DPP BUYER GUIDES

Digital Product Passport Integration: ERP, PIM and PLM

Digital product passport integration starts with deciding which system owns each value and when that value is ready to publish. Connecting an API is only one step. This guide gives product data owners and IT teams a practical way to connect existing systems, test failures and compare a provider’s integration proposal.

Give each source system a clear job

Start with your actual systems, rather than an ideal architecture diagram. ERP means enterprise resource planning; PIM means product information management; PLM means product lifecycle management. The names alone do not tell you where an approved value lives. Ask the owner to show the record used by the business today.

The table is an illustrative allocation, not a rule for every company. For each field, record the system that can approve a change and the systems that only receive it. If two teams claim ownership, settle that conflict before building a sync. Keep the decision beside your data register.

The methods and acceptance checks in this guide are our suggested project approach. They are not an official connector specification or a report of software we tested. Use them to request evidence from your own team and shortlisted providers.

Give each source system a clear job
Possible sourceExample responsibilityQuestion to settle
ERPInternal item codes and production batchesWhich code identifies the exact variant?
PLMReleased design and component revisionDoes this revision cover the manufactured batch?
PIMApproved descriptions and language versionsWhich changes are ready for public display?
Evidence storeSupplier files and review recordsCan the passport retain a stable reference?
DPP publishing serviceApproved view for the permitted audienceCan an import bypass review or access rules?

Map product identity before field names

Create a checked link between the source record and the passport’s product identifier. Keep model, batch and item scope visible in that link. GS1’s DPP guide treats identification, discovery and data sharing as distinct building blocks. Your team still needs to decide how its own internal codes map to those building blocks.

Imagine a fictional chair maker with one design code and separate stock codes for fabric colours. A batch import uses the design code alone and assigns the same care sheet to every colour. The file format is valid, but the result may be wrong. Add both variants to the test set and check the mapping against the released design.

Keep the source key, its system name and the mapping version with each imported record. Avoid quietly reusing an old key after a catalogue merge. Ask the team to show what happens when an identifier is missing, duplicated or reassigned.

References: GS1 standards enabling DPP — v2.2

Separate master data from lifecycle events

GS1 EPCIS distinguishes event data from master data. Events describe things that happen in a business process; master data supplies context for interpreting them. EPCIS defines event capture and query interfaces, but does not prescribe how master data enters a system. An EPCIS capability therefore needs a precise explanation in an integration proposal.

For a fictional appliance pilot, the approved model dimensions and a repair performed on one unit serve different purposes. Do not overwrite the model record with a repair note. Ask where the event sits, how it identifies the unit and how a reader sees the relevant history.

Use a small example in the provider demonstration: one model, two units and one repair. Request the exported records as well as the screen view. Check that the unrepaired unit does not inherit the repair. This is a proposed test, not a claim that every DPP must use EPCIS.

References: GS1 EPCIS 2.0 — sections 2.3 and 6.1 (June 2022)

Define an update contract the team can test

For digital product passport integration, document what triggers a transfer and what counts as success. A saved draft, a released design and an approved public change are different events. State which one sends data, which fields it may change and who receives a failure notice.

Choose a manual file, scheduled transfer or event-driven connection based on the required update time and available support. Ask for the maximum delay the business can accept. A live connection still needs a review boundary where source values require approval.

Include a source revision, a stable transfer identifier and a clear retry rule. Request a duplicate-transfer demonstration: send the same change twice and compare the resulting record and history. Then send an older revision after a newer one. Agree whether to reject it or route it to review; do not leave that choice implicit.

Treat deletion and withdrawal separately from a blank value. Ask what happens when a source record is retired. The receiving system should follow the agreed retention and publication policy, not infer that an empty cell authorises erasing a passport.

Reconcile failures across the whole flow

Use a run report with distinct counts for received, rejected, awaiting review and published records. A successful upload should not be presented as successful publication. Assign someone to investigate the rejected set and preserve the reason for each row.

Pause a test connection after a change reaches the receiving system but before the sender records success. Resume it and check that the retry does not create another product. Also test a valid update beside an invalid one so the team can explain partial failure and recovery.

Compare selected published values with their approved source revisions after each rehearsal. Record discrepancies by product and field. Keep this check independent of the connector’s own success message: a second reading of the destination can expose a wrong mapping that the connector accepted.

Keep Registry integration as a separate acceptance step

The Commission publishes a dedicated Service Account Technical Integration Guide for machine-to-machine access to Registry services. It describes signed client assertions and access-token retrieval. This is a separate connection from an ERP feed into your DPP platform; one working connection does not establish that the other works.

Before approving digital product passport integration, ask for three pieces of evidence: a checked product mapping, a failed-update recovery record and a comparison of approved source values with their destination. Add any Registry acceptance that applies to your project as its own item.

Use the data-requirements guide to settle field definitions, then the implementation checklist to set release owners. If a provider controls the integration, include the mapping files and recovery instructions in your exit plan from the beginning.

References: European Commission — Service Account Technical Integration Guide (listed 30 October 2026)

Sources and review context

Sources were reviewed for this editorial guide on 4 November 2026. Provider descriptions are attributed public statements. This guide does not report hands-on software testing or certify regulatory compliance. Check the current original source when a detail affects your decision.

  1. GS1 standards enabling DPP — v2.2
  2. GS1 EPCIS 2.0 — sections 2.3 and 6.1 (June 2022)
  3. European Commission — Service Account Technical Integration Guide (listed 30 October 2026)
How DPP Atlas reviews evidence