# What to do now, and what to refuse to do now Source: https://activatedigital.ai/knowledge/guides/what-to-do-now Last verified: 28 August 2026 Summary: There is no date by which a textile Digital Product Passport must exist, no adopted act that says what one would carry, and no draft act published for consultation. The Commission publishes a planned adoption for the act and marks the timing as indicative and liable to change. So preparation cannot be sequenced against a deadline, because there is not one. Sequence it by whether a decision can be undone instead. Most of this work is reversible and cheap. Four decisions cannot be undone once goods are made, three of the four get taken by people who do not know they are taking one, and one kind of work stops being possible the moment a season moves on. ## In 60 seconds There is no date by which a textile Digital Product Passport must exist, no adopted act that says what one would carry and no draft act published for consultation. So preparation cannot be sequenced against a deadline. - The ESPR Article 9 duty is conditional on a delegated act existing. For textiles none has been adopted and none published in draft. - Requirements apply no earlier than eighteen months from a delegated act entering into force, except in duly justified cases. The Commission's textile timing is a planned adoption, labelled indicative. - Once passports exist, Article 10 requires the operator to give dealers and online marketplaces a digital copy of the carrier or identifier free of charge within five working days of a request. - The destruction ban under Article 25(1) of Regulation (EU) 2024/1781 has applied since **19 July 2026**. Medium-sized enterprises enter it on 19 July 2030 and micro or small businesses are not reached. - Nothing in the passport provisions carries a size test of any kind. Sequence it by whether a decision can be undone instead. Most of this work is reversible and cheap. Four decisions cannot be undone once goods are made, three of the four get taken by people who do not know they are taking one, and one kind of work stops being possible the moment a season moves on. ## What the framework actually establishes, and what it leaves open The [Ecodesign for Sustainable Products Regulation](https://activatedigital.ai/knowledge/digital-product-passport) is in force. Article 9 establishes the passport and states that information requirements are to provide that products can be placed on the market only if a passport is available in accordance with the applicable delegated acts, and that the data in it is to be accurate, complete and up to date. Article 10 sets the essential requirements for the passport itself. Article 11 governs its technical design and operation. The Article 9 duty is conditional on a [delegated act](https://activatedigital.ai/knowledge/regulation/what-a-delegated-act-decides) existing for the product group. For textiles, none has been adopted, none has been published in draft, and the preparatory work behind one has a milestone whose timing the Commission records as still to be communicated. The delegated act decides the content, the carrier, the level, who may write to the record and who may read it. All of that arrives in one document, at one time. Anything a business commits to before that document exists is committed against somebody's guess at it. There is a route for [taking part before the act is adopted](https://activatedigital.ai/knowledge/guides/influencing-a-delegated-act). The framework separately sets a floor: requirements apply no earlier than eighteen months from a delegated act entering into force, except in duly justified cases. The Commission does publish a timing for the textile act. It appears under a heading calling the timeline indicative, it is described as a planned adoption rather than a date of application, and it carries a note that implementation timelines may evolve as legislative and technical work progresses. A planned adoption plus a floor is arithmetic, not a deadline, performed on a number its own author says may move. This estate publishes the working and not the result. What is adopted, what is in force, what is planned and what has only been signalled is maintained on the status tracker, with the Commission's own framing kept intact. None of that argues for doing nothing. It argues against sequencing by a date: a plan built on a date that moves has to be rebuilt when it moves, and a plan built on reversibility does not. ## The four decisions you cannot undo Each is made once and early, and can't be walked back once goods carrying it are in circulation. ### What is printed on the goods A carrier printed onto a swing tag or a sewn-in label is a physical object with a print run behind it and a garment in front of it. The garment is in somebody's wardrobe years later. Ink, geometry and the number inside are fixed at the moment the label is made, and nothing about the printed square can be edited afterwards. The part of that decision that matters least is which symbology is fashionable this year. The part that matters most is what address the code carries, because the address has to keep answering for as long as the goods exist. ### Whose domain is in that address, and whose account holds the identifier Both are ownership questions dressed as configuration, so they get decided by whoever sets up the system rather than whoever will live with the consequence. If the domain printed into the code belongs to a supplier, the supplier controls whether anything answers a scan of your garment. If the identifier licence sits on a provider's account rather than yours, the licence is theirs, the annual renewal decision is theirs, and the identity printed on your goods depends on a payment you do not control. Every persistence commitment this estate has read carries a condition, and the condition is usually a renewal. Neither is revisitable after the goods ship, which puts both in a category with your company name rather than with your software. ### Committing to a level by printing one The framework names three levels a delegated act may choose between, and none has been chosen for textiles. Printing a serialised carrier commits you to the most expensive of the three before anybody has decided it is required, and the commitment is physical rather than contractual. The counts involved are not a modelling preference either. Under the numbering scheme's own management standard, each style, colour and size variation is a distinct identifier, so one shirt in five sizes and six colours is thirty identifiers whatever anybody thinks about passports. Which of the three levels a passport eventually sits at, and why the