What Is Co-Citation in SEO? Meaning, Examples and How It Differs From Co-Occurrence

Co-citation in SEO means an independent page names or links two brands together. Where the term came from, examples by type, and how co-occurrence differs.

Rankbox Team

November 17, 2026 · 11 min read

On this page8 sections

The short answer

Co-citation in SEO is when an independent page names or links two brands, sites or pages together, which suggests they're related. The more independent pages pair the same two names, the stronger that suggestion gets. A roundup that lists your product beside a category leader is co-citation, whether or not it links to either of you.

Co-occurrence is the term people mix it up with. Co-occurrence pairs one name with topic words, such as your brand and "retainer invoicing" in the same paragraph. Co-citation pairs two names, and a third party does the pairing.

This is the plain definition, with the term's history, examples and a simple way to count it. To earn co-citations next to the leaders in your market, read our co-citation SEO playbook. For what studies show about co-citation and AI answers, see our guide to link building and co-citation.

Key Takeaways

  • Co-citation in SEO means a third, independent source names or links two things together. Its strength is the number of independent sources that pair them.
  • The idea comes from citation analysis. Henry Small and Irina Marshakova proposed it independently in 1973, ten years after M. M. Kessler described its mirror image, bibliographic coupling.
  • Search engineers used link co-citation to find related web pages by 1998 and 1999, in a Xerox patent and a Compaq research paper.
  • In 2012 SEOs argued over the word after Rand Fishkin used it for what he later called co-occurrence. Both terms are still used loosely.
  • For AI answers, a passage that names you beside a leader can be retrieved as one unit. Whether models learn category membership from it is an inference, not a documented fact.

Where the Idea Behind Co-Citation in SEO Came From

The term is more than 50 years old, and it started in information science, not search engines.

1973: two researchers, one idea

Henry Small published "Co-citation in the scientific literature: A new measure of the relationship between two documents" in the Journal of the American Society for Information Science in July 1973. The idea was simple. If later papers often list papers A and B together in their references, A and B are probably about related things, even if neither cites the other. Count how often that happens, and you have a measure of how close they are.

Irina Marshakova reached the same idea the same year, in a paper published in Russian. A 2020 paper in Information Research puts it plainly: "In 1973 Henry Small and Irina Marshakova independently introduced the notion of co-citation." Crossref counted 4,253 citations of Small's paper on 1 October 2026.

1963: bibliographic coupling, the mirror image

Ten years earlier, M. M. Kessler described bibliographic coupling in American Documentation. Two papers are coupled when they cite the same earlier work. The two ideas look alike but point in opposite directions:

Bibliographic coupling (1963)Co-citation (1973)
Who makes the linkThe two documents themselves, through what they citeLater documents, through what they cite together
When it's fixedThe day the two documents are publishedNever; it grows as new sources pair them
What it reflectsThe authors' own choicesOther people's judgment
SEO equivalentTwo sites linking to the same sourceA third page naming or linking both sites

That last row is why co-citation in SEO works as a signal. It measures what others say about you, which you can't simply write yourself.

1998 to 1999: from papers to web pages

On the web, a hyperlink plays the part of a citation. Researchers spotted that early. A Xerox patent by James Pitkow and Peter Pirolli, with a priority date of March 1998, describes "finding related collections of linked documents using co-citation analysis." In 1999, Jeffrey Dean and Monika Henzinger of Compaq's research lab published "Finding related pages in the World Wide Web", whose Cocitation algorithm treated two pages as related when many of the same pages linked to both. That's link co-citation: two sites linked from the same third page.

How SEOs Started Using Co-Citation in SEO

For years the term stayed in research papers. It entered everyday SEO talk in one week of November 2012.

The 2012 Whiteboard Friday

On 16 November 2012, Rand Fishkin recorded a Moz Whiteboard Friday predicting that anchor text was weakening as a ranking signal. In the video he called the replacement "co-citation of terms and phrases along with a brand or a link." His example: many articles mention "cell phone ratings and reviews" together with Consumer Reports without linking to it.

The post now carries his correction. The signal he meant was "co-occurence (which I mistakenly call 'co-citation' in the video)." The post's title changed to say co-occurrence too.

The pushback that sharpened the terms

Two SEO writers explained the mix-up within two weeks. Bill Slawski wrote on SEO by the Sea on 21 November: "Things that co-cite cite things together." Fishkin's examples, he argued, were about words appearing on the same pages, and he tied them to Google's patents on phrase-based indexing. A week later, Joshua Giardino of iAcquire called the effect "Lexical Co-Occurrence", "the nerdy Natural Language Processing term for what he describes as 'Co-Citation'."

