EN 18220:2026, the European standard for digital product passport data carriers, recognises exactly five technologies: QR code, Data Matrix, NFC, HF RFID and RAIN RFID. Digital watermarks are not among them. That single fact reshapes how a textile brand should plan its digital product passport, because the carrier you print, weave or embed today has to survive a delegated act that has not been written yet.
Most articles on this topic stop at "you can use a QR code or an NFC chip." That is true and useless. The real decision involves identifier granularity, wash durability, per-unit cost across a 200,000-piece season, whether a recycler five years from now can read anything at all, and whether one carrier can simultaneously clear the checkout, satisfy the ESPR and drive inventory accuracy. This article works through each of those in turn.
What is a DPP data carrier?
A DPP data carrier is the physical link between a garment and its digital record. Under the Ecodesign for Sustainable Products Regulation (Regulation (EU) 2024/1781, "ESPR"), the passport is not a file that lives only on a server. It must be reachable from the object itself, through something a person or a machine can scan, tap or read.
The regulation splits this into two parts that are easy to confuse. The unique product identifier is the data — a globally unique, persistent string that names this model, this batch or this individual item. The data carrier is the physical encoding of that identifier: the printed square, the woven label, the embedded chip. One is information, the other is the vehicle.
ESPR Article 9 sets the general requirements: a unique product identifier, a data carrier, data in machine-readable format, and conformity with interoperability standards. The carrier must be physically present on the product, its packaging or its documentation. The data behind it must be accurate, complete and up to date.
What the regulation deliberately does not do is name a winning technology. ESPR is written to be technology-neutral, and the specific carrier for each product group is left to that group's delegated act. For textiles, that act does not yet exist.
Why the identifier matters more than the carrier
The identifier decides what the carrier has to be capable of. EN 18219:2026 defines three levels of granularity: model, batch and item. One identifier for a furniture model, one for a production batch, one for each unique physical piece.
A model-level identifier is the same on every garment of that style, so you can print it once on an origination plate and forget about it. An item-level identifier is different on every single piece, which means variable data at the point of label production. That is a manufacturing problem, not a design problem, and it is where most carrier decisions are actually won or lost.
EN 18219 also requires identifiers for economic operators and facilities, not just products. So a full textile passport implementation identifies who made it and where, not only what it is.
Which data carriers does the EU actually allow?
The EU allows any carrier the relevant delegated act permits, but the European standard that operationalises the DPP names five. EN 18220:2026, "Digital product passport — Data carriers", covers QR code, Data Matrix, NFC, HF RFID and RAIN RFID. Each gets a distinct role in the standard.
QR code is described as a high-capacity 2D code readable by most cameras and smart devices. It is the general-purpose option: universally readable, zero-cost to print, and the only carrier a consumer can use without a specific handset feature.
Data Matrix is a high-density 2D code intended for direct part marking where print area is small. In textiles it matters less than in electronics or medical devices, but it becomes relevant on hardware components, trims and small labels where a QR code of adequate size will not fit.
NFC is short-range contactless communication, tap-to-read. It works on modern smartphones without an app, which makes it the most consumer-friendly chip option, and it carries a premium-experience connotation that QR does not.
HF RFID is high-frequency RFID for short-range interactions and shares the 13.56 MHz band with NFC. RAIN RFID is UHF RFID for long-range, fast identification in supply chains — the technology already tagging hundreds of millions of garments a year for inventory, not compliance.
What is not on the list
Digital watermarks are absent from EN 18220. This is worth dwelling on, because watermarking has real industrial validation elsewhere. In the HolyGrail 2.0 trials at the Hündgen material recovery facility in Swisttal, Germany, a digital watermark system recorded 5.66 million detections across 5,949 unique SKUs over 100 days, averaging around 56,000 items per day at detection accuracy consistently above 90%.
That result is for plastic packaging, not textiles, and the technology does not currently appear in the European DPP data carrier standard. A textile brand betting on watermarks as its compliance carrier today would be betting against the standard. As a supplementary sorting aid, that is a different and more defensible proposition.
