ISSN vs DOI: what each one identifies, and why both exist

The ISSN identifies the serial, the DOI identifies the object. A journal title carries an ISSN, and each of the thousands of articles inside it carries its own DOI. The two codes work at different levels, which is why most scholarly publications end up carrying both.

They get confused because both are unique identification codes for publications, close in spirit to the ISBN used for books. The confusion has a practical cost the moment an editor or a conference organizer has to fill in a metadata form and does not know which number goes where.

The difference at a glance

	ISSN	DOI
What it identifies	A serial publication as a title: journals, magazines, proceedings series, newspapers, even blogs, scientific or not	A digital object: an article, a chapter, a figure, a dataset, an entire issue
Granularity	The title, across its whole run	The individual object
Who assigns it	The ISSN National Centre of the country of publication, or the ISSN International Centre where there is no national centre	A registration agency, usually Crossref in scholarly publishing, through a member organization
When it changes	When the key title changes, and separately per medium and per separately published language edition	Never for a given object. The registered URL behind it changes instead
Resolves to a link	No	Yes, through doi.org

The ISSN identifies the title, not the issue

The ISSN is used for serial publications: journals, proceedings, magazines, and other publications that come out in successive parts with no predetermined end.

The important property is that it is stable. It does not matter how many volumes a publication accumulates, the key title keeps the same ISSN throughout the entire life of that title. A monograph series can also carry an ISSN, and each series gets its own code.

In summary, a distinct ISSN is assigned:

• for the whole duration of a title,
• for each separately published language edition (a serial that simply prints articles in several languages does not need one ISSN per language; a distinct edition published in its own right does),
• for each distinct medium.

That last rule is the one that catches people out. A journal published in print and online has two ISSNs, not one, because the print version and the electronic version are different media of the same work. To keep those versions tied together, the ISSN system provides the ISSN-L (linking ISSN), a single designation that groups all the medium versions of the same serial. When an indexer needs to gather the print and the electronic under one title, the ISSN-L is what does it.

The code itself is eight digits split into two groups of four, and the last digit is a check digit that can be an X. That is why an ISSN sometimes ends in a letter without being a typo.

The DOI identifies the object, and keeps identifying it

The DOI was designed for a digital environment, and its job is to secure both the identity and the durability of what it points at. It can apply to almost any level of the record:

• an entire publication, such as a single issue of a journal,
• a chapter or an individual article,
• a database or a dataset,
• components inside a work: figures, tables, charts, and supplementary data belonging to an article, a book, or proceedings.

Structurally, a DOI has a prefix (always starting with 10.) that identifies the registrant, and a suffix chosen by the registrant that identifies the specific object. Written as a link, it becomes https://doi.org/10.xxxx/suffix, and that is the form that belongs in a reference list.

What happens when someone clicks it is where the DOI earns its keep. The DOI does not store the document. It resolves, through doi.org, to a landing page hosted by whoever published the object, and that page carries the bibliographic information and the access route to the full text. When the journal migrates platforms, changes domain, or reorganizes its URLs, the DOI in every citation ever published keeps working, because only the address behind the code changed, not the code.

That last part is worth being honest about, because it is where the promise actually gets kept or broken. The indirection is not self-maintaining. Someone has to update the registered URL with the registration agency when the address changes, and if nobody does, the DOI resolves to a dead page. Crossref is blunt that there is no purely technical fix for this: a persistent identifier is a commitment by whoever registered it, not a property of the string. The permanence is organizational, and the code is just what makes it possible.

A document keeps its DOI throughout its entire existence, and if it is ever withdrawn, that DOI is not reused. The identifier stays bound to that object and to nothing else, which is what allows a citation to remain unambiguous decades later.

Galoa is a Crossref Sponsoring Member Galoa has been a Crossref Sponsoring Member since 2015. The DOIs of the proceedings and journals we publish are registered directly with Crossref, with the metadata deposited and kept current. How our sponsorship works →

Registration is not a one-time formality either. Getting a DOI means depositing metadata with the registration agency: title, authorship, publication date, the ISSN of the serial it belongs to, references, funding. That metadata is what indexers, reference managers, and discovery systems read. A DOI registered with thin or wrong metadata resolves fine and still fails at the job you registered it for, which is being found and counted correctly.

Why the two codes belong together

The most useful arrangement is the one where both codes work in combination. The ISSN establishes which publication issued the work and holds the title together across media and languages. The DOI pins the individual article and keeps it reachable and citable as platforms and addresses change underneath it.

They also reinforce each other inside the metadata. The ISSN of the serial belongs in each article’s DOI record, and Crossref recommends depositing it. It is one of the signals an indexing service uses to connect a single article back to the journal or proceedings that published it, so that the publication accumulates a coherent record instead of a scatter of unlinked documents.

For a scientific conference, the same logic applies, with a wrinkle worth knowing before you start. Proceedings published as an ongoing series behave like any other serial: one title, one ISSN, and each paper inside each edition registered under its own DOI. A one-off volume is a different case and may carry an ISBN instead, and the ISSN does not follow the series automatically if the title or the medium changes along the way. Decide early which of the two you are running. A conference that settles the structure at the first edition spares every future edition the work, and one that does not leaves the output of a whole community harder to cite than it should be.

If you want to see what the metadata behind your own DOIs currently looks like, the Metadata Refiner audits a publication’s Crossref deposits and shows what is missing.