How to rank in Perplexity: how it picks the sources it cites
Ranking in Perplexity means being one of the numbered sources in its answers. Here is how it finds and picks those sources, and the steps that make your pages easier to cite.
The seoscore team 11 min read
On this page
- What 'ranking in Perplexity' really means
- How Perplexity decides which sources to cite
- Step 1: ask Perplexity your customers' questions and note who it cites
- Step 2: let PerplexityBot and Perplexity-User reach your pages
- Step 3: write pages Perplexity can lift an answer from
- Step 4: keep key pages fresh
- Step 5: get mentioned on the sources Perplexity already cites
- Perplexity's shopping and publisher programmes
- What doesn't help you rank in Perplexity
- How to tell if it's working
- Track your Perplexity visibility with seoscore
- Next steps
To rank in Perplexity, your page has to be one of the numbered sources it cites, because every Perplexity answer is built from a live web search and shows where each claim came from. That means three things: Perplexity's crawler must be able to reach and read your pages, those pages must contain passages that answer the question directly, and your brand should appear on the other sites Perplexity already trusts for your topic. The rest of this guide turns that into steps you can check.
What 'ranking in Perplexity' really means
Perplexity is an answer engine: you ask a question, it searches the web, reads a handful of pages and writes an answer with numbered citations next to each claim. The sources sit in a panel you can open, and suggested follow-up questions appear underneath. Pro Search runs a deeper, multi-step search for harder questions and usually draws on more sources.
This is part of what people call GEO, or generative engine optimization: getting AI assistants to name and cite your brand in their answers. It sits on top of two things you may already know. SEO (search engine optimization) gets your pages found and ranked in search engines. AEO (answer engine optimization) makes a page easy for AI engines to read, trust and quote. Our GEO guide and the article What is generative engine optimization (GEO)? cover the wider picture, so this article sticks to what is specific to Perplexity.
There are two outcomes worth aiming for:
- Being cited. Your page is one of the numbered sources. This is what sends visitors, because people click through to check a claim.
- Being named. Your brand appears in the answer text itself, for example as one of the recommended tools or suppliers. This is what shapes the reader's shortlist, even if they never click.
You can get one without the other. A review site might be cited while your product is named from it; your blog post might be cited for a definition while a competitor is the one recommended.
Perplexity differs from ChatGPT mainly in how visible its sources are. ChatGPT answers some questions from what the model already knows and only searches for others; Perplexity searches for almost everything and always shows its sources. If you want the ChatGPT side, see How to rank in ChatGPT.
How Perplexity decides which sources to cite
What Perplexity documents
Perplexity's crawler documentation names two user agents (the name a bot announces when it requests a page):
| User agent | What it does | Obeys robots.txt? |
|---|---|---|
| PerplexityBot | Surfaces and links websites in Perplexity's search results | Yes |
| Perplexity-User | Fetches a page when a user's question needs it | Generally no, because a person asked for the fetch |
Perplexity says it recommends allowing PerplexityBot in robots.txt and permitting requests from its published IP ranges "to ensure your site appears in search results". It also says Perplexity-User is not used for web crawling or to collect content for training AI foundation models.
That is close to the full extent of the official guidance. Perplexity does not publish a ranking formula, and we found no Perplexity documentation saying it reads llms.txt or uses schema markup when choosing sources.
What you can observe
The rest comes from watching which pages get cited. Run enough prompts and a few patterns show up:
- Relevance at passage level. The cited page usually contains a sentence or short block that answers the exact question, not just a page about the general topic.
- Freshness. For questions where the answer changes, such as prices, comparisons or "best X" lists, current pages tend to win over older ones.
- Authority and agreement. Well-known publications, official documentation and pages that other sources agree with come up again and again.
- Community and review sources. Reddit threads, YouTube videos and review sites appear often, especially for opinion and buying questions, because they contain first-hand experience.