Key finding: EN 18220:2026 recognises exactly five data carriers, namely QR code, Data Matrix, NFC, HF RFID and RAIN RFID, and digital watermarks are not among them.
What does EN 18220:2026 require from a data carrier?
EN 18220 organises its requirements into nine categories rather than prescribing one technology. The standard covers data encoding, scannability, accessibility, data carrier design, durability, quality, performance, placement and marking, and a set of further considerations.
The framing that matters most for textiles is the standard's list of who the carrier must serve. It has to work for consumers, businesses, repairers, recyclers, market surveillance authorities and automated systems. Six audiences, with radically different equipment and radically different moments in the product's life.
A consumer has a phone in a shop. A repairer has a workbench and a garment that has been washed sixty times. A recycler has a conveyor moving at speed and a shredded, unsorted bale. Market surveillance has a customs desk and a pallet. Any carrier strategy that satisfies only the first audience is not a DPP strategy, it is a marketing landing page.
Placement and marking is the category most textile brands underestimate. A carrier printed on a hangtag is gone the moment the customer cuts it off, which is typically within an hour of purchase. A carrier on the outer packaging is gone at the same moment. Only a carrier integrated into the garment — care label, woven label, heat transfer — persists into the repair, resale and recycling phases that the standard explicitly names.
The standards family behind the carrier
EN 18220 does not stand alone. CEN/CLC/JTC 24 developed eight horizontal DPP standards under mandate M/604, and they interlock:
| Standard | Scope |
|---|---|
| EN 18216:2026 | Data exchange protocols and data formats |
| EN 18219:2026 | Unique identifiers |
| EN 18220:2026 | Data carriers and links between physical product and digital representation |
| EN 18221:2026 | Data storage, archiving and data persistence |
| EN 18222:2026 | APIs for passport lifecycle management and searchability |
| EN 18223:2026 | System interoperability |
| EN 18239:2026 | Access rights management, information system security, business confidentiality |
| EN 18246:2026 | Data authentication, reliability and integrity |
Six of the eight have been cited in the Official Journal of the European Union, which confers presumption of conformity. The two security standards moved through formal vote and approval during 2026. These are horizontal standards: they define how a DPP works mechanically, not which data each product must carry. The data points stay in the product-group delegated acts.
How do QR, NFC and RFID compare for textiles?
For textiles specifically, QR wins on cost and reach, NFC wins on consumer experience, and RAIN RFID wins on operations. No single carrier wins outright, which is why the honest answer is usually a combination.
| Criterion | QR code | NFC | RAIN RFID (UHF) |
|---|---|---|---|
| Marginal cost per garment | Effectively zero if printed on an existing care label | ~$0.05–0.12 for a woven RFID/NFC care label; $0.15–0.25 for basic heat-seal tags | Same label economics; $0.40–0.75 for sewn-in tags rated 500+ wash cycles |
| Consumer readability | Any smartphone camera, no app | Modern smartphones, tap-to-read, no app | Not readable by consumer phones |
| Read range | Line of sight, ~5–30 cm | ~1–4 cm | Several metres |
| Bulk reading | One at a time | One at a time | Hundreds of items per second |
| Item-level serialisation | Requires variable-data printing | Native — each chip has a unique ID | Native — each tag has a unique EPC |
| Wash durability | Depends entirely on the substrate and print method | Chip survives; antenna and encapsulation are the weak point | Industrial sewn-in tags rated 200–500 cycles at up to 220 °C |
| Works when the garment is a shredded bale | No | No | Partially, until the tag is destroyed |
| Counterfeiting resistance | Low on its own — a QR is trivially copied | Higher — chip UID and crypto features possible | Higher — but tags are clonable without crypto |
| Retail POS use | Yes, via GS1 Sunrise 2027 | No | Increasingly, at self-checkout |
| Regulatory status | Named in EN 18220; QR is the mandated carrier for battery passports | Named in EN 18220 | Named in EN 18220 |
Read that table as three different jobs rather than three competing products. QR is the compliance and consumer job. RAIN RFID is the supply chain and store operations job. NFC is the brand experience job, and it is optional.
