Ask a research-information system who wrote a paper and, thanks to ORCID, it can usually give you a clean answer. Ask it where the author worked and the answer dissolves into a free-text affiliation string that might read “Dept. of Physics, Univ. of Manchester” on one paper and “The University of Manchester, School of Natural Sciences” on the next. Organisations have lacked the one thing authors gained a decade ago: a single, open, persistent identifier. The Research Organization Registry — ROR — exists to fix exactly that, and it is now the standard answer for organisation disambiguation across the research-information-systems domain. For the wider explainer index, see the plain-language explainers.
The affiliation problem ROR solves
A single institution can be written hundreds of ways across a corpus of publications: different word order, abbreviations, departmental sub-strings, translations, historic names, and plain typos. For a librarian trying to count their institution’s output, a funder tracking where grant money produced results, or a ranking body comparing universities, that variation is corrosive. Two records that mean the same place do not match; two records that mean different places sometimes do. The cost is paid in hand-cleaning affiliation data that should never have needed cleaning.
A ROR ID attacks the problem at the root. It is a persistent identifier for a research organisation — a short, resolvable URL of the form https://ror.org/ followed by a unique string — that points to a curated registry record holding the organisation’s canonical name, its known name variants, its location, its relationships to parent and child organisations, and cross-references to other identifier systems. When a publisher records a ROR ID against an affiliation, the messy display string becomes secondary: the identifier carries the meaning, and every variant collapses onto one entity.
What a ROR record contains
The point of a registry, as opposed to a bare identifier, is the metadata attached to each entry. A ROR record typically holds:
- A canonical name and a controlled set of name variants, aliases, acronyms, and labels in multiple languages — so all the ways an institution is written resolve to one record.
- Geographic location, expressed against standard place data, which lets systems aggregate output by city, region, or country reliably.
- Relationships to other organisations — parent, child, related, and successor links — so a research institute inside a university, or a hospital trust within a wider group, can be placed in its hierarchy.
- Cross-references to other identifiers, including ISNI and the Crossref Funder Registry, so ROR can act as a hub rather than yet another silo.
Crucially, ROR is open. The registry, its data, and its API are freely available under an open licence, with no membership barrier to use. That openness is deliberate: an organisation identifier is only useful if everyone — publishers, funders, repositories, CRIS vendors — can adopt it without friction.
The legacy: GRID, ISNI, and how ROR relates to them
ROR did not appear from nowhere, and understanding its lineage avoids confusion with the identifiers it sits alongside.
GRID (the Global Research Identifier Database) was a widely used organisation registry developed by Digital Science. ROR launched in 2019 using a snapshot of GRID as its seed data, and the two were aligned for a period. GRID subsequently moved to a model that is no longer openly maintained for community registration, and ROR became the community-governed, open successor for the disambiguation use case. Where older datasets still carry GRID identifiers, ROR records cross-walk to them, so the legacy data is not lost.
ISNI (the International Standard Name Identifier, ISO 27729) is broader: it identifies persons and organisations across the creative and scholarly economy, not just research bodies. ISNI and ROR are complementary rather than competing — ROR records reference ISNIs where they exist. The practical distinction is scope and governance: ROR is purpose-built for research organisations, community-led, and open, which is why it has become the default in scholarly metadata pipelines even though ISNI predates it.
Where ROR fits in the identifier stack
ROR is most powerful in combination with its sibling identifiers, each answering a different question about a piece of research:
- ORCID answers who — the individual researcher.
- ROR answers where — the organisation they are affiliated with.
- DOI answers what — the output itself.
- RAiD answers which project — the activity that produced it.
Linked together, these identifiers turn a pile of disconnected records into a navigable graph: this person, at this organisation, on this project, produced this output. ROR is the organisational anchor in that graph, and its absence is precisely why “output by institution” has historically been so hard to compute.
Adoption and how to use it
ROR is now embedded in major scholarly infrastructure. Crossref and DataCite accept ROR IDs in their metadata schemas, ORCID supports them on affiliations, and a growing number of publisher submission systems and CRIS platforms either store ROR IDs or match incoming affiliation strings against the registry. For an author, the practical step is simple: when a submission system offers to attach a ROR ID to your affiliation, accept it, and check it resolves to the right organisation. For an institution, the higher-value work is ensuring your CRIS records the ROR ID on every affiliation, and that your registry record’s name variants and relationships are accurate — ROR records can be corrected and updated through its curation process.
Where shared vocabulary fits
“Affiliation”, “organisation”, “institution”, “parent”, and “successor” are used loosely across systems, and that looseness is part of why organisation data has been so hard to reconcile. A shared, federated vocabulary that defines these terms precisely — and points back to ROR for the identifier itself and to ISNI for the broader name standard — is what lets one institution’s record mean the same thing in every system that reads it. Supplying that definitional layer is the role the CASRAI dictionary is designed to play; the relevant terms sit in the research-information-systems domain.