Is there a single portfolio size that makes M2M mandatory?
No statutory M2M cutoff is established by the input-method guidance cited here. The EUDAMED Information Centre gives selection tips: interface entry up to 100 devices, XML bulk around 100–1,000, and M2M above 1,000. Treat those as indicative guidance, rather than a legal obligation to buy an integration.
The overlap at 100 and the broad ranges reinforce the practical nature of the advice. Your choice should also reflect the data you already hold, how often it changes, and who will operate the process. The question is the full lifecycle of entry and maintenance, rather than the fastest demonstration of a single upload.
What are the three available input methods?
Manual entry uses the EUDAMED interface. XML bulk upload prepares structured files that an operator uploads through the interface. M2M exchanges data automatically between an external system and EUDAMED backend services, as described in the official data-exchange guidance.
- Manual interface entry
- An authorized user enters and manages records through the application.
- XML bulk upload
- A file-based process with structured data preparation and a manual upload step.
- M2M exchange
- An integrated exchange process requiring technical setup and ongoing operation.
Keep those options separate when comparing providers. A service described as “automated” may generate XML files while leaving upload and reconciliation to a person. That can be a useful solution, but it is different from a maintained system-to-system connection. Ask exactly which parts are included.
When can manual entry be a reasonable choice?
Manual entry can fit a smaller, stable portfolio when an authorized operator has reliable source data and time to check the resulting records. Its apparent simplicity does not remove preparation or review. Assess the whole work queue before assigning the task to someone with spare administrative capacity.
Try a representative sample during planning. Include records with the relevant variation in fields or relationships, then record preparation, entry, correction, and review effort. This is a project estimation method, not a Commission-prescribed threshold test. Use the observations to understand your own workload without turning a tiny sample into an industry claim.
Also consider continuity. Document where the source information lives and who can make future changes. A manual process understood by one colleague can become difficult to maintain when that person leaves or the portfolio expands. A simple written procedure helps regardless of which tool is used.
When is XML bulk upload worth assessing?
Bulk upload is worth considering when source data can be mapped into the required XML structure and the project benefits from preparing several records together. The upload itself remains a distinct operator action. File generation, validation, error correction, and reconciliation still need clear ownership.
The Commission’s data-exchange documentation states that XML follows the supplied service and entity schemas. Use the current production documentation and rules. An old sample file or a Playground schema is not a reason to assume the same payload will be accepted unchanged in production.
Ask a provider to demonstrate how source rows become XML records and how response errors are connected back to the original data. Preserve a reviewed version of each source file. Without that linkage, corrections can be made in one output while the underlying spreadsheet continues to generate the same problem.
What additional commitments come with M2M?
M2M needs technical onboarding and a supported exchange setup, rather than just a login to the interface. The official prerequisites include actor registration, an active Local Actor Administrator, Access Point details, relevant security keys, and successful Playground onboarding before production.
Confirm who will operate the connection, protect credentials, respond to failed exchanges, and manage future changes. Ask which responsibilities belong to your company, your integration provider, and any third-party service. A working demonstration should be followed by an operating model that covers routine failures and maintenance.
Assess whether your internal source system is ready to supply approved data consistently. A permanent connection can transmit changes efficiently, but it cannot decide whether a new group is justified or whether a source attribute is correct. Keep professional approval and technical transmission as separate controls.
How should you compare the total effort?
Compare preparation, implementation, submission, review, correction, reconciliation, and future maintenance. Include the people who own the data as well as the people who run the upload. A low entry price is difficult to assess if substantial mapping and correction work is excluded from the proposal.
Ask providers to describe the same portfolio boundary and the same expected outcome. Otherwise, one quote may cover file preparation while another covers a maintained integration, and the prices will not answer the same question. Request an explicit list of assumptions and customer responsibilities rather than attempting to infer them from a product name.
For recurring work, consider how frequently records change and what initiates an update. Define the route from an approved product change to a completed database update. The value of integration depends partly on that recurring workflow; a one-time backlog may have a different best option.
What data problems survive every transmission method?
Incorrect actor identity, unsupported grouping, missing attributes, and confused identifiers remain problems whichever method carries the data. Resolve them against authoritative sources and the current EUDAMED dataset before treating a tool as the solution. Read Basic UDI-DI explained for the structural questions.
Build a reconciliation process that compares the intended portfolio with the resulting records. Ask how the team distinguishes excluded rows, failed submissions, and accepted entries. Keep exceptions visible rather than allowing an overall “upload successful” status to stand in for a record-level review.
How do you make a decision you can defend later?
Write a short decision note covering portfolio scope, recurring change volume, source readiness, the chosen method, and who owns maintenance. Include why the alternatives were less suitable for this project. This is practical internal documentation, not a requirement asserted by this site.
Revisit the choice when the portfolio or operating model changes. A process appropriate for an initial backlog may need different support for routine releases. If you need an independent specialist to assess the work, describe your situation. If the first uncertainty is what public records already exist, request a free registration check.