Mapping Digital Product Passport Requirements onto ETIM Fields
In short
Which ETIM fields satisfy Digital Product Passport requirements, and what does a compliant machine-readable product record actually look like?
The Digital Product Passport requires product data to be machine-readable, structured and interoperable, but the regulation deliberately does not name a dictionary — it defers the semantics to standards. For electrical, HVAC, plumbing and building products, ETIM is that dictionary in practice: classes carry EC codes, features carry EF codes, and permitted values carry EV codes, with four data types and no free text on alphanumeric features. ETIM xChange 2.0, released in November 2025, added extended environmental fields for LCA and EPD data plus a packaging element aligned to the PPWR, which covers a substantial part of what the delegated acts are expected to ask for. It does not cover everything: unique identifiers, the data carrier, registry registration and access control sit outside the classification standard and have to be solved alongside it.
The Digital Product Passport regulation is unusually clear about the shape of what it wants and unusually quiet about the content. Product information must be machine-readable, structured and interoperable, resting on internationally recognised open standards. It does not say which standard. That is left to the delegated act for each product group, and most of those acts do not exist yet.
This creates a peculiar situation for a manufacturer of electrical, HVAC, plumbing or building products. You know a requirement is coming. You know roughly what it will demand. You cannot yet know the exact field names. And you have a classification standard sitting in your PIM that already describes most of your product's technical properties unambiguously.
The sensible move is to map what you have onto what you can reasonably predict, build the record now, and leave the regulatory plumbing until the act that governs your products is published. This piece is that mapping — including, importantly, the parts where the mapping fails.
What ETIM actually gives you
ETIM is a dictionary, not a file format. Its job is to make a technical property mean the same thing to a German manufacturer, a Dutch wholesaler and a Polish contractor without any of them agreeing a private schema first.
Three code types carry that meaning:
| Prefix | What it identifies | Example |
|---|---|---|
EC | Product class | EC000001 — miniature circuit breaker |
EF | Feature (a technical property) | EF000007 — colour |
EV | Permitted value for an alphanumeric feature | the entries in that feature's value list |
ETIM 8.0 contains 18,766 features across 5,145 classes. Each feature has one of four data types: numeric, numeric range, alphanumeric, or logical.
The constraint that matters for compliance is easy to miss. Alphanumeric features do not accept free text. Each is bound to a list of permitted values, and each of those values has its own EV code. You cannot write "approximately IP66" into an ingress-protection field. You select a value or you do not populate it.
That single restriction is the difference between data a machine can read and data a machine can evaluate. A regulator checking conformance, a distributor's procurement system filtering on a threshold, and an AI assistant answering "which of these is rated for outdoor use" all interrogate that field identically. This is what "interoperable" means once you stop treating it as a synonym for "shared".
Where the DPP requirements land on that structure
The requirements fall into four groups, and only two of them are ETIM's problem.
Identification — mostly not ETIM
The passport needs a unique product identifier, an issuing scheme for it, and an operator identifier. ETIM classifies what a product is; it does not issue identity. This is GS1 territory, or an equivalent scheme, and it is the first thing to settle because everything else keys off it.
Technical characteristics — squarely ETIM
This is the part ETIM was built for and where the mapping is close to one-to-one. Dimensions, materials, performance ratings, electrical characteristics, ingress protection, temperature ranges: these already exist as EF features on your class, with defined units and permitted values. If your products are classified, most of this is done and you simply have not been thinking of it as compliance data.
Environmental and circularity — ETIM xChange 2.0
Here is the development most people have not registered. ETIM xChange 2.0, released in November 2025, added extended environmental fields for life-cycle assessment and Environmental Product Declaration data, developed with ETIM International's environmental data working group. It also added a packaging material element aimed at the Packaging and Packaging Waste Regulation.
Whether those fields align one-to-one with a given delegated act is unknowable until the act is published. But the structure exists, it is versioned, and manufacturers are already populating it. That is a materially better position than an empty column headed "carbon footprint".
Carrier, registry and access — not ETIM at all
The data carrier on the physical product, registration in the EU registry that went live in July 2026, and differentiated access control — some fields public, some visible only to regulators or recyclers — are all outside the classification standard. ETIM tells you what a property means. It has nothing to say about who may read it.
What a compliant-shaped record looks like
ETIM xChange is JSON with a published JSON Schema, and it separates generic product data from the trade items that carry it. The shape below is that separation, with a miniature circuit breaker as the subject.