Where the comparison usually goes wrong
Brands routinely compare these carriers on "which one holds more data". That is the wrong axis. Under the EU model none of them hold the passport — they hold a pointer to it. The passport data lives in a system the economic operator controls, reachable through an API, with the carrier supplying only the identifier.
This means the storage capacity of an NFC chip is close to irrelevant for compliance. What matters is whether the identifier encoded on it is globally unique, persistent, machine-readable and usable across systems, which is exactly what EN 18219 demands. A 2 KB chip and a printed 30-character URL do the same regulatory job.
Key finding: A QR code printed on a care label you already produce adds effectively nothing to unit cost, while chip-based carriers run from $0.05 to $0.75 per garment.
What QR code specification does a textile label need?
A textile QR code needs more error correction and more quiet zone than a packaging QR code, because the substrate moves and abrades. Four parameters decide whether it still scans after two years of wear.
Error correction level. Use level Q, which tolerates around 25% damage, rather than the level L or M common in print. Abrasion, fading and creasing at fold lines destroy modules unevenly, and the extra redundancy is what keeps a worn label readable.
Module size and physical dimensions. Keep the printed code at 15 mm square or larger for a care label, and never below 10 mm. A GS1 Digital Link URI with GTIN, batch and serial is a longer payload than a plain product code, which raises the module count and therefore the minimum printable size.
Quiet zone and contrast. Leave a clear margin of at least four modules around the code, free of stitching, text and folds. Print dark on light with real contrast, because low-contrast tone-on-tone printing looks elegant in a mock-up and fails under shop lighting.
Placement relative to the fold. Position the code so it does not sit across a seam or a fold line, since repeated folding wears a straight line of modules and defeats error correction faster than general fading does.
Not sure your care label will still scan after fifty washes? We will run a free screening of one of your bestsellers →
Why does GS1 Digital Link matter for textiles?
GS1 Digital Link matters because it lets a single QR code serve retail checkout and DPP access at the same time. First published in 2018, the standard encodes GS1 identifiers as ordinary web URIs, so a product identifier becomes a web address a browser can resolve.
The fundamentals of GS1 Digital Link, the EU registry and DPP access rights are already covered in our complete guide to DPP compliance for textiles. This section assumes them and concentrates on what the resolver layer means specifically for a carrier decision.
The structure is a resolver domain followed by GS1 Application Identifiers and their values. /01/ carries the GTIN, /10/ the batch or lot, /21/ the serial number, so a fully item-level URI looks like https://example.com/01/07501234567890/10/BATCH-2026A/21/SN-001. The identifiers get more granular from left to right, which maps cleanly onto the model, batch and item levels in EN 18219.
A GS1-conformant resolver sits behind that URI. It behaves like DNS for products: it takes the identifier and redirects — usually with a 302 or 307 — to whichever destination fits the request, or returns a LinkSet of everything available. Each destination is tagged with a link type drawn from the GS1 Web vocabulary, so a scan can be routed to a product information page, a warranty registration, a care guide or the digital product passport depending on who is asking.
That routing layer is the practical reason to use Digital Link rather than a bare URL. One physical carrier, many destinations, changeable after the garment has shipped. If a brand prints a hard-coded URL on 200,000 care labels and then migrates its passport platform, those labels are dead. With a resolver in between, the migration is a configuration change.
Sunrise 2027 and the double-duty QR code
GS1's Sunrise 2027 initiative has retailers across 48-plus GS1 Member Organisations committing to accept 2D barcodes at point of sale by the end of 2027. The commercial consequence for textiles is direct: a QR code carrying a GS1 Digital Link URI can replace the linear barcode at the till and serve the passport, from the same printed square.
That is a rare case where a compliance cost has an operational payback. The alternative — a barcode for the till and a separate QR for the passport — means two carriers, two label areas and two sets of data governance, for no benefit. The timing is also convenient, since Sunrise 2027 and the expected textile delegated act land in roughly the same window.
How durable does a carrier have to be on a garment?