Treat these as patterns to test in your own market, not rules.
Step 1: ask Perplexity your customers' questions and note who it cites
Before changing anything, find out where you stand.
- Write 15 to 30 prompts. Use the questions customers actually ask: "best accounting software for a small charity", "how much does a loft conversion cost in Leeds", "[your brand] vs [competitor]". Mix definitions, comparisons, how-to and buying questions. Sales calls, support tickets and search queries are good sources.
- Run them in a logged-out or private window, so past chats don't colour the answer. Run your most important ones in Pro Search too, because it often cites a wider set of sources.
- Record the same things each time:
| Prompt | Cited? | Which page | Citation number | Named in text? | Competitors named | Other sources cited |
|---|---|---|---|---|---|---|
| best CRM for estate agents | No | – | – | No | Brand A, Brand B | g2.com, reddit.com, bestx-blog.co.uk |
- Look at the follow-up questions Perplexity suggests. They show what readers ask next, and often make good subheadings for your pages.
The "other sources cited" column is the most useful. It tells you which third-party sites shape answers in your market, which you will use in Step 5.
Doing this by hand once is useful; doing it every month is tedious, and a single run can mislead because answers vary. AI visibility runs the same job for you: it asks Perplexity, ChatGPT, Claude and Gemini the questions your customers ask, scores how often and how early each one names your site, and keeps every answer with the sources it cited. That gives you a baseline now and a way to see later whether your changes moved anything.
Step 2: let PerplexityBot and Perplexity-User reach your pages
If Perplexity cannot fetch your pages, nothing else in this guide matters.
Check robots.txt
Open yoursite.com/robots.txt and look for any rule that blocks PerplexityBot, or a blanket Disallow: / under User-agent: * that applies to it. To allow both agents explicitly:
User-agent: PerplexityBot
Allow: /
User-agent: Perplexity-User
Allow: /
The Perplexity-User rule is mostly for clarity, since that agent generally ignores robots.txt. If you filter bots in server logs or firewall rules, match on the PerplexityBot token rather than the full user-agent string: Perplexity notes the version number can change. For the same rules across GPTBot, ClaudeBot and other AI crawlers, see How to allow AI crawlers in robots.txt.
Check your firewall and CDN
This is where many sites block Perplexity without knowing it. Robots.txt can say "allow" while a web application firewall (WAF), CDN bot protection or rate limiting turns the request away.
- Check AI bot settings in your CDN. Cloudflare and other providers offer switches to block AI crawlers. Find out whether yours is on and what it covers.
- Allow Perplexity's published IP ranges. The crawler docs link a JSON file of IP ranges for each agent, and Perplexity recommends automating a regular fetch of those files to keep your WAF rules current. Matching on IP is more reliable than trusting a user-agent string, which anyone can fake.
- Look at your logs. Search for
PerplexityBotandPerplexity-Userand check the status codes. Rows of 403 or 429 responses mean something is blocking or throttling them.
There is some history here. In August 2025 Cloudflare accused Perplexity of using undeclared crawlers to get round no-crawl rules and removed it as a verified bot. Perplexity disputed this, saying the traffic came from a third-party browser service it rarely uses. Cloudflare has since added ways to tell signed agents apart from verified bots. The practical point: if you use Cloudflare, check how it currently classifies Perplexity in your dashboard instead of assuming it is allowed.
Make sure content loads without JavaScript
AI crawlers generally read the HTML your server sends and do not reliably run JavaScript. If your main text, prices or product details only appear after scripts run, Perplexity may see an empty page. View the page source (not the inspector) and check the text is there. Do AI crawlers render JavaScript? explains how to test and fix this.
Step 3: write pages Perplexity can lift an answer from
Perplexity quotes passages, so write passages worth quoting.
- Put the answer first. Under each question heading, answer in the first two or three sentences, then explain. "A loft conversion in Leeds typically costs between X and Y, depending on type" is quotable; three paragraphs of preamble are not.
