Guide
Why Perplexity won't cite your site, even with great backlinks
A citation, in Perplexity’s output, is a numbered link the model attaches next to a specific claim it lifted from a specific page. Domain authority barely enters into it. Perplexity answers a query by retrieving a small set of pages at query time, reading each one for a passage that directly answers the question, and attaching a citation only where it actually used a passage. A page with two hundred backlinks and no direct answer on it loses to a three-week-old page with one clear paragraph, because the retrieval step never scores backlinks in the first place. It scores whether the page can be fetched, and whether the fetched text contains something worth quoting.
Why crawlability is the first gate, not the last one
A page that Perplexity cannot fetch cannot be cited, whatever else is true about it. PerplexityBot has to reach the page, render or read its text within a few seconds, and get a 200 response. Nothing else matters yet. Cloudflare reported in August 2025 that Perplexity was in some cases using undeclared crawlers to fetch pages that had explicitly blocked PerplexityBot in robots.txt, which cuts the other way for site owners: a page believed to be excluded may still be reachable, and a page believed to be open may in fact be blocked by an overzealous security rule or a CDN bot-fighting feature with no connection to robots.txt at all. Checking server logs for PerplexityBot requests over the past thirty days answers the question directly, rather than assuming from the robots.txt file alone.
Why backlinks do not directly cause a Perplexity citation
Backlinks are a Google ranking signal built for a different retrieval mechanism: PageRank scores a domain’s accumulated authority across the whole web, then serves ten blue links in a fixed order. Perplexity’s citation step runs per query, not per domain, so a page from a domain with no inbound links can still be cited if it happens to hold the clearest answer to that specific question. A domain’s overall authority correlates loosely with getting crawled in the first place, since a well-linked site tends to be indexed sooner, but that is an indirect effect on reach, not a direct scoring input on citation. Confusing the two leads a marketing team to spend a quarter chasing backlinks for AI visibility while the actual page never gets restructured into anything quotable.
What a citable page structure actually looks like
Take a company’s pricing page as the worked example. A version built for a human sales team buries the number in a “Contact us for a custom quote” form with the actual figure only visible after a call. A version built to be citable states the number in the first sentence under a heading phrased as the question a buyer would type, something like “How much does the starter plan cost”, followed immediately by the figure and the billing period. Perplexity’s retrieval step can lift the second version as a direct quote. The first version has nothing extractable to lift, no matter how much traffic or authority that page otherwise carries.
| Signal | Citable page | Uncited page |
|---|---|---|
| Answer location | First 1-2 sentences under the heading | Buried after several paragraphs or a form |
| Heading phrasing | Matches the buyer’s actual question | Generic label, e.g. “Pricing” |
| Load behaviour | Text present on first response | Text injected later by client-side JavaScript |
| Update signal | Visible last-updated date | No date, looks static or abandoned |
Why the same page gets cited by Perplexity and ignored by ChatGPT
Perplexity and ChatGPT search retrieve and score pages independently, using separate crawlers and separate ranking logic, so a page structured for one is not automatically structured for the other. A page that Perplexity cites today because it answered a specific comparison question may never surface in ChatGPT’s search results if OpenAI’s OAI-SearchBot indexed a different, less complete version of the same page, or has not recrawled it since a recent update. Checking what each engine actually says, rather than assuming a citation on one implies visibility on both, is the same discipline covered in Why does ChatGPT describe my company incorrectly?, where accuracy problems turn out to differ by model for the same underlying reason: each one is reading its own version of the page.
Common questions
Does joining the Perplexity Publishers’ Program guarantee citations? No. Perplexity launched the Publishers’ Program in 2024 with partners including TIME and Fortune, sharing ad revenue when a member’s page is cited, and in August 2025 it announced Comet Plus, a separate subscription with a $42.5 million pool paying publishers 80% of that subscription’s revenue. Both govern how money is split once a citation happens. Neither has any bearing on whether a page gets selected as a citation in the first place, which still depends on the content matching the query.
Do HTML tables actually help get a page cited? Yes, for comparison and pricing queries specifically. A table’s rows and columns give the retrieval step a compact, unambiguous chunk to quote, which is a large part of why comparison pages tend to be cited more often than narrative prose covering the same numbers.
How often does Perplexity recrawl a page after it changes? There is no published fixed interval, and behaviour varies by domain and by how often PerplexityBot has found the page changing in the past. A page with a visible, accurate last-updated date gives the crawler a reason to treat a refetch as worthwhile.
Sources
- Cloudflare (4 August 2025). Perplexity is using stealth, undeclared crawlers to evade website no-crawl directives.
- Bloomberg (25 August 2025). Perplexity to let publishers share in revenue from AI searches. Comet Plus, an 80/20 split and a $42.5 million pool.
Sourcive measures whether ChatGPT, Claude, Perplexity and Gemini mention your business when your buyers ask, and gives you the files to fix it. Run the free audit.