How-tos · November 12, 2026 · 11 min read
Perplexity SEO: What Perplexity Cites and How to Get Cited
Perplexity searches the web for almost every question and shows numbered sources. What our own sample shows about the sources it cites, and a playbook to become one of them.
Perplexity SEO is the work of getting your brand into Perplexity’s answers. Perplexity searches the web for almost every question and shows numbered sources with its answer. So the work has two parts: be one of the sources it retrieves, and be named in those sources. There is no rank to buy and no published formula to game.
This post explains how Perplexity answers, what our own sample showed about the sources it cites, and a playbook you can run by hand. About the numbers: they come from Perplexity’s API (the sonar model), 321 answers to 46 of our prompts, from 23 August to 4 September 2026. That is part of a larger sample of 6,069 answers from seven engines to 48 prompts, from 23 August to 1 October 2026. Every answer is about brands we track in our own workspace. About 92% of the runs are in one developer software niche, and 78% are for one brand. Results in your category will differ.
How Perplexity answers
Perplexity is built around search. When it searches, it rewrites the question into one or more web searches (this is query fan-out), reads the pages it retrieves, and writes an answer. Numbered markers in the text point to the sources, and the full source list sits with the answer. The Perplexity app often shows the searches it ran, which is the easiest place to see fan-out happen.
Two things follow. The sources change with the web: a new or updated page can appear once Perplexity can find and fetch it, without waiting for a new model. And you can see what it trusts, because every answer lists its sources.
Perplexity’s public documentation does not publish a ranking formula, so nobody outside Perplexity can state the rules. Lists of “Perplexity ranking factors” are guesses from samples, like ours. Treat any claimed rule as a hypothesis to test.
What our sample showed
We read Perplexity through its API and as the answer the app shows. The app data is too thin to quote (84 usable answers, for one brand), so every Perplexity figure below is the API. The API is not the same thing as perplexity.ai.
About ten sources per answer, on every answer
| Engine, through the API | Answers | Answers with citations | Mean citations | Median citations | Median distinct domains |
|---|---|---|---|---|---|
| Perplexity | 321 | 100.0% | 11.46 | 10 | 9 |
| ChatGPT | 321 | 94.7% | 6.08 | 6 | 4 |
In this sample, every Perplexity answer had citations, with a median of 10 links from 9 distinct domains. ChatGPT through its API cited a median of 6 links from 4 domains.
What to do: about ten source slots per answer leaves room for more than the market leader. Aim for one slot on the questions that matter, and check that the count holds for your own questions.
Reddit in two thirds of answers
Share of answers that cite the platform at least once, through each API:
| Platform | Perplexity, API | ChatGPT, API |
|---|---|---|
| reddit.com | 67.9% | 4.0% |
| linkedin.com | 13.4% | 0.3% |
| g2.com | 7.2% | 0.9% |
| youtube.com | 3.7% | 0.0% |
In this sample, Perplexity cited reddit.com in 67.9% of its answers, and Reddit was 11.2% of all its links. Social and community sites together were 13.9% of its links. Most of the sample is in developer software, where community threads are common. A category with fewer active communities may show less.
What to do: read the community threads that Perplexity cites for your questions. Note which ones describe your category, whether they name you, and whether what they say about you is still true.
Mostly third-party sources
We sort every cited domain into three groups: the tracked brand’s own site, a competitor site we configured, and every other site (third parties).
| Perplexity, API | Own site | Competitor sites | Third parties |
|---|---|---|---|
| All seven brands | 5.8% | 3.2% | 91.0% |
| The three brands with a competitor list | 7.1% | 5.2% | 87.6% |
In this sample, about nine in ten Perplexity links went to third parties: blogs, list-style posts, directories, communities and other vendors. Own sites still appeared often, as one source among many. On the 43 prompts both APIs answered, the most cited tracked brand site appeared in 48.9% of Perplexity answers and 19.6% of ChatGPT answers. We wrote the prompts about the tracked brands’ niche, so own-site figures run higher than they would for neutral questions.
What to do: most of the answer is built from other people’s pages. Plan more of your effort for those pages than for your own site.
Spread across many domains
| Perplexity, API | Citations | Distinct domains | Top domain | Top 10 domains | Top 50 domains |
|---|---|---|---|---|---|
| All answers | 3,680 | 360 | 11.2% | 38.7% | 66.0% |
Perplexity’s top 10 domains took 38.7% of its links, against 46.3% for ChatGPT’s API over the same weeks. Of all the routes we read, Perplexity’s API was the least concentrated on its top 50 domains. A small prompt set asked daily inflates every concentration figure.
One caveat on the list itself: github.com was cited in 47.4% of Perplexity answers. Across the whole sample, 67.3% of github.com links have a tracked brand’s or competitor’s name in the URL. Most link to the tracked brands’ or their competitors’ own GitHub pages. That is an artifact of the sample’s niche.
What to do: a long tail means smaller sites can get in. Look for niche directories and comparison posts on modest blogs in your own citation list.
Little overlap with ChatGPT
For the same 43 prompts in the same weeks, Perplexity’s API and ChatGPT’s API shared about 1 cited domain in 9 (mean Jaccard overlap 0.114). When we compared one answer from each on the same prompt and day, 26.4% of the 575 pairs shared no domain at all. We show the full comparison in which sources ChatGPT cites.
What to do: build a separate citation list for each engine your buyers use. Our ChatGPT SEO guide covers the ChatGPT side.