Durability is the requirement that separates textiles from every other DPP product group. A battery is not laundered. A garment is washed, tumbled, ironed, worn against skin, folded, resold and eventually shredded, and the passport is supposed to remain reachable through all of it.
EN 18220 names durability as an explicit requirement category but, as a horizontal standard, does not set a textile-specific wash-cycle figure. That number will come from the textile delegated act, and until it does, brands are choosing durability targets on commercial judgement rather than legal instruction.
The industrial evidence is reassuring on the chip side. Industrial-grade sewn-in RFID tags used in rental uniforms and workwear survive 200 to 500 wash cycles at temperatures up to 220 °C. Soft woven RFID labels integrated into care labels withstand normal consumer laundering, which is a far gentler duty cycle than industrial laundry.
Printed QR codes are the more variable case, because durability depends entirely on substrate and print method. A thermal-transfer print on a cheap satin label can fade to unreadability well inside a garment's useful life. A carefully specified print on a durable woven or heat-transfer substrate performs very differently, and this is a specification question brands should be putting to their label suppliers now rather than after the delegated act lands.
The redundancy argument
The most robust textile approach is redundancy rather than a single perfect carrier. A printed QR on the care label handles consumers, repairers and market surveillance. A chip in the same label handles bulk operations and survives print abrasion. Both encode the same identifier, so both resolve to the same passport.
This costs more per unit than QR alone, which is why it makes sense for higher-value and longer-life products first. RFID and NFC tags are proportionate where the cost of embedding a chip is small relative to product value, and that logic will hold until chip prices fall further.
What does a data carrier cost per garment?
Carrier cost ranges from effectively zero to roughly $0.75 per garment, and the spread is driven by whether you are printing or embedding. Printing a QR code onto a care label you were already producing adds no material and no process step, so the marginal cost is close to nothing.
| Option | Indicative unit cost | When it makes sense |
|---|---|---|
| QR printed on existing care label | ~$0.00 marginal | Every product, always — this is the baseline |
| QR with item-level variable data | Low, plus a print-process change | When the delegated act or resale strategy requires item-level identity |
| Woven RFID/NFC care label | ~$0.05–0.12 | Mid-to-premium products, brands already running RFID |
| Basic heat-seal UHF tag | ~$0.15–0.25 | Retrofit onto existing production, seasonal lines |
| Sewn-in tag rated 500+ wash cycles | ~$0.40–0.75 | Workwear, rental, uniform, high-value long-life garments |
These are indicative ranges from the tag supply market and move with volume, chip supply and label construction. Treat them as an order of magnitude for budgeting, not a quotation.
The number that should actually drive the decision is not unit cost but total programme cost. Across a 200,000-unit season, a $0.08 woven RFID label is $16,000 in labels. The data infrastructure, supplier data collection, LCA modelling and passport hosting behind those labels will usually cost considerably more than the carriers themselves. Brands that agonise over cents per tag and then underinvest in supply chain data have optimised the wrong line.
How does the carrier actually get onto the garment?
The carrier enters production through the label supplier, and that is a longer and less flexible path than most brands expect. A care label is specified in the tech pack, approved as part of the trim package, and produced weeks or months ahead of the garment. Changing what is printed on it is not a software deployment.
For a model-level identifier the process is straightforward. The identifier is the same on every unit of that style, so it is fixed at artwork stage, printed with the rest of the care and composition information, and needs no per-unit handling. This is why model-level carriers are cheap: nothing in the existing label workflow has to change.
Item-level identity is a different operation. Every label carries a different code, which means variable-data printing and a system that assigns, records and hands off identifiers to the label supplier in the right sequence. The identifiers then have to be reconciled back to the garments they ended up on, or the passport points at the wrong item.
This reconciliation step is where item-level programmes usually fail. A brand generates a million serials, the label supplier prints them, and nobody records which serial went onto which size, colourway or production lot. The passport then exists but says nothing specific, which is the expensive version of doing nothing.