Where that left the term

SEOs still use co-citation in two senses: the classic link sense, and brand co-citation, where two brands are named together in one passage. Google's documentation doesn't define either. The closest it comes is its How Search Works page, which says one quality factor is "if other prominent websites link or refer to the content." A 2012 Google patent describes an "implied link" as "a reference to a target resource ... which is included in a source resource but is not an express link." A patent shows an idea Google protected, not a system it runs, and Google Patents lists this one as expired over unpaid fees.

Co-Citation in SEO vs Co-Occurrence

Here is how the two terms differ in practice. Tallyfold is a fictional invoicing app for agencies, and Brindlework and Kestrelyn are its fictional rivals.

Co-citationCo-occurrence
What gets pairedTwo entities: brands, sites, pages or peopleOne entity and the words of a topic
Who does the pairingA third source, independent of bothAny text, including your own pages
ExampleA roundup names Brindlework and Tallyfold in one paragraphA paragraph puts "Tallyfold" and "retainer invoicing" together
What it suggestsThe two belong in the same groupThe entity is about that topic
Where the idea comes fromCitation analysis, 1973Linguistics and language processing

Both matter for AI search, for different reasons. Co-occurrence teaches what you are. Co-citation places you among the others in your category. Our entity SEO strategy guide covers choosing the topics you want to co-occur with.

The Pair Test

When you aren't sure whether a page shows co-citation in SEO terms or plain co-occurrence, run the Pair Test, our three-question check:

  1. Are two named things paired? If it's one name plus topic words, it's co-occurrence.
  2. Is the source independent of both? If one of the two brands published it, it's a self-made pairing. That's useful context, but it's your claim, not a co-citation.
  3. Are both names in the same passage? One paragraph is a strong pairing. The same page in different sections is a loose one.

Yes to all three is a strong co-citation. Yes to the first two and no to the third is a loose one. A no on either of the first two means it isn't co-citation at all.

Examples of Co-Citation in SEO, by Type

Each example uses the fictional brands above. The first six are co-citation in SEO terms. The last two look similar but aren't.

  1. Link co-citation. A resource page for agency owners links to both brindlework.example and tallyfold.example. This is the classic sense from the 1999 research.
  2. Brand co-citation in a roundup. A "best invoicing software for agencies" article says Brindlework, Kestrelyn and Tallyfold all send recurring invoices. No links are needed for the pairing to count.
  3. An unlinked news mention. A trade article about late payments quotes Tallyfold's founder in the same paragraph that names Brindlework.
  4. A review-site category page. A software review site lists all three tools on one category page, each with ratings. Our glossary entry on brand mentions covers this kind of mention.
  5. A podcast episode page. The show notes name the guest's company and the two rivals the host asked about.
  6. A forum thread. A user compares Kestrelyn and Tallyfold for a five-person agency, and other members reply.
  7. Not co-citation: your own comparison page. Tallyfold's "Tallyfold vs Brindlework" page pairs the names, but Tallyfold wrote it. Our comparison page formula shows how to make that page fair and useful.
  8. Not co-citation: a topic pairing. A blog post that puts "Tallyfold" next to "late client payments" is co-occurrence.

How to Count Co-Citation in SEO: A Worked Matrix

Small's original measure was a count, and you can apply the same arithmetic to brands. Take a set of independent pages, note which brands each one names in a single passage, and count the pairs. The numbers below are a toy sample of eight pages, invented to show the method.

PageBrands named in one passage
P1Brindlework, Kestrelyn, Tallyfold
P2Brindlework, Kestrelyn
P3Brindlework, Tallyfold
P4Brindlework, Kestrelyn
P5Kestrelyn, Tallyfold
P6Brindlework, Kestrelyn, Tallyfold
P7Brindlework
P8Brindlework, Kestrelyn

Now build the matrix. Each cell counts the pages that name both brands. The diagonal counts the pages that name that brand at all.

BrindleworkKestrelynTallyfold
Brindlework753
Kestrelyn563
Tallyfold334

To compare pairs fairly, divide each pair count by the pages that name either brand. That's our simple co-citation strength score:

  • Brindlework and Kestrelyn: 5 ÷ (7 + 6 − 5) = 5 ÷ 8 = 0.625
  • Brindlework and Tallyfold: 3 ÷ (7 + 4 − 3) = 3 ÷ 8 = 0.375
  • Kestrelyn and Tallyfold: 3 ÷ (6 + 4 − 3) = 3 ÷ 7 = 0.429

The two leaders are paired most strongly, as you'd expect. Tallyfold sits a little closer to Kestrelyn than to Brindlework, so pages that cover Brindlework without Tallyfold, such as P2, P4, P7 and P8, are where its next co-citations would come from. For a real audit across the pages AI engines cite, use the method in our co-citation SEO playbook, and find the pages with prompts from our free AI Visibility Prompt Kit.