The Perplexity SEO playbook
1. Find the questions
Write 15 to 25 questions in the words your buyers use: “best [category] for [audience]”, “alternatives to [leader]”, “[rival] vs [rival]”, “how do I [solve the problem]”. Leave your brand name out of most of them. Sales calls and support tickets give you the real wording.
2. Read the cited sources for each question
Ask each question in Perplexity several times, on different days. Record the numbered sources and the searches the app shows. Group the sources: communities, comparison posts, directories, review sites, vendor sites, your own site. Count how often each domain appears. That count is your to-do list.
3. Earn a place in community threads, honestly
If community threads appear in your list, take part as yourself. State your affiliation every time. Follow each community’s rules on self-promotion, which are often strict. Answer the question that was asked, and mention your product only where it is a real answer. Never post from fake accounts.
4. Get into third-party comparisons and directories
For each comparison post or directory that Perplexity cites, check whether you appear and whether the facts are right. Send authors something useful: a correction, a feature they missed, an account to test. Complete your directory listings. Do not pay for placement you do not disclose.
5. Publish answer-first pages
We cannot see how Perplexity picks a passage, but a page that answers the question in plain text near the top is easier to quote. Give it that passage:
- Put a complete answer in the first paragraph.
- Answer one question per page. Use the buyer’s words in the title and headings.
- Put facts in text, not images: prices, plans, limits, integrations, who the product is for.
- Show dates: when the page was published and when it was last updated.
- Publish honest comparison pages that name rivals and say where they are better.
- Add a FAQ with the real questions, each answered in one paragraph.
6. Keep pages fresh
An out-of-date page gives an out-of-date answer. Update prices, plan names and screenshots when they change. Change the “last updated” date only when the content changes.
7. Make the site crawlable
Perplexity can cite only pages it can fetch and read. Serve the main content as HTML, not only through browser scripts. Keep answers out from behind logins and forms. Keep pages fast. Publish a sitemap.
8. Do not block the crawlers you want to be read by
Check your robots.txt, your firewall and any bot-protection settings at your host or CDN. A rule that blocks all unknown bots can block the crawlers you want. Perplexity’s documentation lists its current crawlers; check it before you change a rule, and test that your pages still load for them.
What does not work
- Fake community activity. Sock-puppet accounts, bought upvotes and planted threads break community rules and get removed.
- Fake or bought reviews. They hurt you with buyers once found.
- Keyword stuffing and hidden text. Text aimed at AI readers and not at people makes a page worse to quote.
- Asking Perplexity about your brand again and again. Your questions do not change what it retrieves for others.
- Judging from one screenshot. Answers vary from ask to ask.
- Blocking crawlers and expecting citations. If Perplexity cannot fetch the page, it cannot cite it.
How to measure it
Keep a fixed prompt set and sample it on a schedule. Read three numbers per engine, over many answers:
- Visibility rate: the share of answers that name your brand.
- Average rank: your position among the tracked brands named, when you are named. A position in a single answer means little, because the lists vary between runs, so rank is useful only as an average over many answers and next to visibility rate.
- Citation share: your domain’s share of all the links in the answers.
Count named and cited separately. Wait for at least 10 answers per prompt before you trust a percentage, and do not change the prompt set in the same week you change the site.
Perplexity brand monitoring in CitePrism samples Perplexity on a schedule, through the API and as the app shows the answer. AI citation tracking groups every cited link by domain and category, and the Queries page shows the searches an engine reports running, where the engine exposes them. Metrics grey out below 10 answers, and every chart clicks through to the stored raw answers.
What this sample can and cannot tell you
Our Perplexity figures come from 321 API answers to 46 prompts, 23 August to 4 September 2026, inside a sample of 6,069 answers to 48 prompts about brands we track in our own workspace, mostly in one developer software niche. Each prompt repeats daily, so the effective sample is the prompts, not the answers. The data shows large contrasts inside this sample, such as Perplexity against ChatGPT on Reddit. It cannot tell you how Perplexity picks sources in general, how perplexity.ai behaves for real users, or what your category looks like.
FAQ
How does Perplexity choose its sources?
Perplexity searches the web for almost every question, reads the pages it retrieves, and cites the ones it uses in its answer. Its public documentation does not publish a ranking formula, so the selection rules are not known outside Perplexity. What you can see is the result: the numbered sources on each answer. Sample your buyer questions and count which domains appear.
How do I rank on Perplexity?
There is no rank to climb. You aim to be one of the sources Perplexity cites, and to be named in the other sources it cites. Find your buyer questions, read the sources Perplexity cites for each, earn an honest place in those communities, comparisons and directories, and publish answer-first pages it can fetch.
Does Perplexity cite Reddit a lot?
In our sample, yes: Perplexity’s API cited reddit.com in 67.9% of answers, against 4.0% for ChatGPT’s API on the same questions. Our sample is one developer software niche, where community threads are common. Check your own questions before you plan a community effort.
How long does it take to get cited by Perplexity?
Perplexity reads pages when it answers, so a new or updated page can appear soon after Perplexity can fetch it. We have no data on how long this takes. Measure weekly, and expect changes on third-party pages to take longer than changes on your own pages.
Is Perplexity SEO different from ChatGPT SEO?
The method is the same: find the questions, read the cited sources, improve your presence there and on your own site. The sources differ. In our sample, the two engines shared about 1 cited domain in 9 for the same questions, and Perplexity leaned much more on community threads. Build a separate source list for each engine.