How to Prepare Product Data for a Digital Product Passport

The identifier and the carrier are decisions made once. The data is a continuing operation. The EU’s central DPP registry has been live since 20 July 2026 (European Commission), so the infrastructure question is largely answered. The harder question is whether a brand can fill a passport accurately.

Five drivers recur:

• Data sits with teams that rarely work together: legal, production, sourcing, product development, sustainability.
• Suppliers hold the hardest fields (material origin, recycled content, substances of concern) and answer by email, in different formats.
• Data exists at the wrong level, for example a model-level spreadsheet used to describe batch-specific materials.
• Nobody is accountable for a field once it is published.
• Claims live in marketing copy, not in auditable records.

Timing adds pressure. Textiles requirements are indicatively expected from 2027, with obligations applying at least 18 months after each delegated act. The battery passport applies from 18 February 2027. Supplier data takes the longest to collect, so the work has to start before the final rules for a product group are published.

The distinction that matters: a passport is not a document a brand writes. It is a dataset a brand maintains, for as long as the product is in use.

It is the structured information that the Ecodesign for Sustainable Products Regulation, Regulation (EU) 2024/1781 (ESPR), allows delegated acts to require for a product group. Each act selects its own subset and sets access rights.

The categories give a planning framework. They are indicative, and the delegated act for a product defines the actual data points.

Two things stand out. Most categories belong to teams with no shared data model. And almost nothing varies by item.

1. Map required fields to owners. For each field in the delegated act, name the team and the system that holds it today.
2. Define the level. Model, batch, material lot or item. Collect at that level only.
3. Close supplier gaps. Request missing data with a standard template and a deadline tied to purchase orders.
4. Validate. Check completeness and plausibility before anything is published.
5. Publish with access rules. Serve each audience its view from the same identifier.
6. Maintain. Version on change, correct by batch, review when acts are updated.

Before step one, run a short readiness audit, product line by product line. Which ESPR product group does each product fall into? Which team holds each likely field? Is the value supplier-declared, certified or estimated? Is it current, versioned and traceable to a source?

Most brands find the same split. Product, operator, compliance and repair data mostly exist internally. Material origin, recycled content and substances of concern depend on suppliers and are incomplete. That split tells you where to put effort first.

At the level where it actually varies. An item-level passport does not require item-level data entry.

A record for a leather bag shows why:

Only the identifier is unique to the item. Everything else is inherited from the model, the batch or the material lot, and the item adds its own identity and history. Identifiers typically combine a GTIN with a serial number, expressed as a GS1 Digital Link.

Projects that copy model data into every item record create a maintenance problem with no benefit. One correction then means millions of edits instead of one.

Different audiences see different views of the same identifier. Delegated acts specify access rights for each product group. Brands cannot make restricted data public, and they should not make public data harder to reach.

Access control is not a hosting detail. It decides whether supplier identities and facility details stay restricted. Build it into the data model from the start. Adding it to a published dataset later is far harder.

Four data types cause most of the delay. Each has a fix that works at scale.

Material origin. Tier-two suppliers do not share it. Ask for batch-level supplier declarations and attach them to every item in the batch. Collect once per batch, not per item.

Recycled content. Claims exist in marketing, not in auditable data. Link supplier certificates to the material lot used in each batch. Claims then become traceable evidence.

Substances of concern. The data is spread across compliance files and supplier emails. Centralise it per model, map it to the passport field and version it on change. The result is one maintained source instead of a yearly scramble.

Repair information. It sits in after-sales, not in product data systems. Link the after-sales documentation to the model record, and both the passport and service teams reuse it.

Where data is thin, resist the temptation to estimate silently. Mark each value as declared, certified or estimated. A passport that shows its own evidence level is easier to defend than one that implies certainty.

A passport is not published once. Products change, suppliers change and errors surface.

Versioning is the control that matters. If a market surveillance authority asks what a passport said on a given date, the brand has to be able to show it.

Sometimes it is. A brand with a few dozen references, one factory and stable suppliers can hold passport data in a well-governed spreadsheet or its existing PIM. The discipline matters more than the tool: named owners, a defined level per field, validation and version history.

The limits appear with scale. Several hundred references, seasonal batches, multiple tiers of suppliers and several delegated acts quickly overwhelm manual files. Spreadsheets also struggle with the access tiers and per-batch inheritance described above. The test is simple: can you reproduce a passport as it stood on a past date, and correct one batch without touching the rest? If not, the tooling has to change before the volumes do.

Leather goods and textiles deserve early attention. Clothing, footwear and leather goods account for 62% of seized counterfeits (OECD/EUIPO, 2025), so the sectors with the heaviest data burden also face the strongest copying pressure.

Yes, in two ways. The carrier must stay with the product, not on packaging that is thrown away. And for goods exposed to counterfeiting, it should resist copying, or a fake will carry your carefully prepared data.

The ESPR is carrier-neutral. In practice, a GS1 Digital Link QR is emerging as the default visible carrier, and NFC is the trust layer: NFC for trust, visible carrier for compliance. The EU recognised NFC as a compliant passport data carrier in 2026 (Implementing Decision (EU) 2026/1736, referencing EN 18220).

Chip class matters for trust. NTAG 424 DNA (ISO/IEC 14443-A, NFC Forum Type 4, AES-128) generates a unique cryptographic response on each tap through SUN messaging. A standard NTAG213-class tag lists around €0.19 to €0.23 at 1,000 to 10,000 units, and cannot do this. Secure NTAG 424 DNA-class labels cluster around $0.45 to $0.65 per unit at five-figure volumes, with on-metal variants adding 7% to 30% (public catalogues, 2026). iPhone XS and XR and later read tags in the background without an app, and Android reads natively with the screen on. Standard tags detune on metal, so on-metal tags with a ferrite layer are needed there.

Linking this data to each unit is also where passports meet product authentication. The same record can serve compliance, authenticity and traceability, and projects that treat product identity as an afterthought tend to rebuild their data model twice.

The mandatory set is a floor. The same passport can carry provenance stories, care guidance, authenticity status, ownership registration and repair booking. Keep regulated fields clearly separated and accurate. Optional content must not blur what the regulation requires.