Who owns the carrier internally
Carrier decisions sit awkwardly between product development, sourcing and IT, and unowned decisions drift. Product development owns the tech pack and the trim specification. Sourcing owns the label supplier relationship and the cost line. IT or the sustainability team owns the identifier scheme, the resolver and the passport data.
The workable pattern is to make one team accountable for the identifier scheme and let the others consume it. The identifier scheme is a standard the brand publishes internally: what granularity, what format, who mints identifiers, how they are registered and how they reach the label supplier. Once that document exists, the carrier becomes a routine trim specification rather than a recurring negotiation.
Lead time is the reason to start this in 2026 rather than 2027. Trim specifications for a season are locked well before production, so a carrier decision made in one calendar year typically first appears on garments in the next. Two seasons of lead time is normal, and that alone consumes most of the gap between now and a 2028 compliance date.
Which carrier works for repairers, resellers and recyclers?
QR codes work for repairers and resellers; nothing currently works well for recyclers at scale. This is the least comfortable finding in the whole data carrier discussion, and it deserves a straight answer rather than an optimistic one.
A repairer or a resale platform handles an intact garment, one at a time, with a phone. A printed QR on the care label is exactly the right carrier for them, provided it is still legible after years of wear. Resale in particular is a strong argument for item-level identity, because a passport that can carry ownership history, repair records and authentication is worth considerably more to a resale platform than a static model-level datasheet. This is the part of the circular economy where the DPP has the clearest near-term commercial value.
Recycling is where it breaks down. The EU has required separate collection of textile waste since 1 January 2025, and sorting facilities are scaling up, but they identify fibres optically rather than by reading passports. Fibersort by Valvan in the Netherlands and SIPTex in Malmö, Sweden — the latter running Tomra's Autosort NIR/VIS and industrially operational since 2020 — use near-infrared spectroscopy, where each fibre type reflects light at characteristic wavelengths to produce a material fingerprint matched against a database.
NIR reads the surface of a textile. It does not read a bill of materials, it is well documented as struggling with blends, and it cannot tell a sorter about chemical finishes, dyes or trim composition. A DPP could supply exactly that missing information, which is precisely why EN 18220 lists recyclers among the audiences a carrier must serve.
The gap is physical. By the time a garment reaches a sorting line it may have lost its labels, and once material is shredded there is no carrier left to read at all. Bridging that gap needs either carriers that survive further into end-of-life than today's do, or a sorting model where the optical read looks up a passport rather than replacing it. Neither is solved, and a brand that claims its QR code closes the recycling loop today is overstating the case. The wider picture of where EU recycling capacity actually stands is in our article on circular economy in textiles.
Key finding: No carrier reliably survives to the sorting line today, so automated textile recycling still identifies fibres optically instead of reading a passport.
How does the DPP registry change carrier decisions?
The registry turns the identifier into something registered with the Commission rather than something a brand simply invents. Commission Implementing Regulation (EU) 2026/1778 of 16 July 2026 laid down the implementation arrangements for the digital product passport registry established under ESPR. It was published on 17 July 2026 and entered into force on 6 August 2026.
The Commission launched the registry itself, together with a separate testing environment, on 20 July 2026. Economic operators must first be verified through eIDAS identity means, with verification valid for up to three years. The registry currently covers batteries, construction products, toys and detergents, with future ESPR categories — including textiles — to follow.
For carrier decisions this has two consequences. First, the identifier you encode has to be one you can register and keep persistent, which rules out ad hoc internal SKU strings that get reissued between seasons. Second, the existence of a live registry and test environment means textile brands can now test the plumbing before their delegated act exists, using a real system rather than a mock-up.
The architecture behind it is a public-private split worth understanding. The Commission runs central services — the passport registry, unique identifier services, a web portal and a semantic repository. The passport data itself stays decentralised with the economic operator or its service provider, reachable through APIs, with a backup service for persistence. The carrier points into that structure; it is not the structure.
What can textile brands learn from the battery passport?
