No textile delegated act exists, and nobody knows the final list of fields a passport will carry. That is the reason most often given for doing nothing.
It does not hold up, for two reasons. The preparatory study released in May 2026 already describes the expected structure, precisely enough to work from. More to the point, most of the data it lists is already required elsewhere, on the label sewn into the garment. Gathering it is not anticipating a future constraint. It is tidying what you already owe.
This article takes that data field by field, and says at which level each one lives.
What the preparatory study is, and what it is not
The document is titled Study on DPP content for textile apparel products under ESPR. It is an external study prepared for the Commission's Joint Research Centre, released in May 2026. It still carries placeholder publication numbers and states itself that it does not necessarily reflect the Commission's position.
So it is not a text, and that matters for reading what circulates about it. The lists of mandatory data points and the season dates found on some vendor sites are extrapolations. The study is nonetheless the document the delegated act will be built on.
Its proposed scope rests on a threshold: products containing at least 80 percent textile fibres by weight, across ten apparel categories, workwear and sportswear included.
Its exclusions are more instructive than its scope. Out go smart and electronic textiles, personal protective equipment, medical devices and toys. And above all, every intermediate product, meaning fibres, yarns and fabrics sold as such. The obligation lands on the finished product only.
In other words, the passport is not an upstream traceability tool. That is the opposite of what the platforms selling supplier data collection promise.
The three levels, and the one the study settles on
The study proposes three granularities and picks between them.
Model level carries the bulk: fibre composition, weight, environmental footprint, a recyclability score, care and repair instructions, warranty. These values are identical for every item of one reference.
Batch level carries what changes from one run to the next: operator and facility identifiers, substances of concern, recycled content amounts.
Item level carries an individual identifier, which the study puts explicitly in the voluntary column, with an inheritance mechanism from the batch data.
The recommendation itself is clear: require the batch identifier as the minimum, with each data point then populated at batch or model level according to its nature.
For a store, that means something simple. The passport is identified by production run, not by the item sold, and its content comes almost entirely from the model. You have to serialise nothing, and keep no per-customer register.
It also means that changing manufacturer partway through a model's life creates a new run, so a new passport, rather than an edited field on the old one. Garments already sold keep pointing at the data that was true when they were made.
One more line from the study deserves quoting, because it cuts against what is written elsewhere. It notes that more than 99 percent of the companies in the EU textile supply chain lack the digital infrastructure these requirements assume, and recommends aligning them with established commercial practice, observing that model-level data already matches what is published today, starting with fibre composition.
Composition, the one field already governed by law
This is the passport's central data point, and the only one you already publish in a form European law prescribes. Regulation (EU) 1007/2011 on textile fibre names sets precise rules, worth knowing before filling a structured field.
- Every fibre, as a percentage by weight, in descending order (article 9(1)), using the harmonised names in annex I.
- Minor fibres may be grouped as "other fibres" followed by their total percentage, up to 5 percent for a single fibre and 15 percent for several (article 9(2)). This is not a drafting convenience: the text reserves it for fibres that cannot easily be stated at the time of manufacture.
- "100%", "pure" and "all" are reserved for products made of a single fibre (article 7). Technically unavoidable extraneous fibres are tolerated up to 2 percent, and up to 5 percent for carded products.
- Visible decorative fibres up to 7 percent and antistatic fibres up to 2 percent are left out of the stated composition (article 10).
- A manufacturing tolerance of 3 percent is allowed between the declared composition and the one obtained by analysis (article 20(3)).
- The label is written in the language of the member state where the product is made available to the consumer, unless that state provides otherwise (article 16(3)).
Two practical consequences for a passport. First, the value already on your label is the right one: nothing to recompute, only to lift out of free text. Second, "other fibres" reads badly in a structured field. Where you know the detail, write it, even below the thresholds that would let you group it.
Worth watching: the Commission has a review of this regulation under way, and digital labelling is one of its topics. Nothing is settled, but that is the text to follow.
Care instructions, mandatory nowhere and expected anyway
A point that surprises people: care labelling is required by no European text. The 2011 regulation covers composition only. Its article 24 called for a report on possible harmonisation, and the conclusion was that it was not needed, the voluntary EN ISO 3758 standard already covering the ground.
The symbols themselves belong to GINETEX, which holds them as trademarks and brought them into the ISO standard. They are not public-domain pictograms, which is worth knowing before redrawing them for a web page.
The study nevertheless places care and repair instructions at model level, among the expected data. Preparing for that costs nothing: write the instructions out in words. "Wash at 30 degrees, do not tumble dry" is text, it translates, it reads on a phone, and it raises no question of usage rights.
Some markets do impose care labelling in national law. If you sell outside your own country, that is worth checking market by market.
Origin, two questions that get confused
There is no European obligation to mark the origin of a textile. The 2005 proposal on mandatory origin marking was withdrawn, and the 2013 attempt did not survive the Council.
That said, once you state an origin it has to be accurate, and accurate has a precise customs meaning. Non-preferential origin is defined by article 60 of the Union Customs Code: the country where the product was wholly obtained, or the one where it underwent its last substantial, economically justified processing. It is not the country of the brand's head office, nor necessarily the country of making up.
The French AGEC law asks a different question. Its article 13 requires not a label but a digital disclosure, and it asks for three countries: dyeing or printing, weaving or knitting, and making up, each being where that stage is principally carried out.
So a passport reading "made in Portugal" does not answer AGEC, and three stage countries do not mechanically give a customs origin. They are two distinct pieces of information, and it is worth planning both fields.
That French obligation only applies above cumulative thresholds, lowered each year and standing since 1 January 2025 at 10 million euros of turnover and 10,000 units placed on the French market. Below that it is not required, which is very widely misunderstood.
What not to prepare now
Item-level serialisation. Voluntary in the study, and it assumes stock handling that has nothing to do with yours.
Buying identifiers. The regulation names no vendor. The study mentions standardised identifiers among the options, and nothing is settled.
Scores. The environmental footprint points at a sector calculation method, and the recyclability score has no published reference framework. As for the robustness score, the study gives it as possibly voluntary and even suggests de-prioritising it, the benefits not matching the burden. Nobody can sell you a score backed by a text that does not exist.
A realistic order of work
A catalogue is not handled product by product, or the work scales with the number of references and never ends. It is handled in layers, widest first.
- The store. What is true of nearly everything you sell.
- The category. Composition per product family. This layer does most of the work, since one fabric serves dozens of references.
- The product. Exceptions only, plus values that are unique by nature, such as the batch number.
That is exactly the order of the DPTag defaults engine: the product attribute wins, otherwise the category value, otherwise the store value. The wizard walks those three layers, then creates the catalogue's passports in batches.
For the starting point, see the passport seen from WooCommerce. For the dates, see what is already in force and what is not.