One deliberate omission needs explaining. The ETIM model is licensed ODC-BY, and this article does not reproduce it. EC000001 and EF000007 appear because ETIM International and its national organisations publish them as illustrative examples in their own documentation. Every other code below is shown as a placeholder. Pull the real ones from your ETIM release — that is the licensed, attributed, always-current source, and republishing a dictionary you do not own is a poor foundation for a compliance record.
{
"product": {
"manufacturerProductNumber": "MCB-C16-1P",
"manufacturerGLN": "<GS1 Global Location Number>",
"productIdentifier": {
"scheme": "GS1",
"value": "<GTIN>"
},
"classification": {
"system": "ETIM",
"release": "ETIM-8.0",
"class": "EC000001",
"features": [
{ "code": "EF000007", "type": "A", "value": "<EV code for colour>" },
{ "code": "EF<rated-current>", "type": "N", "value": 16, "unit": "A" },
{ "code": "EF<poles>", "type": "N", "value": 1 },
{ "code": "EF<tripping-characteristic>", "type": "A", "value": "<EV code for C>" },
{ "code": "EF<ip-rating>", "type": "A", "value": "<EV code>" }
]
},
"environmental": {
"epd": {
"declarationNumber": "<EPD registration>",
"validUntil": "2029-11-30",
"gwpTotal": { "value": 2.41, "unit": "kg CO2e" }
},
"recycledContent": { "value": 18, "unit": "%" }
},
"packaging": {
"materials": [
{ "material": "corrugated board", "weight": { "value": 42, "unit": "g" }, "recycledContent": { "value": 82, "unit": "%" } }
]
}
},
"tradeItems": [
{
"gtin": "<GTIN>",
"packagingQuantity": 1,
"orderUnit": "PCE"
}
]
}Three things about this record are worth drawing out.
The classification block carries its release. "release": "ETIM-8.0" is not decoration. Features move between releases, and a record that does not state which dictionary it was written against becomes ambiguous the moment the next version ships. This is the single most common defect in real-world ETIM data.
Every feature carries its type. A consumer can validate "type": "N" against a numeric and "type": "A" against a permitted-value list without knowing anything about circuit breakers. That is what makes the file checkable rather than merely parseable.
Generic data and trade items are separate. One product, many pack sizes, one set of technical truths. Flattening these — which is what most spreadsheet-based exchanges do — is how a 10-pack ends up claiming the weight of a single unit.
The gaps, stated plainly
A mapping article that claims complete coverage is selling something. Four honest gaps:
- 01No delegated act exists for most product groups. The only firm date is 18 February 2027 for certain industrial, EV and light-transport batteries, under the Batteries Regulation rather than ESPR. Anyone quoting you a single DPP deadline is oversimplifying.
- 02ETIM's environmental fields are new. Released November 2025, they have not been tested against a published delegated act, because there is not one to test against.
- 03Access control is unsolved in the format. The regulation contemplates tiered visibility. ETIM xChange does not express it. That will need to live in whatever serves the passport.
- 04Classification coverage is uneven. 5,145 classes is a lot, but if your product sits in a sparsely-featured class, the dictionary will not carry your differentiators and you will be back to attachments and free text.
Why this record is worth building anyway
Strip the regulation away entirely and the same file is still the most valuable thing you could build for your catalogue.
A structured, typed, dictionary-backed product record is what lets a distributor ingest your range without a mapping project, what lets a procurement system filter on a threshold rather than a keyword, and what lets an AI assistant answer a specification question about your product and cite you rather than guess. We measured 984 industrial catalogues on that last point and the median score was 50 out of 100 — the constraint is almost never the product, it is that the data describing it is trapped in PDFs and JavaScript.
The compliance deadline is a forcing function for work that pays for itself. That is an unusually comfortable position, and it is worth recognising it as one.
Where to start
In this order, because each step makes the next cheaper:
- 01Classify against your current ETIM release, and record which release. If you are already classified, verify the release is stated in your exports — often it is not.
- 02Export to ETIM xChange 2.0 and validate against the published JSON Schema. Validation failures here are data-quality problems you already had and had not seen.
- 03Populate the environmental block with whatever EPD and LCA data exists today, even if partial. Partial structured data beats complete unstructured data.
