A lot of marketers treat Perplexity as a completely separate world from Google. They assume it needs its own special "GEO" playbook, separate from normal SEO. When I look at which sites Perplexity actually cites, I see something different: Perplexity behaves more like Google than any other AI assistant I have data on, with one big twist. It loves Reddit and YouTube.
That is useful to know. Perplexity shows numbered sources under its answers, so a citation there is a visible, clickable link to your page. If you know which sites it trusts, you know where your brand needs to appear and how much of your normal SEO work already carries over.
In this article I go through the latest Ahrefs ranking of the most-cited websites in Perplexity, explain what it measures, and show how I would turn it into a mention plan. This is a source analysis. If you want the step-by-step on getting your own pages cited, read my guide to Perplexity SEO.
How Ahrefs counted Perplexity's sources
Ahrefs used Brand Radar, its tool for recording what AI assistants say and cite. For Perplexity, the September 2026 snapshot covers more than 3.1 million US queries across all topics. Ahrefs counted how often each domain appeared as a source and ranked them by mention share, which is a domain's citations as a share of the combined citations of the top 50 sources.
Ahrefs explains why this list is especially telling for Perplexity:
“Perplexity built its whole reputation on citing its sources, so the sites it leans on tell you a lot about how it answers.”
— Ahrefs, The 50 Most-Cited Websites in Perplexity
Perplexity says the same thing about itself in its help content:
“Every answer includes footnotes linking to the original sources.”
— Perplexity, How does Perplexity work?
| Detail | What we know |
|---|---|
| Data source | Ahrefs Brand Radar |
| Sample | Over 3.1 million US queries, all topics |
| Snapshot | September 2026, refreshed monthly by Ahrefs' AI agent |
| Main metric | Mention share of the top 50 sources |
| Not stated | How queries were chosen, exact dates, and whether data came from the web app or the API |
Reddit and YouTube take 42%: Perplexity's leaders
Here are the first twelve domains, with the monthly rank change Ahrefs reports. The source-type column is my own label.
| Rank | Domain | Mention share | Monthly move | Source type (my label) |
|---|---|---|---|---|
| 1 | reddit.com | 21.6% | Up 1 | Community / forums |
| 2 | youtube.com | 20.8% | Down 1 | Video |
| 3 | en.wikipedia.org | 6.3% | No change | Encyclopedia |
| 4 | facebook.com | 4.5% | No change | Social network |
| 5 | amazon.com | 3.6% | Up 1 | Marketplace |
| 6 | healthline.com | 2.6% | Up 3 | Health publisher |
| 7 | alibaba.com | 2.0% | Down 2 | B2B marketplace |
| 8 | oreateai.com | 1.8% | Up 4 | Unclear |
| 9 | goodhousekeeping.com | 1.7% | Up 11 | Product testing magazine |
| 10 | nytimes.com | 1.6% | Up 28 | Newspaper |
| 11 | tripadvisor.com | 1.5% | Up 6 | Travel reviews |
| 12 | consumerreports.org | 1.3% | Up 7 | Independent testing |
Source: Ahrefs, September 2026 snapshot, US queries. Labels are mine.
Reddit and YouTube together hold 42.4 points of mention share, by my addition. Wikipedia is a distant third. After that, the list is a mix of marketplaces, health publishers, testing magazines and a newspaper.
One difference from Google's AI features is worth noticing: Perplexity's list has no google.com and much less Facebook and Instagram. Instagram is down at fifteenth with 1.2%.
What Amazon, Alibaba and Target say about shopping questions
Three retail sites sit in Perplexity's top 15: Amazon at 3.6%, Alibaba at 2.0% and Target at 1.3%. Alibaba is the interesting one. It is mainly a business-to-business sourcing marketplace, and Ahrefs' top 50 even includes a separate Alibaba subdomain for car interior products. My reading is that Perplexity gets a fair number of sourcing and product-specification questions, and Alibaba has a matching page for almost any part or material.
If you are a manufacturer or wholesaler, that is worth noting. Your listings on sourcing marketplaces may be read by Perplexity more often than your own catalog pages. Keep specifications, materials, minimum order details and certifications accurate everywhere they appear, because Perplexity will quote whichever version it finds.
Why Perplexity's sources look so much like Google's results
This is the finding I think most marketers miss. In a September 2025 Ahrefs study, Louise Linehan compared the sources ChatGPT and Perplexity cited for about 3,311 head terms with Google's results for the same terms. About 65% of Perplexity's cited URLs also appeared in Google's top 10, and domain overlap was about 81%. For ChatGPT, only about 10% of cited URLs were in Google's top 10.
“Short-tail queries show closer SERP-AI alignment than natural language prompts—especially when it comes to Perplexity.”
