a plan that funds the wrong end of the chain. The plants will still be built. The plants will still be unable to operate at design throughput. The feedstock paradox repeats.
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 ~15% 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, because the trip the consumer is making anyway is the trip the in-flight behaviour piggybacks on.
For the brands whose products are entering the stream, the same observation cuts a different way. Retail is also the route by which a brand's own materials become identifiable as the brand's own when they arrive at recyclers. This matters for circular-content claims, for closed-loop programmes, and for the verifiable evidence that emerging EU disclosure standards now require. Communal collection produces aggregate weight, untraceable to brand. Retail collection produces brand-attributable circular outcomes from the moment of return.
This 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. What follows are policy implications, offered for the reader's consideration rather than as advocacy.
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 in three respects.
The bulk-versus-tail mental model treats communal collection as the volume channel and retail as the long-tail channel. The bulk-versus-tail framing holds for the period in which retail take-back has been narrow, brand-specific, voluntary, and operationally minimal. Under those conditions it is not unreasonable. The framing fails when retail take-back is designed differently. 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 policy-design implication is that the volume contribution from a properly designed retail channel is materially larger than the bulk-versus-tail framing predicts, and the policy framework should reflect this.
EPR fee structures designed around volume produce a different system than those that recognise the quality of the material recovered. A tonne of communal-collected mixed material that arrives at a sorter at 36% yield is not the same as a tonne of retail-collected identified material that arrives at a much higher yield. Eco-modulated fees that reward higher-quality collection routes differently from volume-based collection would change the incentive structure on the retailer side, the recycler side, and the brand side simultaneously. This is a policy design question, not a commercial one, and is offered for the reader's consideration.
The data captured at the take-back counter has a parallel destination on the disclosure side. ESRS E5 [6] requires disclosure of resource outflows, the destination of waste, and the extent of circular economy practices, in a form that is verifiable. Aggregate weight reports produced by recyclers do not satisfy the requirement directly, because they are not attributable to specific products at specific points in the operating boundary. Retail take-back records (brand, model, category, store, timestamp, circular outcome) sit at the right granularity for E5 disclosure. The policy framework should consider whether the rules under which retail take-back is permitted, and required, adequately reflect the data-attribution function that retail uniquely produces.
The artefact's contribution on policy is to surface the evidence that makes these questions worth asking. The questions will be answered differently in different jurisdictions; the structural argument that makes them serious questions is the same in all of them.
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 platform deployment with active retail partners in the Netherlands across nine months from 22 August 2025 to 9 May 2026. Industry cost-per-email benchmarks [9] [10] [11] are cited as commercial comparators, not as findings.
Deployment data is reported as aggregate patterns from across the active partner footprint. Specific store-level statistics, individual partner outcomes, per-store breakdowns, and brand-level data are commercial-in-confidence and are not disclosed here. The 1.5-items-per-customer figure, the 90% identification rate, and the 27% repeat-customer share are presented as the operational signature of the drip model in deployment, not as a claim about long-run unit economics at scale. The cost-per-email range of €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 steady-state long-run figure.
The artefact distinguishes schematic from empirical charts. The bell curve and the four-stage flow charts (stages 1, 2, 3, and 4) are labelled schematic. The trajectory line chart is labelled empirical. Schematic charts are honest visual interpretations, not measurements; empirical charts are sourced and defensible. Where a schematic chart includes an illustrative figure (the 85% retail sort yield in stage 3), the qualifier appears inside the chart so screenshots travel with the qualifier intact.
The 2071 and 2113 figures derive from a linear projection of the EEA-published rate of improvement (10.7% in 2016 to 15.0% in 2022, a gain of 0.72 percentage points per year) extrapolated forward to the dates at which the projection intersects the 50% and 80% ambition lines. The arithmetic is mathematically straightforward. The projection assumes the rate of improvement continues at its historical pace; both faster and slower scenarios are possible, and the figures are presented as the trajectory under the rate observed over 2016–2022, not as a forecast. Where future EEA data points become available, the projection should be re-anchored.
The EEA capture rate (~15% in 2022) and the BCG/ReHubs collection rate (~33% in 2025) are both correct against their own definitions. The EEA figure measures the share of all household textile waste that is separately collected, with mixed waste fractions in the denominator. The BCG/ReHubs figure uses a more restrictive denominator that excludes some mixed-waste fractions. The artefact uses 15% as the primary baseline because it is the stricter and more conservative figure, the one most cited in EEA reporting, and the one from which the trajectory arithmetic is derived. Where the 33% figure appears, the denominator difference is named and explained in place.
The peer-reviewed studies cited here measure effects under their own experimental conditions. Buell and Kalkanci's transparency-premium finding [1] is from field and online experiments in retail contexts that are not specifically take-back contexts. McKie et al.'s clothing take-back finding [2] is measured under information and convenience treatments designed to isolate those variables, not in a complete operational deployment of the kind the platform represents. Wollbrant, Knutsson and Martinsson's S-curve finding [14] 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 operational evidence is interpretive, not statistical.
The argument in sections 3, 4, 5, and 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.
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.
See the impact thesis built on this paper's findings →
Sources cited
1. Buell, R. W., & Kalkanci, B. (2021). How transparency into internal and external responsibility initiatives influences consumer choice. Management Science, 67(2), 932–950. doi:10.1287/mnsc.2020.3588
2. McKie, E. C., Sáez de Tejada Cuenca, A., & Agrawal, V. (2025). The role of information, rewards, and convenience in take-back programs for clothing. Manufacturing & Service Operations Management, 28(2), 362–380. doi:10.1287/msom.2023.0561
3. European Environment Agency (2024). Management of used and waste textiles in Europe's circular economy. EEA Briefing no. 03/2024. Copenhagen: EEA. eea.europa.eu
4. European Environment Agency. Textile waste generation per person in the EU per year. Circularity Metrics Lab (2022 data). eea.europa.eu
5. European Commission. Waste Framework Directive (Directive 2008/98/EC, as amended). environment.ec.europa.eu
6. European Financial Reporting Advisory Group (EFRAG). ESRS E5 Resource Use and Circular Economy. efrag.org
7. European Topic Centre on Circular Economy (ETC CE). Textile waste management in Europe's circular economy. ETC CE Report 2024/5. eionet.europa.eu
8. Utilitarian. Netherlands deployment, in-store shoe take-back with INTERSPORT, Runnersworld, EK Sport, and FastFeetGrinded (2025–2026). utilitarian.world/case-study
9. Data & Marketing Association. Marketer Email Tracker 2026. dma.org.uk
10. Opensend. Cost Per Acquisition by Channel Statistics for eCommerce Stores. opensend.com
11. WordStream. Facebook Ads Benchmarks 2025. wordstream.com
12. Utilitarian deployment data (2025–2026). Internal cost-per-verified-email figures from Netherlands deployment, reported as a directional €1–3 range.
13. Boston Consulting Group and ReHubs (2026). Austruy, E., Giraud, T., Manuelli, N., van de Kerkhof, R., Wiener, E., Wong, R., Sauvaget, M., and Hamidouche, Z. Advancing Textile Circularity in Europe: The Case for System-Level Scale-Up. BCG Publications, 23 March 2026. bcg.com/publications/2026/france-advanc…
14. Wollbrant, C. E., Knutsson, M., & Martinsson, P. (2022). Extrinsic rewards and crowding-out of prosocial behaviour. Nature Human Behaviour, 6(6), 774–781. doi:10.1038/s41562-022-01293-y
One conversation. We'll walk you through the data, the deployment, and the unit economics for your specific footprint.
Book a call →