- 04Settle identity — GTIN or equivalent, and the issuing scheme — because the passport keys off it.
- 05Leave carrier and registry until your delegated act is published. Building against a draft is how you end up doing it twice.
The first three steps produce an asset that is useful the day it exists, regardless of what the delegated acts eventually say. That is the test to apply to any DPP work someone proposes to you: if it is worthless without the regulation, it is probably premature.
Common questions
Does the Digital Product Passport regulation require ETIM specifically?
No, and this is the most common misreading. The ESPR framework requires data to be machine-readable, structured and interoperable, and says the semantic level should rest on internationally recognised open standards — but it does not name ETIM, ECLASS, GS1 or any other dictionary. That choice is left to the delegated act for each product group and, in practice, to the sector's existing habits. For electrical, HVAC, plumbing and building products in Europe, that habit is overwhelmingly ETIM, which is why mapping to ETIM is a reasonable bet even before the acts land. It is a bet, not a certainty.
What is the difference between ETIM and ETIM xChange?
ETIM is the dictionary: the classes, features and permitted values that give a technical property an unambiguous meaning across languages and suppliers. ETIM xChange is the transport: a JSON format, with a JSON Schema for validation, for moving product records between manufacturers, wholesalers and distributors. You need both. A perfectly classified product that only exists inside a supplier portal is not interoperable, and a beautifully structured file whose attributes are free text is not machine-readable in any useful sense.
Which ETIM data types matter for compliance?
ETIM defines four: numeric, numeric range, alphanumeric and logical. The important property for compliance is that alphanumeric features do not accept free text — each is bound to a list of permitted values with their own EV codes. That constraint is precisely what makes the data checkable by a machine rather than merely readable by one. A regulator, a customer's procurement system and an AI assistant can all evaluate the same field the same way, which is what interoperable means in practice.
Does ETIM xChange 2.0 already carry the environmental data the DPP will ask for?
Partly, and that is a meaningful head start. Version 2.0, released in November 2025, introduced extended environmental fields for life-cycle assessment and Environmental Product Declaration data, developed with ETIM International's environmental data working group, plus a packaging material element aimed at the Packaging and Packaging Waste Regulation. Whether those fields map one-to-one onto a given delegated act cannot be known until that act is published. What can be said is that the structure exists, is versioned, and is already being populated by manufacturers.
What does the DPP require that ETIM cannot give me?
Four things. A unique product identifier and its issuing scheme, which is a GS1 or equivalent question. The data carrier — the QR code or similar physically on the product — which is a labelling and print question. Registration of the passport in the EU registry, which went live in July 2026 and is a directory rather than a data host. And differentiated access control, because the regulation contemplates that some fields are public while others are visible only to regulators or recyclers. ETIM tells you what a property means. It does not tell you who is allowed to read it.
Should we wait for the delegated act covering our products before doing this work?
The work that pays off regardless is the classification and the structured record, because that asset is already valuable without any regulation: it is what makes a catalogue readable to procurement systems, marketplaces and AI assistants today. The work worth deferring is anything that hard-codes assumptions about a specific act — field names lifted from a draft, a registry integration built against a preview API, a data carrier printed onto packaging before the content set is final. Build the record now, wire the compliance plumbing when the act exists.
Sources
- 01ETIM International — ETIM xChange version 2.0 officially released
- 02ETIM International — Model information (class, feature and value structure)
- 03ETIM UK — Structure of the ETIM data model
- 04Regulation (EU) 2024/1781 — Ecodesign for Sustainable Products Regulation
- 05Regulation (EU) 2023/1542 — Batteries Regulation, battery passport from 18 February 2027
- 06ETIM International — classification licence terms (ODC-BY 1.0)
See exactly what AI assistants can and cannot read of your products today — crawler policy, catalog coverage, datasheet access — scored and benchmarked against 984 distributors and manufacturers worldwide.
Grade my catalogRelated field notes
Digital Product Passports and Machine-Readable Data
What the EU Digital Product Passport actually mandates, the confirmed dates, how ETIM, ECLASS, AAS and GS1 fit…
TechnicalAI Crawlers Do Not Run JavaScript
GPTBot, ClaudeBot and PerplexityBot read raw HTML and never execute scripts. How to test whether your catalog …
MarketThe Agentic Web and Industrial Product Data
The web is splitting into a human surface and a machine surface. Crawl-vs-refer economics, Cloudflare's toll g…