Cite a source without a DOI

Fallback priorities for persistent links

Registered identifier: Handle, URN, ARK or another persistent scheme documented by the repository.
Stable repository record: institutional landing page containing title, creator, version and files.
Official publication page: publisher, university, public body, organisation or project.
Direct document address: where no record page exists; also name the parent source.
Ordinary webpage: complete URL, title, operator and access date.
Archived state: where lawful and useful, with original address and archive date.
No dependable link: use full offline metadata, a library record or archive shelfmark.

Prefer the address that explains identity and version, not merely the shortest one.

Build identity for the source type

A book requires creator, title, edition, publisher and year according to the chosen style. A chapter adds chapter author, chapter title, edited volume and page range. A report needs issuing body, title, report number, year and version. A webpage normally needs responsible party, page title, publication or update date, URL and access information. Archive material is located by collection, shelfmark and object description.

Missing DOI is not a reason for thin metadata. Inspect title page, imprint, catalogue and official landing record. A filename such as “report_final.pdf” is not a publication title. Represent group authors consistently.

Prefer records to fleeting downloads

Repositories commonly provide both a record and a direct file link. Prefer the record because it explains version, creator, licence and related files. The download path may change after an update. Still preserve the file used and record its filename, version and checksum in the research log.

Remove tracking parameters, session identifiers and search paths where the clean official address identifies the same object. Test the link in a private browser without authentication. An institutional subscription link may fail for readers even though the bibliographic identity remains correct.

Use access dates and archives proportionately

Access dates matter for changeable pages but do not replace creator, title and publication date. Stable published objects may be treated differently by your citation style. Do not invent a date when none is stated; follow the style's method for missing dates.

Web archives can preserve a state, yet dynamic applications may be incomplete and archiving may not always be appropriate. Retain a lawful local copy or provenance-rich screenshot for your evidence without republishing protected content. Distinguish a captured state from the continuing official page.

Findability, permission to access and scholarly quality are separate. A persistent link is not a quality mark, and a paywall does not make a publication unciteable.

Pre-submission workflow

Confirm through DOI, catalogue and repository searches that no DOI exists.
Determine source type and complete core metadata.
Select the most stable official landing page.
Match version, date and file state.
Test the URL without personal or temporary parameters.
Check publication status and corrections.
Format the reference under the required style.
Run a link check and content sample.

Keep a short decision note: “No DOI in Crossref or publisher record; institutional Handle used” or “No landing page; official PDF address and report number used.” It demonstrates that the identifier was investigated rather than forgotten.

The source evaluation hub joins identity and quality checks. Check a DOI explains registry matching. The citing sources overview covers full references.

Weak fallbacks to avoid

A search-engine result instead of the original.
A local file path or personal cloud share.
A URL containing an expired session token.
An organisation homepage without the work.
A shortening service that conceals the destination.
A link to a version other than the one read.
An invented DOI borrowed from a similar article.

If no URL appears durable, a complete catalogue or archive reference can offer better findability. Never cite a different object merely because its address looks stable.

During final checking, open the link and compare title, creator and version with the bibliography. A technically live link may have been redirected to a replacement document. Record changes and update the reference only after confirming that the cited claim remains present.

Build an archive trail for changeable web sources

For a web page without a DOI, the live URL is an address rather than a promise of long-term retrievability. Record the responsible author or institution, page title, publication or revision date, relevant section and access date. Check whether the institution also provides a citable PDF, repository record or official archive link. That first-party, more stable representation is normally preferable to a private mirror or a search-results URL.

If the page is continuously updated, the reference and research notes should identify the state actually used. Capture a version number or “last updated” date where the publisher supplies one. When neither exists, an institutional web archive can preserve evidence of the consulted state. Keep the archive URL alongside, rather than silently substituting it for, the original address. This distinction preserves both provenance and the frozen representation.

A compact provenance record makes later checking faster: responsible body, document type, original URL, any alternative persistent identifier, archived snapshot, access date and a note identifying the cited passage. For a dynamic table or dashboard, save the filter choices and download date as well. The record remains useful even when navigation or filenames change. If authorship, issuing body and publication state cannot be established, disclose that limitation and consider replacing the item with a stronger primary source instead of treating the missing DOI as the only problem.

Official foundations

Sources reviewed 4 September 2026.

How to cite a source without a DOI and keep it findable