How do you get cited by Perplexity?

There is one structural difference from classic SEO worth internalizing first: Perplexity searches the web in real time for every query and cites the handful of sources its model actually used to write the answer. There is no permanent rank to win on Perplexity. That cuts both ways — you can lose a citation to a fresher page tomorrow, but a page published this week can be cited within days, not months.

Perplexity’s own documentation describes the process in four steps: it interprets the question, searches the internet in real time “gathering insights from top-tier sources,” compiles the most relevant material into an answer, and attaches numbered citations linking to the sources it used.

Which pages those are can be counted. In our page-types study, our classification of 569 links from 120 answers of four assistants, 14–15 September 2026 (Perplexity: 311 links, 24 answers, logged in with personalisation off), 23 of Perplexity’s 24 answers (96%) cited at least one vendor’s own page, and none of its 311 links pointed to platform documentation. Of those links, 32% went to agency and vendor articles, 19% to directories and comparison sites and 9% to forums. For a company, that makes its own service or product page the page Perplexity is most likely to cite, while articles and directories carry what others say about it. Nine links (3%) went to an unrelated “GEO”, such as GEO Group or graduate student unions: if your name or category is an acronym, define it on the page in the first sentence that uses it.

The practical consequences:

This works differently from ChatGPT, where citation runs through OpenAI’s separate search crawler and its own index — we covered that pipeline in How do you get cited by ChatGPT?

Advice about Perplexity often mixes official documentation, research findings and assumptions. Separate them before treating a claim as a ranking factor.

Two misreadings of the Ahrefs data circulate widely. First, the “12%” figure: only 12% of links cited by ChatGPT, Gemini, and Copilot rank in Google’s top 10 — but that aggregate does not include Perplexity, whose overlap is 28.6%, the highest of any answer engine. Second, the reverse mistake: treating a top-10 ranking as the mechanism. Perplexity leans on classic search more than its competitors, yet 71% of its citations still come from outside the top 10. Ranking well helps; it is neither sufficient nor strictly necessary.

Access is binary: if the crawler can’t fetch the page, nothing downstream matters. In robots.txt, allow the indexing crawler explicitly:

User-agent: PerplexityBot
Allow: /

Three checks beyond robots.txt:

Perplexity needs passages it can quote or paraphrase without substantial revision. Clear writing provides a practical starting point:

This article follows the same pattern: an answer at the start, focused paragraphs, sourced figures and dated studies. Those choices give retrieval systems passages they can use and readers claims they can check.

The third bar is the least mechanical: Perplexity describes its sources as “top-tier,” and what makes a source top-tier is mostly established off your page.