The battery passport is the working prototype for everything textiles will face, and it chose QR. Under the EU Battery Regulation (EU) 2023/1542, from 18 February 2027 every EV battery, light means of transport battery and industrial battery above 2 kWh placed on the EU market must carry a battery passport accessible via a QR code. A battery without a passport cannot legally be placed on the market from that date.
Three details transfer directly. The QR code and unique identifier must follow recognised standards, with the regulation pointing to ISO/IEC 15459 or equivalent. The passport is a structured, machine-readable record covering carbon footprint, recycled content, chemistry, state of health and due diligence. And access is tiered rather than fully public, so competitively sensitive data is exposed only to the roles entitled to see it.
That last point is the one textile brands most often miss. A digital product passport does not mean publishing your bill of materials to the internet. It means role-based access, where a consumer sees care and composition, a recycler sees material detail, and a market surveillance authority sees compliance evidence. EN 18239 covers exactly this — access rights management, information system security and business confidentiality.
The pattern across battery, construction products, toys and detergents is consistent enough to plan against: a 2D code on the product, a registered unique identifier, decentralised data, tiered access. Textiles will differ in what data is required, not in how the plumbing works. You can see what this looks like in practice before the textile rules are final.
What are the security risks of DPP data carriers?
The main carrier-specific risk is that a QR code is trivially replaceable, and the DPP security standard says so directly. The threat model behind EN 18239 explicitly names identity and deception attacks including QR quishing and system faking, data interception through man-in-the-middle and transport layer attacks, and non-trustworthy insider or partner access.
Quishing in a DPP context is straightforward. An attacker covers or replaces the printed code on a garment, or on shelf-edge material, with one pointing at a lookalike domain. The consumer scans, sees a convincing passport page, and hands over data or credentials. The brand's own passport was never touched, which makes it hard to detect and easy to deny.
The defences are unglamorous. Use a resolver on a domain you control and that customers can recognise. Do not print raw redirect-shortener links. Serve the passport over TLS with a certificate that matches the brand. Monitor for lookalike domains. Where product value justifies it, add a chip whose unique identifier can be cryptographically checked, since cloning a chip UID is substantially harder than photocopying a square.
Data authentication is the complementary problem, covered by EN 18246 — data authentication, reliability and integrity. A passport that can be silently altered is worse than no passport, because it launders bad data through an official-looking channel. Signing passport data and keeping an auditable change history is the mitigation, and it also happens to be what makes supplier-declared data defensible under scrutiny.
What data sits behind the carrier for textiles?
Nobody knows the final list yet, and any article claiming otherwise is inventing it. The textile data points will be defined in the textile delegated act, indicatively expected in 2027. What can be reasoned about today is the shape of that list, because ESPR sets out the categories delegated acts draw from and the battery, construction product, toy and detergent passports show how the pattern is applied.
The categories ESPR works with are consistent across product groups: durability and reliability, reusability, upgradability and reparability, presence of substances of concern, recycled content, resource use, carbon and environmental footprint, and expected generation of waste. For textiles those translate into fibre composition, country of processing steps, chemical inputs, care and repair information, recycled content share, and quantified environmental impact.
The footprint element is the one with the longest lead time, because it cannot be answered from a bill of materials alone. Producing a defensible figure means primary supplier data, a recognised methodology and an audit trail, which is a different order of work from filling in a composition field. A worked example of what that produces is in our full lifecycle breakdown of a cotton t-shirt, where a screening under PEFCR v3.1 lands at 2.71 kg CO₂e per garment. Brands that treat environmental impact modelling as a 2027 task will not have credible numbers by 2028.
Why this makes the carrier decision look small
Set against that data programme, the carrier is a rounding error in effort. The identifier scheme, resolver and printed square can be specified in weeks. Collecting verified data from tier 2 and tier 3 suppliers who have never been asked for it, and keeping that data accurate, complete and up to date as ESPR requires, takes years and touches every sourcing relationship a brand has. The methods, the tier-by-tier data map and an eighteen-month collection schedule are set out in our guide to collecting supplier data for digital product passports.
The reason to settle the carrier early is therefore not that it is hard. It is that settling it removes a dependency, unblocks the label supplier conversation, and lets the organisation move on to the part that actually takes the time.