— Ahrefs, ChatGPT Doesn’t Follow Google’s Sources, Perplexity Does
Let me explain what this means for you in plain words. If you already rank in Google's top 10 for an important short query, there is a good chance Perplexity will consider your page too. Your classic SEO work, such as clear pages, good internal links and real backlinks, carries over to Perplexity more directly than to ChatGPT.
The note about natural language matters, though. The overlap is strongest for short queries. For long, conversational questions, Perplexity's sources drift further from Google's top results. That is where community threads, videos and detailed reviews pick up more of the citations.
Here is a quick way to use this overlap. Take your 20 most important short queries and check where you rank in Google:
- Top 10 already: your page is a realistic Perplexity source. Make sure the answer to the query sits near the top of the page in plain words.
- Positions 11 to 20: these are your best bets. A push into the top 10 may help both Google and Perplexity at once.
- Not ranking at all: for these, earning a mention on a page that already ranks, such as a review site or a Reddit thread, is often faster than ranking yourself.
If you want to understand why each assistant ends up with a different source mix, read my piece on why each AI assistant cites different sources.
The rise of review magazines and newsrooms in Perplexity's list
The September 2026 update showed a clear move toward edited, tested content. Look at the climbers:
- New York Times: up 28 places into the top ten.
- Good Housekeeping: up 11 places to ninth.
- WebMD: up 10 places to thirteenth.
- Consumer Reports: up 7 places to twelfth.
- Tripadvisor: up 6 places to eleventh.
Further down the top 50 there are more review and testing brands, such as RTINGS, PCMag, CNET, TechRadar and Trustpilot. I don't want to claim a permanent trend from one month. But it fits what Perplexity is designed for: people asking for comparisons and "which one should I buy" answers, and wanting sources they can check.
For a business, this points to a very traditional kind of work: getting your product tested and reviewed by credible, independent publications. That is digital PR, not a technical trick. I cover how to approach it in digital PR for AI visibility, and why reviews matter so much in how review data shapes AI recommendations.
What PerplexityBot and Perplexity-User mean for being cited
Perplexity uses two different agents to read websites, and they behave differently. Its crawler documentation explains both.
| Agent | What it does | Follows robots.txt? | What I suggest |
|---|---|---|---|
| PerplexityBot | Crawls to surface and link websites in Perplexity's search results; not used for training foundation models | Yes | Allow it if you want to be cited |
| Perplexity-User | Fetches a page when a user's question needs it, then links to that page in the answer | Generally no | Make sure your firewall doesn't block it by accident |
“Since a user requested the fetch, this fetcher generally ignores robots.txt rules.”
— Perplexity, PerplexityBot and Perplexity-User crawler documentation
Two practical points come out of this. First, if you blocked "AI bots" as a group some time ago, check that PerplexityBot isn't in that block, or you may be removing yourself from the index Perplexity cites from. Second, Perplexity publishes IP ranges for both agents, so your security team can allow the real ones and block fakes. I explain how in my guide to verifying AI crawler requests.
Sorting Perplexity's sources by how you can earn a place
Here is how I group the source types in Perplexity's list, from easiest to hardest to influence.
| Source type | Examples | How you earn a place | Difficulty |
|---|---|---|---|
| Your own Google-ranking pages | Your site | Classic SEO for short head terms, since Perplexity tracks Google closely | Medium |
| Video | YouTube | Clear review, comparison and how-to videos with accurate titles and captions | Low to medium |
| Marketplaces | Amazon, Alibaba, Target | Complete, honest listings with real reviews | Low if you sell there |
| Review and travel platforms | Tripadvisor, Trustpilot | Ask real customers for reviews; respond to them | Low to medium |
| Communities | Honest participation; never fake accounts | Medium | |
| Testing magazines | Good Housekeeping, Consumer Reports | Offer products for testing; you can't buy a verdict | High |
| News and health publishers | New York Times, Healthline, WebMD | Expert commentary, original data, credible PR | High |
| Encyclopedias | Wikipedia | Only through genuine notability | Very high |
For a broader version of this triage across every assistant, see which AI citation sources you can actually influence.
Perplexity compared with ChatGPT, Gemini and Google's AI features
Here is how Perplexity's top sources compare with other assistants, using the matching Ahrefs lists from the same September 2026 update.
| Assistant | YouTube | Wikipedia | Reddit + YouTube | |
|---|---|---|---|---|
| Perplexity | 21.6% | 20.8% | 6.3% | 42.4% |
| Gemini | 28.5% | 14.6% | 8.8% | 43.1% |
| Google AI Overviews | 18.5% | 22.9% | 4.0% | 41.4% |
| Google AI Mode | 17.9% | 17.8% | 4.0% | 35.7% |
| ChatGPT | 16.8% | 2.5% | 7.0% | 19.3% |
Sources: Ahrefs lists for Perplexity, Gemini, AI Overviews, AI Mode and ChatGPT. Combined column is my addition.