How Co-Citation Relates to AI Answers

AI search makes co-citation in SEO more concrete than it was in 2012. Google says its AI features retrieve pages through its ranking systems, and then its systems "review the specific information from those retrieved pages". A passage that names three brands reaches the answer with all three names in it.

Inside the models, research points the same way. Kang and Choi found language models are "vulnerable to the co-occurrence bias, defined as preferring frequently co-occurred words over the correct answer." So what appears together in training text shapes what a model recalls.

The step from there to "co-citation teaches a model that you belong in the category" is an inference. No AI vendor documents it, and no study isolates it. Our co-citation evidence guide grades each claim. Google also warns, in the same AI guide, that "seeking inauthentic 'mentions' across the web isn't as helpful as it might seem." Only earned pairings count.

Earned co-citations come from pages other people choose to write, so you can't buy or automate them. What Rankbox can do is write your side of the story. Its Citation-Ready Writer researches the live web and writes 2,000 to 3,500-word source-backed articles, such as a fair comparison page or a data write-up an editor could cite. Rankbox doesn't pitch editors or track AI citations. The Business plan is $49.50 a month with a 7-day trial (pricing).

Frequently Asked Questions

What is co-citation in SEO?

Co-citation in SEO is when an independent page names or links two brands, sites or pages together, which suggests they're related. The more independent pages pair them, the stronger that signal. A roundup that lists your product next to a category leader is an example, with or without a link.

What is the difference between co-citation and co-occurrence?

Co-citation pairs two named things, such as two brands, in a third party's page. Co-occurrence pairs one name with topic words, such as a brand and "retainer invoicing" in one paragraph. Co-citation places you among rivals. Co-occurrence ties you to a subject.

Who invented co-citation?

Henry Small and Irina Marshakova introduced co-citation independently in 1973, in citation analysis of scientific papers. Small's paper appeared in the Journal of the American Society for Information Science. Web researchers at Xerox and Compaq applied the idea to links in 1998 and 1999.

Not in the sense SEOs mostly use today. The classic link sense needs one page linking to both sites, but brand co-citation only needs a page that names both brands in one passage. Unlinked pairings still reach readers and AI answers, though links also help pages get found and ranked.

Is co-citation a Google ranking factor?

Google doesn't name it as one. Its How Search Works page says prominent sites that "link or refer to" content are a quality factor, and a 2012 Google patent describes implied links. A patent isn't proof of use, so treat co-citation in SEO as a plausible signal, not a confirmed factor.

How do I get co-cited with industry leaders?

Earn mentions on the pages that already name the leaders: roundups, comparison pages, review sites, podcasts and forum threads. Send editors accurate facts, invite real customer reviews, and take part in communities under your own name. Our co-citation SEO playbook has a source-by-source method and the outreach rules.

References

  1. 1.Co-citation in the scientific literature: A new measure of the relationship between two documents (Small, 1973), Journal of the American Society for Information Sciencedoi.org ↗
  2. 2.Bibliographic coupling between scientific papers (Kessler, 1963), American Documentationdoi.org ↗
  3. 3.A document co-citation analysis method for investigating emerging trends and new developments (Wei and Zhang, 2020), Information Researchinformationr.net ↗
  4. 4.US6457028B1: Method and apparatus for finding related collections of linked documents using co-citation analysis, Google Patentspatents.google.com ↗
  5. 5.Finding related pages in the World Wide Web (Dean and Henzinger, 1999), Computer Networksdoi.org ↗
  6. 6.Prediction: Anchor Text is Weakening and May Be Replaced by Co-Occurrence, Moz (2012)moz.com ↗
  7. 7.Not All Anchor Text is Equal and other Co-Citation Observations, SEO by the Sea (2012)seobythesea.com ↗
  8. 8.It's Not Co-Citation, but it's still awesome, iAcquire (2012), via the Internet Archiveweb.archive.org ↗
  9. 9.US8682892B1: Ranking search results, Google Patentspatents.google.com ↗
  10. 10.How Search works: ranking results, Googlegoogle.com ↗
  11. 11.Optimizing for generative AI features on Google Search, Google Search Centraldevelopers.google.com ↗
  12. 12.Impact of Co-occurrence on Factual Knowledge of Large Language Models (Kang and Choi, 2023), arXivarxiv.org ↗

Written by

Rankbox Team

The team behind Rankbox. We study how ChatGPT, Perplexity, Gemini, and Google AI Overviews choose their sources, and publish what we learn so you can put it to work.

Who we are and how we work
All articles
Keep exploring

Become an entity AI engines recognize

See where AI cites you today

Enter your site to see how often ChatGPT, Perplexity, Gemini, and Google cite your brand, and exactly what to publish next.

7-day free trial · Card required · Cancel anytime