How should a textile brand actually decide?
Start from identifier granularity, then let the carrier follow. Granularity is a business decision with regulatory and commercial consequences; the carrier is an engineering consequence of that decision.
Step one: choose your granularity. Model level is cheapest and sufficient for basic composition and care disclosure. Batch level lets you tie a garment to a specific production run, which matters for supply chain claims and recalls. Item level is required for resale history, authentication and per-item ownership, and it is the only level that supports a genuine second-life proposition.
Step two: pick the baseline carrier. For essentially every textile brand this is a QR code carrying a GS1 Digital Link URI, printed on a durable care or woven label rather than a hangtag. It is the only carrier readable by consumers without special hardware, it is the carrier the battery passport precedent chose, and it doubles as your Sunrise 2027 POS code.
Step three: decide whether you need a chip. Add RAIN RFID if you already run or plan item-level inventory, since the compliance identifier can ride on infrastructure you are funding for operational reasons anyway. Add NFC if consumer tap-to-experience is part of the brand proposition, or if authentication of high-value pieces matters. Skip both if neither applies — a chip added purely for compliance is currently solving a problem the regulation has not posed.
Step four: specify durability with your label supplier. Ask for wash-cycle test data on the exact substrate and print method you intend to use, not on a generic sample. This is a question label suppliers can answer today, and the answers vary enormously.
Step five: put a resolver in front of everything. Never encode a destination URL directly. Encode an identifier that resolves through infrastructure you control, so the destination can change without reprinting a single label.
Common mistakes
The most expensive mistake is printing the passport URL directly onto labels. It converts every future platform decision into a reprint of the entire season's labelling.
The second is putting the carrier on a hangtag. It satisfies the retail moment and nothing after it, and the standard's own list of audiences — repairers, recyclers, market surveillance — starts after the hangtag is in the bin.
The third is treating the carrier as the project. The carrier is perhaps five percent of the work. Collecting verified supplier data, modelling environmental impact to a defensible methodology, and maintaining that data as accurate, complete and up to date is the other ninety-five.
The fourth is waiting for the delegated act. The carrier decision has a long lead time because it touches label suppliers, garment construction and print processes, and those change slowly. Brands that start after the act is published will be specifying labels while their competitors are already collecting data.
What should brands do before the textile delegated act?
Work backwards from a realistic date rather than a rumoured one. The ESPR Working Plan 2025–2030, adopted on 16 April 2025, identifies textiles as a priority group and preparatory work has started. The delegated act for textiles is indicatively expected in 2027, and under ESPR a delegated act cannot apply earlier than 18 months after entry into force.
That arithmetic puts mandatory textile DPP compliance at 2028 at the earliest, with 2029 entirely plausible if the act slips within 2027. Anyone giving you a precise month is guessing. What is not a guess is that the regulation is moving: the registry implementing regulation is in force, six horizontal standards carry presumption of conformity, and the acts on destruction of unsold goods adopted on 9 February 2026 bring disclosure obligations touching textiles and footwear from February 2027.
| Period | What to do |
|---|---|
| Now (2026) | Decide granularity. Specify and durability-test a care-label QR. Stand up a GS1 Digital Link resolver. Start supplier data collection — this is the long pole. |
| 2027 | Read the delegated act on publication. Adjust data model to the mandated data points. Run a pilot collection on one product family, end to end. |
| 2027–2028 | Roll the carrier into standard label specifications across the range. Register identifiers. Test tiered access with real supplier data. |
| 2028–2029 | Compliance at scale. By this point carrier choice should be a solved, boring part of your tech pack. |
The carrier is the easy half. A QR code on a care label is a solved problem that your label supplier can quote this week. Knowing the fibre composition, country of processing, chemical inputs and carbon footprint behind that code — verified, from suppliers who have never been asked before — is the part that takes two years.
