The DPP data you cannot get later: who holds which fact

The garment you cannot currently describe

Answer four questions about a style you shipped last season, from your own files, without asking anyone. Which mill knitted the fabric, under what registered name? Which dye house put the colour in? What did the fibre split measure on that lot, rather than what the tech pack asked for? Who supplied the waistband elastic? Most brands stop at the second question.

That is not a scandal and nobody hid anything. The facts were never written down anywhere the brand could reach, because nobody was ever asked to write them down. A purchase order records style, quantity, price and delivery date; a packing list records what went in the cartons. Neither describes how the product was made, or by whom, and neither was ever meant to.

A Digital Product Passport changes what a brand is expected to be able to say about a garment: not marketing copy but structured, product-level detail somebody else can check. The exact fields and the timing are unsettled, and we come back to that; the direction is not. So the useful question today is not what the rules will require. It is: if somebody asked now, whose filing cabinet holds the answer?

The custody map: five or six parties, none of them you

Walk the chain and the facts sort themselves by owner. The yarn supplier holds fibre origin and content. The mill holds construction — knit structure, gauge, finished weight, any finishing it applied itself. The dye house holds the colour: recipe, chemical inputs, dye lot, date. Each trim supplier holds its own component: elastic, thread, drawcord, zip, labels. The garment factory holds assembly, and often none of the facts above it, having bought fabric finished and trims to specification.

The brand sits at the end of that line with a commercial invoice, a packing list, a tech pack and a photograph. The first two describe a transaction, the third an intention, the fourth the outside of the garment. Not one describes the production.

This is not one supplier's file you can request. It is five or six independent companies' files, and none of them keeps anything for your benefit. A mill keeps lot data because the mill needs lot data; when it stops needing it, it stops keeping it. Nobody holds the whole picture unless one party is deliberately made its custodian — and the only candidate is whoever placed the orders, because it is the only party that dealt with all of them.

The cost of a fact rises steeply after the goods ship

At the moment you place the order, every fact on the map is nearly free. The mill is quoting, the dye house is booking your colour, someone is buying trims this week. Everyone involved wants the business, and your order is the thing in front of them. Asking then costs an email.

At shipment the price has gone up but the facts are still live, and there is usually a balance outstanding — the only leverage that reliably works. Ask after the final payment clears and you are not invoking a term of the deal; you are asking a favour of people busy with somebody else's season.

A year later the situation differs in kind, not degree. Records age out: facilities keep lot data for their own operational reasons and their own retention periods, not yours, and when the reason expires so does the record. Staff move on. A mill that ran your fabric one season may genuinely not hold the lot detail two seasons later, and is not lying when it says so.

Trims are the worst case. Elastic, drawcords and zips are often bought through a middleman or off a local market, sometimes at short notice to protect a delivery date. If nobody recorded the supplier on the day, that supplier may not be identifiable afterwards at all — not slow to find, not expensive, simply gone. A fact that was free on the day does not exist a year later.

What to record, per order, starting with the next one

Start with identity: the legal name of every facility that touched the product, and what each one did. Not “our supplier in Fujian” but the registered company name, the city and the step — knitted the fabric, dyed it, applied the finish, cut and sewed, supplied the elastic. A legal name is the only version of a facility's identity that survives a new trading name or a new sales contact.

Then the product facts as they happened, rather than as specified. Fibre composition as measured on the production lot, not the nominal figure from the tech pack — the labelling consequences of that gap are covered in a separate guide. The construction and finished weight on that lot. The colour and the dye house that produced it. Trim components with their suppliers. Any finish or treatment and which entity applied it, because the finisher is frequently not the mill. Production dates and lot references, which tie a fact to a run rather than a style.

Attach the record to the order and the style, because the question, when it comes, will be about a product rather than a season. A spreadsheet row per order gets you most of the way. Framed as a compliance project this needs a budget, an owner and a decision, and will not happen. Framed as one more field on the order form, filled in by whoever is already emailing the supplier, it costs nothing.

Put it in the purchase order, or it stays a favour

Whether you get this data comes down to when you ask, and whether the ask is a term or a request. A request made after shipment competes with everything else the supplier has on. A term written into the order was part of the job that was quoted, and gets done for the same reason the labels get sewn in.

So write the clause, as two plain obligations. First, disclosure: the partner facility identifies by legal name the sub-suppliers used on this order — mill, dye house, finisher, trim suppliers — and gives written notice before substituting any of them mid-programme. Second, the record: the lot facts above are supplied as a condition of shipment or of final payment, travelling with the packing list and the inspection report rather than being chased afterwards.

Expect friction on the first obligation. Some facilities treat their mill and dye house relationships as commercially sensitive, and a brand asking for a full sub-supplier list on a first small order may be told no — a negotiation about volume and trust, not a moral failure. The second is different: production-lot facts about your own goods cost a supplier nothing. A facility that will not give you the measured composition or the production dates of garments it is invoicing you for has told you how it keeps records. Put the clause in the first order rather than the fifth; adding it later is a renegotiation.

What is not settled, and what not to buy yet

Be straight about the state of the rules. The Ecodesign for Sustainable Products Regulation puts the framework in place and names textiles a priority category. The detail that would bind you — which data fields, in what format, carried how, and from when — sits in a product-specific act for textiles that has not been adopted. Until it is, any date attached to it is an estimate, and we are not going to give you one.

That has a direct consequence for spending. Do not build an expensive system against a specification that has not been finalised. Software procurement, a consumer-facing passport interface, a decision on carrier technology, a commitment to a data standard — all get cheaper and safer once the requirements exist. Moving early does not buy a head start; it buys a migration.

What is safe now is the cheap thing, and it is cheap because it is not technical. Capture supplier identity and production facts per order. That data is worth holding whatever the final rules ask for — you need it for a recall, a complaint, a tariff query or a retailer's questionnaire. And it is the one part you cannot retrofit. A software decision made late can be corrected in a quarter; a dye house nobody recorded this season cannot be recovered two years from now.

Where a coordinating partner fits, and where it does not

SEAMDANCE has been a buyer-side trading and supply-chain management company in Xiamen since 2018, coordinating independent specialist factories, mills and dye houses. Because we place the work, we are the party that knows which mill knitted a given fabric, which dye house coloured it, which facility applied a finish and where the trims came from — not as a reconstruction afterwards but as a consequence of having arranged it. For an order we run, the custody map can be written down while it is happening and handed over as a record.

The working rhythm helps here. Quotes inside 24 hours, stock samples in three to four days, a custom first sample in around seven, MOQ from 100 pieces per style on stock and 300 to 500 for custom. Decisions about who runs what get made early, on small quantities — exactly when each facility's identity is easiest to fix and cheapest to write down. Final inspection to AQL 2.5 checks workmanship and measurements; it says nothing about provenance, and no inspection standard does.

The limits matter as much. We do not certify anything and hold no certification of our own — certifications belong to the partner facilities that hold them, and holder, scope and validity are verified before an order is placed. We cannot compel a supplier we did not place work with to disclose anything: a factory a brand appointed directly is outside our reach. And we cannot reconstruct production facts that predate the arrangement — for goods made elsewhere, in a season we did not run, the record does not exist and we cannot manufacture one.

This is general information, not legal advice. The rules are still being written: the textile requirements are not adopted, the data fields are not final, and nothing here says a brand or a product is or will be compliant. Confirm your own obligations for each market you sell into, with advisers who answer for it. Whatever the specification finally says, the claim this guide is built on holds: the facts are cheapest on the day the order is placed, and some are not available at any price a year later.