gap between them runs to three orders of magnitude on an ordinary range, is set out on [how many passports a range needs](https://activatedigital.ai/knowledge/guides/granularity). ### Putting the wrong kind of value into an identifier field An internal stock code written into a barcode field is the most common identity defect in this subject. It's free to fix today and not fixable at all once it has travelled into artwork, into a channel feed and onto a garment, because a corrected identifier passes the arithmetic and may now name somebody else's product. What allocating an identifier commits you to, and why two identical catalogues can face different bills depending on where the business is incorporated, is on [identifiers](https://activatedigital.ai/knowledge/fields/identifiers). ## The work that expires, which is the only genuinely urgent part Everything above is about not deciding too early. This is the opposite, and it is why waiting is not free. Some facts about a product exist only at the moment it is made and cannot be recovered afterwards from anything anybody kept. If they are not captured then, they are gone, and no software bought later reconstructs them. **Which production run a variant came from.** Reading upward is easy: from a batch you can always name the model. Reading downward is the direction a claim needs, and from a model record, or from a garment in a customer's hand, you cannot say which run it came out of unless something was captured at the time. A certificate covering one dyehouse, a recycled input used in one run, a finish that met a restriction on four colourways out of six: each is a fact that is true of part of a style and has no home unless the join was recorded when it existed. **What a document actually covered.** A certificate arrives, is filed, and the scope printed on its face is never written down beside it. A year later that scope is a matter of somebody's memory. Writing the coverage onto the row when the document arrives takes a minute and cannot be done later without reading every document again. **What is attached to the fabric.** Trims, coatings, laminates and the elastane fraction are known at the point of design and forgotten within a season. Nobody currently asks a brand to record them, and they are the material every downstream process turns on. **What you asked, of whom, on what date.** Without a request log an absence cannot be typed at all. There is no way to tell a fact nobody has looked for from a fact somebody looked for and did not find, and those two license completely different sentences. It is the cheapest record to start keeping and the only one whose past cannot be recovered. That is the honest cost of waiting, and it is not a licence fee. Most of what you defer here you can do later at broadly what it costs today. These four get more expensive every season they are left, and some of what they capture stops being recoverable altogether. Their value does not depend on which fields a delegated act eventually names. For the dates that do exist, see [every dated rule, in one table](https://activatedigital.ai/knowledge/data). ## The obligation that arrives with the passport and nobody prepares for One provision creates a workflow rather than a data set, and workflows take longer to build than fields do. Article 10 requires the [economic operator](https://activatedigital.ai/knowledge/regulation/who-carries-the-obligation) placing a product on the market to give dealers and providers of online marketplaces a digital copy of the [data carrier](https://activatedigital.ai/knowledge/passport/choosing-a-carrier-that-still-works) or the [unique product identifier](https://activatedigital.ai/knowledge/fields/the-three-identifiers), so that they can make it accessible to customers who cannot physically handle the product. It requires that copy, or a webpage link, to be provided free of charge and in any event within five working days of a request. That is a service level, not a field. It applies once passports exist for the goods concerned, and it lands on whoever answers the phone to a wholesale customer or a marketplace. A business selling through its own store will barely notice it. A business selling through fifty stockists will need a route from an incoming request to a correct answer in five working days, and that route is a process, an owner and a record rather than a metafield. Nothing requires it yet. It's the one future obligation whose shape a business can usefully know now, and building a request-and-answer process is work that survives whatever the delegated act says. ## The afternoon's work that survives every outcome Five things. Four cost an afternoon each and none requires buying anything. **Find out which markets you place goods on, and record it once for the business.** One answer settles the applicability question underneath several duties at a stroke. Shipping settings are not that answer: a shop that will post to Belgium has not thereby placed goods on the Belgian market. **Find out which role you occupy in each of them.** Manufacturer, importer, distributor and dealer are defined terms carrying different duties, one business can occupy different roles in different markets, and the role decides the duty and does not change when an act arrives. **Look at what your own product data actually describes.** It's the largest of the five and changes every later estimate, because most of the difficulty is not collecting new values but establishing what the values you already hold are about. Where they live, and the four questions that decide whether one is usable, is worked through on [the product data you already have](https://activatedigital.ai/knowledge/guides/the-data-you-already-have). **Structure the attributes that are already governed.** Fibre composition is regulated now, by an instrument in force, independently of any passport. It is also one of the harder values to reconstruct later and one of the easier ones to fix today. **Write down what you cannot answer, and why.** A recorded absence with a reason is a usable artefact that somebody else can pick up. A blank cell is a question nobody knows has been asked. A sixth item costs more than an afternoon and still belongs here, because it does not get cheaper by being deferred. Asking your suppliers once, properly, is a different exercise from sending a questionnaire, it produces a record of having asked even when nobody answers, and the one statutory route compels rather less than most people assume. It is set out on [getting evidence out of suppliers](https://activatedigital.ai/knowledge/evidence/supplier-evidence). If you want to know how long that afternoon actually is, you can [see which of them your current data already answers](https://activatedigital.ai/dpp-readiness-checker) before you start. ## What to refuse to do yet The shorter list, and the one a vendor will not write for you. **Do not print anything that assumes an answer.