For most readers, a journal is no longer something they find on a shelf or in a printed catalogue. It is something a system finds for them. A researcher types a query into a discovery layer, a citation database, a general search engine or an AI-assisted literature tool. A librarian checks whether a title is covered by a knowledge base before activating it. An author looks for a suitable venue through a journal finder. Each of these routes depends on machine-readable records describing the journal and the articles it publishes.
That shift changes what "visibility" means for a journal. A well-edited title with thin, inconsistent or missing metadata can be harder to find than a weaker journal with clean, complete records. Discovery has become, to a large extent, a data-quality and interoperability problem, and it is one that publishers control more directly than they often realise.
This matters most for small and independent publishers, who rarely have dedicated metadata staff but whose journals depend on being found through the same systems as the largest portfolios. The good news is that the standards and infrastructure needed to do this well are mostly open, documented and inexpensive to adopt.
The record behind the article
When a publisher registers a DOI with Crossref, it deposits metadata describing the work: title, authors, publication details and, optionally, a great deal more. The optional elements are where discovery gains or loses ground.
Crossref's Participation Reports show, for each member, what percentage of its records include eleven key metadata elements: references, abstracts, ORCID iDs, affiliations, ROR IDs, Funder Registry IDs, funding award numbers, Crossmark, text-mining URLs, licence URLs and Similarity Check URLs. The reports are free and open to anyone, which means a publisher's metadata completeness is visible to librarians, funders, indexers and competitors alike.
Each of these elements does practical work downstream:
- Abstracts give search systems text to match against queries; without one, a record may be found only by its title.
- References allow citation links to be built, so an article can be reached by following citations from other work.
- ORCID iDs connect articles to the right authors, avoiding confusion between researchers with similar names.
- ROR IDs tie affiliations to a persistent identifier for an institution, so a university can find its own output and a funder can track outputs from the organisations it supports.
- Funder IDs and award numbers let funders and research offices link publications to grants.
- Licence URLs tell machines whether and how content can be reused, which matters for open access discovery and text mining.
A journal can publish excellent research and still deposit records with none of these elements. The articles will have DOIs and will resolve, but they will be far less connected to the rest of the scholarly record.
Identifiers that hold the system together
Persistent identifiers are what allow different systems to agree that they are talking about the same thing. DOIs identify works, ORCID iDs identify people, ROR IDs identify organisations, and the ISSN identifies the journal itself.
The ISSN is easy to overlook, but it is the key many systems use to recognise a serial publication. According to the ISSN International Centre, an ISSN is an eight-digit code with no intrinsic meaning: it says nothing about a journal's content or quality, only which publication it is. An electronic version of a print journal must have its own ISSN, even if the title is identical, and a significant change to a publication calls for a new one. Publishers that display the wrong ISSN for a format, or keep using an old ISSN after a significant change, create mismatches that ripple through library catalogues, knowledge bases and indexing records.
Libraries see journals through knowledge bases
In most academic libraries, users do not reach a journal through the publisher's homepage. They reach it through a discovery service and a link resolver, which rely on a knowledge base recording which titles and date ranges the library can access.
The standard that governs much of this exchange is KBART (Knowledge Bases and Related Tools), a NISO Recommended Practice for transferring holdings metadata from content providers to knowledge base suppliers and libraries. NISO explains that metadata problems in that supply chain make link resolvers and electronic resource management systems inaccurate, and KBART exists to standardise the data. Its third phase, NISO RP-9-2026, was approved in August 2026, with goals including better support for additional content types and global content.
For a publisher, the implication is concrete. If a journal does not supply accurate KBART holdings files, or supplies them irregularly, a library that has paid for access or wishes to activate an open access title may find that its users cannot reach the content. Researchers then conclude that the journal is unavailable, when the real failure is a missing or out-of-date file.
Open aggregators and the AI layer
A growing share of discovery runs through services that aggregate open metadata at scale. OpenAlex, an openly licensed catalogue of the research system, is a clear example. Its documentation lists its core inputs as Crossref and DataCite, ORCID and ROR, PubMed and PubMed Central, DOAJ and the ISSN International Centre, alongside repositories, publisher feeds and its own extraction from open access content, and it releases all of its data under a CC0 public-domain dedication.
Services built this way inherit the quality of their inputs. If a publisher's Crossref records lack abstracts, affiliations or references, an aggregator has less to work with, and so does every tool built on top of it. The same logic applies to the AI-assisted search and summarisation tools that researchers increasingly try: a system can only retrieve, rank or recommend articles that it can find and interpret. Open, structured, complete metadata gives a journal a fair chance of appearing in those results. Sparse or closed metadata does not.
It is worth being precise about what this is not. Appearing in an aggregator or an AI tool is discovery, not indexing in a selective database such as Scopus or the Web of Science Core Collection, and neither is the same as long-term archiving or citation metrics. Good metadata helps a journal be found. It does not substitute for the editorial evaluation that selective indexes apply. Our explainer on why journals lose indexing covers that separate process.
The policy push towards open research information
The infrastructure behind discovery is increasingly shaped by explicit principles about openness and governance.
The Principles of Open Scholarly Infrastructure (POSI), first published in 2015 and revised as version 2.0 in October 2025, set out commitments on governance, sustainability and insurance for organisations that provide scholarly infrastructure. Adopters listed on the POSI website include Crossref, DataCite, ROR, DOAJ, OpenAlex, OpenCitations and the Public Knowledge Project, the organisation behind the widely used Open Journal Systems software.
The Barcelona Declaration on Open Research Information, published on 16 April 2024, asks research-performing organisations and funders to make openness the default for the research information they use and produce, to work with services and systems that support open research information, to support the sustainability of the infrastructure, and to act collectively to accelerate the transition. Crossref, OpenAlex and ORCID are listed on the declaration's website as supporters.
Taken together, these initiatives signal where the institutions that buy, fund and assess research are heading. Universities and funders that commit to open research information will increasingly rely on open sources to understand publications, and publishers whose records are incomplete in those sources will be less visible in the analyses that inform library and funding decisions.
What small publishers can do
None of this requires a large budget. It requires attention and routine.
- Check your Crossref Participation Report. It shows exactly which metadata elements are missing across your records. Prioritise abstracts, references, ORCID iDs and affiliations with ROR IDs.
- Deposit references and abstracts as standard practice, and make sure your production workflow or platform passes them through to Crossref rather than leaving them on the article page only.
- Audit your ISSNs. Confirm that print and online ISSNs are registered and displayed correctly, and check with your ISSN national centre whether any significant change to the journal requires a new one.
- Supply KBART holdings files to the knowledge base providers your library customers use, and update them whenever coverage changes.
- Record licences and funding information in machine-readable form, especially for open access content.
- Correct errors at source. Fixing a record in Crossref improves every downstream service that harvests it, whereas fixing it in one database fixes only that database.
Conclusion
Journal discovery now happens largely between machines before a human reader arrives. Researchers, librarians and authors meet a journal through records assembled from Crossref deposits, identifier registries, knowledge base files and open aggregators, and increasingly through AI tools built on the same foundations. Publishers that treat metadata as a core editorial responsibility rather than a production afterthought give their journals a better chance of being found, read and connected to the rest of the scholarly record. For small publishers in particular, complete and accurate metadata is one of the few visibility advantages that does not depend on size.