Frequently asked questions
Is a QR code enough for a textile digital product passport? For most textile products, yes, provided it is durable, item-appropriate in granularity and backed by a resolver. QR is named in EN 18220:2026, it is the carrier the EU battery passport mandates from 18 February 2027, and it is the only option a consumer can read with no special hardware. A chip becomes worthwhile when you need bulk reading, stronger authentication or a tap-to-experience layer.
Does the ESPR require NFC or RFID for textiles? No. ESPR is technology-neutral, and the specific carrier requirements for textiles will be set in the textile delegated act, indicatively expected in 2027. EN 18220:2026 recognises QR code, Data Matrix, NFC, HF RFID and RAIN RFID, so all five remain on the table until the act narrows them.
Can one QR code serve both the checkout and the digital product passport? Yes, if it carries a GS1 Digital Link URI. Under GS1's Sunrise 2027 initiative, retailers across 48-plus GS1 Member Organisations are committing to accept 2D barcodes at point of sale by the end of 2027, so the same printed code can clear the till and open the passport. This is the strongest argument for choosing GS1 Digital Link over a plain URL.
Where on the garment should the data carrier go? On a permanent care or woven label rather than a hangtag. EN 18220 requires the carrier to serve repairers, recyclers and market surveillance authorities in addition to consumers, and all of those interactions happen long after a hangtag has been cut off and discarded.
How much does a DPP data carrier cost per garment? A QR code printed on a care label you already produce adds effectively nothing. Woven RFID or NFC care labels run roughly $0.05–0.12 per unit, basic heat-seal UHF tags $0.15–0.25, and sewn-in tags rated for 500-plus wash cycles $0.40–0.75. For most brands the carrier is a small fraction of total programme cost compared with supplier data collection and impact modelling.
Will recyclers be able to read my digital product passport? Not reliably today. EU separate collection of textile waste has applied since 1 January 2025, but automated sorting lines such as Fibersort and SIPTex identify fibres by near-infrared spectroscopy rather than by reading carriers, and labels are often missing or destroyed by the time material reaches them. EN 18220 names recyclers as an audience the carrier must serve, so this gap is recognised — it is simply not yet closed.
What happens if I change DPP platforms after printing labels? If you encoded a direct URL, those labels are dead and the garments become unreachable. If you encoded an identifier resolved through a GS1-conformant resolver you control, migration is a configuration change and the printed carriers are unaffected. This single decision is the highest-leverage choice in the whole carrier discussion.
Does a DPP mean publishing confidential supply chain data? No. Access is tiered by role, as it already is under the EU Battery Regulation, and EN 18239:2026 governs access rights management, information system security and business confidentiality. A consumer, a recycler and a market surveillance authority each see a different slice of the passport.
Do I have to register my product identifiers with the Commission? For product groups already covered, yes. Commission Implementing Regulation (EU) 2026/1778 sets out the registry arrangements, the registry went live on 20 July 2026, and economic operators must be verified through eIDAS identity means. Textiles will join once the textile delegated act applies, so identifiers should be designed now to be globally unique and persistent.
Getting the carrier decision right
The data carrier is the most visible part of a digital product passport and the least difficult. The decisions that will actually determine whether a textile brand is compliant in 2028 are about identifier granularity, supplier data, methodology and access control — and every one of them has a longer lead time than choosing between a printed square and a woven chip.
Choose the QR code on a durable care label as your baseline, put a GS1 Digital Link resolver behind it, add RFID only where operations already justify it, and then spend your remaining attention on the data. That sequence is defensible today and adaptable to whatever the textile delegated act says in 2027.
Three steps to take next
1. See a working passport before you specify a carrier. Scan through a live digital product passport and look at what a consumer, a repairer and an auditor each get from the same identifier. It makes the granularity decision concrete in about five minutes, which is faster than any internal workshop on the subject.
2. Line the carrier up against the data you will have to carry. The carrier is only worth what sits behind it, so read our complete guide to DPP compliance for textiles next and check your data model against the requirements before you lock a label specification.
3. Get a free screening of one product. We take one of your bestsellers, work through granularity, carrier specification and supplier data readiness, and tell you where the gaps actually are. Book a free screening with cyrcID and we will build the digital product passport picture for that product with you.
