Your DOI Deposit Pipeline Has a Sunset Date

The quietest risk in a journal operation is often the one that has worked for years. A manuscript is accepted, production creates a DOI deposit, the vendor sends XML to Crossref, the article page goes live, and nobody asks which schema the pipeline is using. Success becomes the evidence that the workflow is healthy.

That assumption now needs a dated review. Crossref updated its metadata deposit schema documentation on June 15, 2026 to say the current recommended schema version is 5.5.0 for complete metadata and resource schema 5.4.0 for resource-only updates. The same page says schema versions 4.4.2 and earlier are currently supported but will be deprecated by the end of 2027: https://www.crossref.org/documentation/schema-library/metadata-deposit-schema/.

For journal leaders, this is not just a technical footnote. A deprecation window turns DOI registration from a background utility into a governance question. Which journals are still depositing through older XML templates? Which vendors know? Which metadata fields are being dropped because the pipeline never moved forward? Which failures would authors, editors, and indexers see first?

A Schema Version Is A Policy Choice

Schema changes matter because Crossref metadata is no longer a bare DOI wrapper. It carries contributors, affiliations, funders, references, licenses, relations, updates, clinical trial information, archive locations, and the signals that let other systems connect the article to the wider research record.

Crossref made that direction explicit in its July 9, 2026 Schema 5.5 announcement: https://www.crossref.org/blog/schema-5.5-now-available-adding-credit-new-record-types-for-blogs-and-posters-and-more/. The release added richer contributor-role support, including compatibility with CRediT, multiple roles per contributor, and a way to mark the corresponding author. It also expanded support for report series metadata, added blog and poster subtypes for posted content, added archive location support, and extended clinical trial metadata across more record types.

A journal does not need every new field on day one. Crossref says adoption can be gradual and that existing deposits remain supported. But gradual adoption is different from accidental stagnation. If the journal has already collected ORCID iDs, CRediT roles, funder details, license terms, or post-publication update information, and the deposit pipeline cannot carry them, the journal is losing value after acceptance.

The Deadline Will Arrive Through Vendors

Most editors will not personally touch a Crossref XSD file. The practical risk sits in platform contracts, production vendor templates, custom scripts, inherited OJS plugins, one-off XML generators, and manual web deposit habits. A society may think its DOI workflow is modern because the public article page looks modern. Behind that page may be a fragile template built around a much older deposit model.

This is where the end-of-2027 deprecation notice should change procurement conversations. A platform demo that says "we handle DOI deposits" is too vague. Journal leaders should ask which schema versions are supported, what the upgrade path is, whether the vendor validates against current Crossref schemas before deposit, how failed submissions are surfaced, and whether metadata updates are free, auditable, and easy to rerun.

The question is not whether a vendor can register a DOI today. The question is whether the journal can keep registering, correcting, enriching, and redepositing records as Crossref moves the infrastructure forward.

Metadata Manager Is A Signal, Not An Escape Hatch

Crossref is also making richer registration easier for members that do not want to work directly with XML. On August 20, 2026, it announced that Metadata Manager now supports books and chapters, reports, dissertations, and post-publication updates for journal content: https://www.crossref.org/blog/registering-rich-metadata-for-books-chapters-reports-dissertations-and-updates-just-got-easier/. Crossref said the tool can capture identifiers for contributors and affiliations, funding information, references, and more, and can edit existing metadata without re-entering the full record.

That is useful, especially for smaller publishers. But it should not become an excuse to leave the core publishing workflow under-specified. Manual tools help when a title has low volume, a special record type, or an exceptional correction. They are weaker as the main control for a busy journal that needs DOI metadata to match submission, production, article pages, JATS, PDF, HTML, and indexing feeds.

Treat Metadata Manager as a benchmark for what staff should be able to see and correct, not as a place where every missing field gets patched after the fact.

Run The Audit From The Article Backward

The fastest useful audit starts with published articles, not architecture diagrams. Choose a recent article with funding, ORCID iDs, contributor roles, references, a data availability statement, a license, and at least one related object such as a dataset or preprint. Then trace what survives in each place.

• Article page: are contributor roles, funders, license terms, corrections, and related outputs visible to readers?
• Production output: do JATS, PDF, HTML, and internal production records carry the same facts?
• Crossref record: does the DOI metadata include the fields the journal already collected?
• Update path: can staff correct a deposited record without rebuilding the article from scratch?
• Failure handling: when a deposit is rejected or partially incomplete, who is alerted and what evidence is kept?

This backward audit is uncomfortable in a productive way. It shows whether metadata quality is being designed into the workflow or rescued by one person who knows where the vendor portal is.

Crossref Is Rebuilding Too

The schema notice is part of a wider infrastructure moment. On August 12, Crossref said its board approved a three-year project funded by about $4.9 million of surplus to accelerate a redesign of the Crossref system: https://www.crossref.org/blog/investing-4.9-million-of-our-surplus-in-rebuilding-the-crossref-system/. Crossref described the current Content System and REST API as large, tightly coupled monoliths and said it plans to move toward separate services for different functions while continuing to maintain the current system.

That does not mean journal teams need to predict every future interface. It does mean the direction is clear: more structured metadata, more editable records, more service-level separation, and less tolerance for old workflows that cannot adapt. If the central infrastructure is being modernized, publishers should not leave their local deposit paths frozen in an earlier generation.

Practical Takeaway For Journal Leaders

Ask for a DOI deposit readiness note before the next production planning meeting. It should name the schema versions your platform or vendor uses, the journals and article types covered, the fields currently lost between submission and deposit, the owner for Crossref submission failures, and the timeline for moving away from any 4.4.2-or-earlier dependency before the end of 2027.

Do not wait until the deprecated path fails. By then the problem will arrive as delayed DOI registration, incomplete article metadata, confused authors, and support tickets from systems that depend on the record. The cheaper moment to fix a deposit pipeline is while it still appears to be working.