Perplexity sits with Gemini and Google's AI features in leaning hard on Reddit and YouTube. ChatGPT is the clear outlier. You can read the details for each in my breakdowns of the sites ChatGPT cites most, the sites AI Overviews cite most, the sites AI Mode cites most and the sites Gemini cites most.
One small correction to the Ahrefs page itself. It describes Reddit's 21.6% in Perplexity as the most concentrated top source of any assistant Ahrefs looked at. But Ahrefs' own Gemini list, from the same update, shows Reddit at 28.5%. I would not rely on that particular claim.
A worked example: a boutique hotel builds its Perplexity source list
Here is how I would use this data for a hypothetical 40-room boutique hotel in a popular US city.
- List 25 questions travelers ask. For example: "best boutique hotels near the old town," "quiet hotel with parking downtown," "pet-friendly hotels with a rooftop bar."
- Run each in Perplexity and log the sources. Suppose the 25 answers cite 140 sources: Tripadvisor 26, Reddit 24, YouTube 18, two local newspapers 15, a travel magazine 9, the hotel's own site 3.
- Check Google for the short versions of the same queries, such as "boutique hotel [city]." Since Perplexity tracks Google closely for short terms, note where your site ranks in the top 10 and where it doesn't.
- Pick the influenceable sources. Tripadvisor (collect more recent reviews and reply to them), YouTube (a short room tour and neighborhood guide), local press (pitch a story about the hotel's restored building).
- Fix the robots.txt and firewall basics. Confirm PerplexityBot is allowed and Perplexity-User isn't blocked.
- Re-run the 25 questions monthly. If the hotel's own site and its Tripadvisor page together go from 8 of 140 citations (5.7%) to 20 of 140 (14.3%) in two quarters, that is a real gain.
Caveats before you act on Perplexity's list
- US only. The data covers US queries; other markets may look very different.
- Relative share. Percentages are shares of the top 50, not of all citations.
- Unknown sites. oreateai.com is eighth, and Ahrefs doesn't explain what it is.
- Big monthly swings. A 28-place jump for the New York Times shows how much the middle of the list can move.
- Method gaps. We don't know how queries were selected or whether results came from the app or the API.
- Dated usage figures. Ahrefs cites "hundreds of millions of queries a month" from a TechCrunch report published in October 2024, so treat that as old context.
If you publish news or content at scale, Perplexity also runs programs for publishers. I cover what they offer in my article on Perplexity's publisher programs.
Questions people ask about the sites Perplexity cites
Which website does Perplexity cite most?
Reddit. In Ahrefs' September 2026 snapshot of over 3.1 million US queries, Reddit had 21.6% mention share in Perplexity, just ahead of YouTube at 20.8% and well ahead of Wikipedia at 6.3%.
Does ranking on Google help me get cited by Perplexity?
Yes, more than with most assistants. In an Ahrefs study from September 2025, about 65% of the URLs Perplexity cited also ranked in Google's top 10 for the same head terms. For ChatGPT the figure was about 10%.
Should I block PerplexityBot in robots.txt?
Only if you don't want to appear in Perplexity's results. Perplexity says PerplexityBot exists to surface and link websites in its search results and is not used to train foundation models. Blocking it reduces your chance of being cited.
Why does Perplexity cite oreateai.com?
Ahrefs lists it eighth with 1.8% share but doesn't explain what the site is. I would treat it as a reminder that Perplexity will cite less familiar sites that have many matching pages, rather than as a target in itself.
How stable is Perplexity's top-cited list?
Not very, below the top three. In the September 2026 update the New York Times rose 28 places and Good Housekeeping 11, while Target fell six. Watch trends over several months before acting on one month's moves.
How I would start earning Perplexity citations this quarter
In brief: Perplexity cites communities and video heavily, it is moving toward tested reviews and newsrooms, and it follows Google's top results more closely than any other major assistant. So keep doing solid SEO, allow PerplexityBot, and put effort into the review sites, videos and publications your buyers already trust.
Start with 25 real questions, write down every source Perplexity cites, and choose three you can realistically influence. Then measure monthly.
If you would like our team to do this research and build the plan with you, take a look at our SEO services or get a quote. We are happy to help.
Further reading on websites Perplexity cites most
Sources for this Perplexity citation analysis
- The 50 Most-Cited Websites in Perplexity (Ahrefs)
- ChatGPT Doesn’t Follow Google’s Sources, Perplexity Does (Ahrefs)
- How does Perplexity work? (Perplexity)
- PerplexityBot and Perplexity-User crawler documentation (Perplexity documentation)
- Perplexity says it’s now serving 100M search queries a week (TechCrunch, October 2024)
- The 50 Most-Cited Websites in ChatGPT (Ahrefs)
- The 50 Most-Cited Websites in Gemini (Ahrefs)
- The 50 Most-Cited Websites in Google AI Overviews (Ahrefs)
- The 50 Most-Cited Websites in Google AI Mode (Ahrefs)