** A serialised carrier commits you to a level nobody has chosen. A code carrying somebody else's domain commits you to their continued existence. Printed decisions do not get revised. **Do not buy a platform against a requirement that has not been written.** The fields, the carrier, the level and who may write to the record all arrive together in one document. A system selected before that document exists is selected against a guess. Data structured for its own sake survives whatever the act says; a system chosen to satisfy a guess may not. **Do not commission a full life cycle assessment because a passport might want a number.** Nothing requires an environmental figure on a textile product today, and the proposal that exists permits secondary data and default values rather than requiring measurement. The composition and the weight underneath any such figure are worth fixing first, because every modelled figure is built from one and divides by the other. **Do not buy item-level identity as compliance.** It may be a good idea for anti-counterfeit work, warranty or a service you want to offer, and those are real arguments on their own merits. Today item-level identity is not a requirement, and a proposal presenting it as one describes a decision nobody has taken. **Do not treat a marketplace as the deadline.** Nothing read in any platform or marketplace rule requires passport data as a condition of listing, and the largest supplier manual this estate has read describes a passport as expected rather than obligatory across three versions of the document. ## Two duties that do already reach a clothing business Neither is a passport duty and both get lost in conversations about one. A prohibition on destroying unsold apparel, clothing accessories and footwear is already in application, made under the same framework. It is graded by the size of the business, which is the one place in this architecture where size changes what is owed. If you are a micro or small business it does not reach you at all. Medium-sized businesses are not reached until 2030, and large ones are bound now. That grading is the most plausible origin of a belief this estate meets constantly: that small businesses will be carved out of the passport. Nothing in the passport provisions carries a size test of any kind. The instrument is Article 25(1) of Regulation (EU) 2024/1781, its scope is Annex VII and it has applied since [19 July 2026](https://activatedigital.ai/knowledge/data). Medium-sized enterprises enter it on 19 July 2030. For the prohibition and its derogations in full, use [EU ban on destroying unsold textiles](https://activatedigital.ai/knowledge/regulation/destroying-unsold-stock). A second instrument reaches any business making environmental claims to consumers in the Union, with no size relief at all, and its date is close. For most brands the work it creates is a copy and packaging change rather than an evidence project, and it is considerably cheaper to do before it starts than after. Both dates, and which instruments are in force, adopted or merely indicated, are kept on [what already applies today](https://activatedigital.ai/knowledge/regulation/what-applies-today). This page carries no date of its own, deliberately, because a page about sequencing is the easiest place in an estate for a date to be read as a deadline. ## Four questions any quotation should be able to answer Each has a checkable answer rather than a persuasive one, and each works against anything this business puts in front of you as well as anybody else's. **Which level does the price assume, and which document says so.** The unit being counted is styles, sellable variants or garments, and those differ by three orders of magnitude on the same range. A quotation that cannot say which it counted is not a quotation. **Which of these charges recur, and which stop being optional once goods carrying a code are in circulation.** The published costings this estate has looked at price a first year. Very little of the real work has the shape of a first year. **Whose account holds the identifier licence, and whose domain is printed on the garment.** Both answers should be yours. If either is not, the reason should be written down before anything is printed. **Which half of this quotation is data work.** Tidying, deduplicating and attributing product data is not a new market and has had a published price for years. A reader being quoted for a passport programme is partly being quoted for that, and asking which half is which is fair and answerable. ## What would change this page A draft or adopted textile act, at which point the reversibility analysis is replaced by a requirement and this page is rewritten around it rather than defended. A published requirement from a marketplace or a large buyer making passport data a condition of supply, which would be the first genuine commercial deadline in this subject and which this estate has looked for and not found. A decision on the level at which a textile passport sits, which is the single input that most changes what is safe to print. ## Related - [How a DPP Requirement Moves from EU Planning to an Applicable Product Rule](https://activatedigital.ai/knowledge/regulation/how-dpp-requirements-take-effect) — How a planned DPP rule becomes applicable - [Where Do I Start With a Digital Product Passport?](https://activatedigital.ai/knowledge/start) — Where to start, and the six routes through Knowledge ## Keep exploring The question this one usually raises next. - [Another angleRules & scopeEvery date this estate holdsWhat is in application, what is adopted and what is merely indicated, with sources and check dates.→](https://activatedigital.ai/knowledge/status/textile-passport) Also worth reading - [CompareImplementationWhat this work costs, and who can tell youThe published estimates, who published them, and the half of the work nobody has priced.→](https://activatedigital.ai/knowledge/guides/cost) - [Another anglePassport technologyWhen the link diesThe questions to settle about domains, licences and persistence before anything is printed.→](https://activatedigital.ai/knowledge/passport/when-the-link-dies) - [CompareSmall businessDigital Product Passport Readiness for Small Businesses: What to Check FirstRunning a smaller product business with a small team or simple systems?→](https://activatedigital.ai/knowledge/guides/small-business-dpp-readiness)