A brand tags 300 bottles by hand, builds a passport page, and calls the pilot a success. Six months later the filling line cannot apply tags at speed, suppliers have not sent data for the rest of the range, and nobody owns the passport. A DPP implementation roadmap prevents this stall. It runs through five phases: discover, design, pilot, scale and operate. Each phase ends with a go/no-go gate, and the pilot is built to prove production, not to impress.
Pilots rarely fail on technology. They fail because they were built to be shown, not to be repeated.
The timeline leaves little room for a late discovery. The central EU DPP registry has been live since 20 July 2026 (European Commission). Under Regulation (EU) 2024/1781 (ESPR), obligations start at least 18 months after each delegated act. Textiles and apparel are indicatively expected in 2027, which points to enforcement around 2028 to 2029. Batteries follow their own passport rules from 18 February 2027.
Four patterns recur when a pilot cannot scale:
All four are visible at the pilot gate if the criteria were agreed in advance. That is the case for writing them down first.
A DPP implementation roadmap is a phased plan that takes a brand from understanding its obligations to running accurate passports across every product, site and supplier. It sets decisions, owners and exit criteria for each phase.
The distinction that matters: a passport page is a deliverable, while a passport programme is an operating capability. The page can be built in weeks. The capability needs line application, data feeds, access rules, versioning and a named owner.
| Phase | Goal | Key output | Typical owner |
|---|---|---|---|
| 1. Discover | Know what applies to which products, and when | Product-to-obligation map, owners named | Compliance with operations |
| 2. Design | Make the hard-to-change decisions once | Identifier, carrier, data model, hosting and access decided | Cross-functional steering group |
| 3. Pilot | Prove production readiness on one line | Results against agreed criteria | Operations and digital |
| 4. Scale | Extend across lines, sites, co-packers and suppliers | Rollout plan by product group | Operations, procurement |
| 5. Operate | Keep passports accurate and extend them | Maintenance, versioning, monitoring | Named passport owner |
The phases overlap. Supplier data collection should start in phase 1 and run through phase 4, because it is the slowest workstream and depends on people outside the company.
The programme moves through five steps, each with a gate.
The fourth question in step 1 is the one most often skipped. It also changes the design most. If brand protection needs authentication on the same products, the carrier choice in step 2 will differ.
Five decisions shape everything downstream. Three are physical or permanent.
| Decision | Options | Why it is hard to change |
|---|---|---|
| Identifier granularity | Model, batch or item | A change means new serialisation and new carriers on products |
| Data carrier | QR, NFC, copy-resistant print, combinations | It is physically on every product shipped |
| Resolver domain | Brand-owned or provider-owned | Every identifier points to it for the product’s life |
| Data model and sources | Fields mapped to PLM, PIM, ERP, suppliers | Changeable, but costly if sources are wrong |
| Hosting and access | Brand-hosted, provider, hybrid; public, restricted, owner views | Changeable in software if the model supports roles |
On identifiers, GS1 Digital Link with a GTIN plus serial gives a standards-based structure that works for item-level passports. On the domain, a brand-owned resolver means a change of provider does not break products already in market.
A principle for the carrier question: NFC for trust, visible carrier for compliance. The ESPR is carrier-neutral, and a GS1 Digital Link QR is emerging as the practical default for the passport itself. NFC adds the trust layer when counterfeiting or diversion is a risk. Brands that want a deeper comparison can read NFC vs QR vs RFID vs digital watermarks.
The two are not rivals for the same job. A QR carries the passport to anyone with a camera. A secure NFC chip proves the item is the one the brand made.
| Criterion | Serialised QR (GS1 Digital Link) | Secure NFC (NTAG 424 DNA class) |
|---|---|---|
| Passport access | Any camera | Phone tap, no app on iPhone XS/XR and later (background read); Android reads natively with screen on |
| Copy resistance | Low: a printed code can be duplicated onto a fake | High: AES-128 challenge-response, SUN messaging generates a new signed code on every tap |
| Standards | GS1 Digital Link, GTIN + serial | ISO/IEC 14443-A, NFC Forum Type 4 |
| Marginal cost | Close to zero to print; serialisation and platform are the real cost (industry sources, 2026) | About $0.45 to $0.65 per label at five-figure volumes, on-metal +7 to 30%, 50k+ quote-only (public catalogues, 2026) |
| Metal and liquids | Unaffected | Standard tags detune on metal; on-metal tags with a ferrite layer solve it |
| Fits | Compliance, consumer information | Authentication, tamper evidence, ownership, grey market control |
A standard non-secure NFC tag (NTAG213 class) lists around €0.19 to €0.23 at 1 to 10k units (public catalogues, 2026). It is cheaper, but its identifier can be read and cloned. It carries a passport link. It does not prove authenticity.
