Conference Record

Citing a Conference Paper

Conference citations are the ones people get wrong, partly because the elements differ from a journal citation and partly because the source has frequently disappeared by the time anybody needs it. For teams comparing day-to-day work practices, further details provides further context.

Reviewed August 9, 2026. Exact formats vary by style guide; the elements below are what every style needs. A useful external reference point is Wiley .

The six elements

Authors.

Title of the paper.

The proceedings title, which is not the conference name — it is usually "Proceedings of the Nth ..." and may differ from how the conference brands itself.

Conference name, number, dates and place. The edition number matters: a series has many conferences with the same name.

Publisher and year, where proceedings exist.

And an identifier, a DOI or permanent URL, which is what makes the citation survive the venue.

Page numbers where the proceedings have them, which many electronic proceedings do not.

What people get wrong

Citing the conference instead of the proceedings. The paper appeared in a publication; the conference is where it was presented.

Omitting the edition number, which makes a citation in a long-running series ambiguous.

Using the conference website URL as the source. It is the least durable thing available and it is what most people paste.

And citing a talk as a paper when no paper exists — a distinct situation, below.

Citing a talk with no paper

Legitimate and frequently necessary, particularly in industry contexts where proceedings are not the model.

Cite it as a presentation, with speaker, title, conference, date and place, and say it was a conference presentation.

Do not imply a publication. A talk is a talk, and dressing it as a paper is the kind of thing that gets noticed in a review.

And link to the durable copy if one exists — a deposited slide deck or recording rather than the programme page.

When the venue has vanished

The common case for anything more than a few years old.

Cite what the paper was, not where it currently is. A citation identifies a work; a dead link does not invalidate the identification.

Add an archived version where one exists, marked as an archive copy with the date it was captured.

And note if you consulted a preprint or an author's copy rather than the published version, because they can differ.

Do not silently substitute a different version. If you read the author's deposited copy, say so — it is accurate and it helps the next person find the same thing you read.

For your own work

Record the citation at the time. Ninety seconds while the details are on screen against an hour of reconstruction later, and it is the single highest-return habit in this section.

Include the identifier if one was assigned, and note if none was.

And keep the version you actually presented, because the proceedings version and the delivered talk sometimes differ, and somebody asking about the talk means the talk.

The short version