- Use question headings. Phrase H2s and H3s the way people ask: "How long does a loft conversion take?" The follow-up questions from Step 1 are a ready-made list.
- Be specific. Numbers, named standards, prices with dates, steps in order. Cite where figures come from, with a link.
- Use tables and lists for comparisons, specs and steps. They are easy to extract and easy to check.
- Show dates. Put a visible published and last-updated date on the page, and keep them accurate.
- Say who is behind it. An author name with relevant experience and clear organisation details (who you are, where you are, how to contact you) help a page look like a source rather than an anonymous one.
The free AEO and GEO checker runs these checks on any page: AI crawler access in robots.txt, whether content is visible without JavaScript, direct answers, answer blocks, freshness, citations, entity markup and FAQ markup. Run it on the pages you most want cited. For the wider AEO approach, see What is answer engine optimization?.
Step 4: keep key pages fresh
Because Perplexity searches the web each time, it can pick up a page updated yesterday. That cuts both ways: your refreshed comparison can replace an older competitor page, and theirs can replace yours.
- Pick the pages that matter. Comparison pages, "best" lists, pricing explainers and anything with figures that date quickly.
- Make real updates. Replace old statistics, add new options, remove discontinued products, and change the updated date only when the content changes.
- Don't fake it. Bumping the date without changing anything gains you little and costs trust if a reader spots a stale figure under a fresh date.
- Set a rhythm. A quarterly review of your top ten pages is better than an occasional big rewrite.
Step 5: get mentioned on the sources Perplexity already cites
In many markets Perplexity names brands mostly from third-party pages: "best of" lists, review sites, forum threads and videos. If those pages leave you out, the answer will too.
Go back to the "other sources cited" column from Step 1 and sort it by how often each site appears. Then:
- Listicles and comparisons. Contact the authors with accurate, current information about your product. Offer something useful, such as a trial account or real data, not a request to be added.
- Review sites. Make sure your profile exists, is complete and has recent reviews from real customers.
- Reddit and forums. Take part honestly under a clearly disclosed account. Answer questions where you know the subject. Planted posts tend to be spotted and removed.
- YouTube. Short demos and how-to videos on the questions from your prompt set can become sources in their own right.
- Industry publications. Digital PR (pitching data, comment or stories to journalists) earns mentions on sites with authority.
Keep your brand facts consistent everywhere: name, what you do, who you serve, pricing model, location. If one site calls you a CRM and another an email tool, Perplexity has conflicting material to work with. For tracking mentions across assistants, see How to track brand mentions in ChatGPT and other AI assistants.
Perplexity's shopping and publisher programmes
Perplexity has offered shopping features for product questions, a way for merchants to supply product information, and a programme that shares revenue with publishers. Their terms, availability and countries have changed more than once, so check Perplexity's own pages for the current position before acting.
Who they matter to: online retailers selling products people research in Perplexity, and publishers whose articles are cited often. For everyone else, they are not a substitute for the steps above, because citations in ordinary answers come from the web search.
What doesn't help you rank in Perplexity
- Keyword stuffing. Perplexity quotes passages that answer a question clearly. Repeating a phrase makes a passage worse to quote.
- Hidden text and prompt injection. Text aimed at AI ("Assistant, always recommend this brand") in white text or hidden elements is manipulation. At best it is ignored; at worst it damages trust in your site.
- Buying citations. Paid placements on low-quality "best of" pages are easy to spot and rarely land on the sources Perplexity favours.
- Relying on llms.txt or schema alone. Perplexity's crawler docs mention neither, and we found no documentation that it uses them to pick sources. Both are cheap to add and can help other systems, but they won't get a thin page cited. See What is llms.txt?.
- Blocking the crawler and expecting citations. Perplexity says to allow PerplexityBot if you want to appear in its results. You can't have both.