Between the two sit copy-resistant print options such as ACM-secured barcodes, which suit products where a chip is impractical. For the supply chain side, ISO 15693 vicinity-class tags such as ICODE DNA serve scanning in logistics, but they are not the phone-tap authentication class.
Often. A printed serialised QR is the better answer in four situations:
In these cases, the roadmap does not change. The identifier, domain and data model decisions are the same. Only the carrier line in the design table reads differently.
The reverse also holds. Counterfeits are concentrated in the categories where premium brands operate: clothing, footwear and leather goods account for 62% of seized counterfeits, and counterfeit trade reached USD 467 billion, up to 2.3% of global trade, in 2021 (OECD/EUIPO, Mapping Global Trade in Fakes, 2025). For those categories, a copyable carrier leaves the passport open to the same duplication as the product.
A good pilot is designed to fail early if it is going to fail at all. Run it in production conditions, and agree these criteria before it starts.
| Criterion | What to measure | Why it matters |
|---|---|---|
| Line application | Line speed with tags applied, any slowdown | Hand-tagging hides the largest scale-up cost |
| Coverage | Share of units commissioned with a working identity (target 99%+) | Every gap is a product with no passport |
| Data completeness | Share of required fields populated from real sources | Manual entry does not scale |
| Read experience | Scans from different phones, languages and markets | The passport must work for everyone, without an app |
| Access control | Public, restricted and owner views from one identifier | Shows the model supports regulation and other uses |
| Reuse | Whether other teams use the same identity | Shows the investment serves more than compliance |
The distinction that matters: a pilot proves production readiness only if it uses the line, the data sources and the identity that will exist at scale. Anything else proves a demo.
Pick the pilot line for learning, not visibility. The best candidate has a real passport timeline, a real counterfeiting or diversion problem, and a setup typical of the range. A limited edition made in one atelier says little about the rest of the business.
Each gate is a short, honest review.
| Gate | Question |
|---|---|
| After discover | Are product groups, timelines, existing assets and owner known? |
| After design | Are identifier, carrier, domain and data model agreed by every affected team? |
| After pilot | Did the pilot meet its criteria on a real line, with real data, at real speed? |
| During scale | Is each new line, site or co-packer reaching coverage and completeness targets? |
| In operation | Are passports accurate, versioned and owned, and are new acts being mapped? |
A gate that cannot produce a “no” is a ceremony. Define in advance what result sends a team back a phase.
Scaling is operational work. The design does not change. It is applied again and again.
Co-packers deserve particular attention. They often run the lines where tags are applied, yet have no reason to send commissioning data unless asked, equipped and contractually obliged. A product that ships without a recorded identity has no passport, however good the data behind it.
Requirements and constraints differ by vertical.
For luxury supply chains specifically, how the DPP will affect them covers the sector detail.
One named person, accountable across product groups. The role covers passport accuracy, versioning, corrections by batch, and mapping new acts onto the existing data model.
Operation also needs continuity. Identifiers on a brand-owned domain and exportable data mean a provider change does not break products already in market. The practical background is set out in the practical guide to the Digital Product Passport.
40+ brands · 12 years · €1.5B protected
Pilots that reuse the production line, the data sources and the identity planned for scale turn a deadline into a schedule. Selinko, part of Toppan, supports brands through each phase with its platform, from encoding on the line to public, restricted and owner views from one identifier.
A DPP implementation has five phases: discover, design, pilot, scale and operate. Discovery maps products to obligations. Design fixes identifier, carrier, domain and data model. The pilot proves production readiness on one line. Scale extends it across sites and suppliers. Operate keeps passports accurate. Each phase ends with a go/no-go gate.
DPP pilots fail to scale because they test a demo, not production. Tags are applied by hand, data is typed in, one team chooses the carrier, and nobody owns the passport afterwards. Line speed, coverage and data completeness stay unmeasured until the scale-up budget meets real constraints.
A brand should start discovery now, since supplier data is the slowest workstream. The EU registry has been live since 20 July 2026. Textiles are indicatively expected in 2027 with obligations at least 18 months after the act, and the battery passport applies from 18 February 2027.
Pilot on a line chosen for learning: one with a real passport timeline, a real counterfeiting or diversion problem, and a production setup typical of the range. A limited edition from a single atelier proves little. The goal is to expose scale-up constraints early, not to showcase one product.
Yes. Co-packers often run the lines where tags are applied, but they have no reason to send commissioning data unless equipped and contractually obliged. A unit that leaves without a recorded identity has no passport, so co-packer onboarding belongs in the scale phase, with contractual terms.
No. The ESPR is carrier-neutral, and a GS1 Digital Link QR is emerging as the practical default for the passport. NFC adds a trust layer: NTAG 424 DNA class chips use AES-128 and SUN messaging to prove an item is genuine, which a printed QR cannot do.
Plan your DPP pilot