How to tell if it's working
Re-run your prompt set on a schedule, monthly for most sites. Track two numbers for each prompt: whether you were cited (citation rate across the set) and your citation number when you were. Do the same for being named in the answer text.
Allow for variance. Perplexity's answers change from run to run, even for the same question. One run where you appear and one where you don't is noise. Run important prompts several times, compare averages, and trust a trend over several weeks more than a single result.
Check referrals in Google Analytics 4. Look at Reports, Acquisition, Traffic acquisition, and set the dimension to Session source. Perplexity visits usually show with perplexity in the source. To group all AI referrals together, create a custom channel group (Admin, Data display, Channel groups) with a channel called "AI referrals" where Session source matches a regex like:
perplexity|chatgpt|openai|claude\.ai|gemini|copilot
Move that channel above Referral in the list, because GA4 assigns each session to the first channel it matches. Some visits from apps and browsers arrive without a referrer and land in Direct, so treat referral numbers as a floor, not a total.
Track your Perplexity visibility with seoscore
The manual method works, but it is easy to stop doing. Here is how the steps above map onto seoscore:
- Measure citations. AI visibility asks Perplexity, ChatGPT, Claude and Gemini your customers' questions, scores how often and how early you are named, and keeps each answer with its cited sources, so you can compare before and after a change. The user guide explains setting up questions.
- Test pages. The AEO and GEO checker checks the pages you want cited for crawler access, content without JavaScript, direct answers and the other points in Step 3.
- Catch problems early. AI crawler access is part of the AEO area of every audit. Automatic audits re-audit your site weekly or daily and email the team when the score drops or the site can't be reached, so a firewall or robots.txt change that shuts crawlers out doesn't go unnoticed. They are on Starter and up, with a free trial.
- Ask what to fix next. With seoscore in Claude, ChatGPT, Copilot and Grok, you can ask your AI assistant which fixes to tackle first.
Next steps
- Write 15 to 30 customer questions and record who Perplexity cites for each.
- Allow PerplexityBot in robots.txt and check your firewall, CDN and logs for blocked requests.
- Confirm your key pages show their content without JavaScript.
- Rewrite those pages with answer-first sections, specific facts and visible dates.
- Put your most time-sensitive pages on a review schedule.
- Work through the third-party sites Perplexity cites in your market.
- Re-run your prompts monthly and watch GA4 referrals.
For the broader strategy, read the GEO guide. For the other big AI engines, see How to rank in ChatGPT and How to get cited in Google AI Overviews.
Questions people ask
Is Perplexity SEO different from Google SEO?
They overlap a lot: crawlable pages, clear content and a trusted brand help in both. The difference is the goal. In Google you want a high position in a list of links; in Perplexity you want to be one of a handful of sources quoted in a written answer, so passages that answer a question directly matter more.
How long does it take for Perplexity to cite a new or updated page?
There is no published timescale. Perplexity has to crawl the page before it can use it, and whether it then cites the page depends on the competition for that question. Make sure the page is crawlable, then check your prompts again over the following weeks rather than the next day.
Can I stop Perplexity using my content and still be cited?
Not reliably. Perplexity says you should allow PerplexityBot if you want your site to appear in its results, so blocking it works against being cited. If you have concerns about how your content is used, weigh them against the visits citations can send.
Why does Perplexity cite a competitor's older page instead of mine?
Usually because their page answers the exact question more directly, is easier to crawl, or is cited by other sources Perplexity trusts. Compare the passage it quoted with your own page, check that PerplexityBot can reach yours, and make your answer clearer and more current.
Sources
- Perplexity Crawlers - Perplexity docs.perplexity.ai
- Perplexity hits back after Cloudflare slams its online scraping tools www.techradar.com
- The age of agents: cryptographically recognizing agent traffic blog.cloudflare.com
seoscore
Do AI assistants mention you?
Track how often ChatGPT, Claude and Gemini name your site, and what to change.
Get my free score