OwnlyticsPerplexity traffic: what to expect and how to track it
Perplexity shows its sources as links, and clicks on them usually arrive with perplexity.ai as the referrer, so they show up as their own source in your analytics. What volume to expect is a different matter: it depends on your topic and your pages, and nobody can promise a number or a citation.
Fig.Perplexity next to the other AI sources. Sample data, not a real site.
What does Perplexity traffic look like?
Perplexity answers a question and lists the pages it drew from. A reader who wants more can click one. Compared with a search results page, an answer with sources may give the reader less reason to click, so small numbers are plausible for many small sites.
When a click does happen, it normally carries a referrer of perplexity.ai. That makes Perplexity easier to count than assistants whose apps strip referrers. Ownlytics recognizes it and files it in an AI channel next to ChatGPT, Claude, Gemini and Copilot. Tagged links such as ?utm_source=perplexity or ?ref=perplexity map to the same source. The wider view is on our Perplexity traffic guide.
What are PerplexityBot and Perplexity-User?
Two different things appear in server logs, and mixing them up causes confusion:
- PerplexityBot indexes pages for Perplexity search and, per Perplexity, follows robots.txt.
- Perplexity-User fetches pages as part of answering a person's request.
A bot fetch is not a visit. It will not appear in a normal analytics report that runs on a JavaScript tracking script, and it does not mean you were cited. If you want to allow or block these, check Perplexity's own documentation for current crawler names, because bot names and behavior change. We cannot verify them for you here.
Blocking a crawler in robots.txt is a decision about whether you want to be eligible to appear. It is not a measurement tool.
How do you check Perplexity traffic in a few steps?
- Open your sources report for the last 90 days. Use a longer window than for search, because volumes are small and noisy.
- Filter for
perplexity.ai. - Look at the landing pages. These are the pages Perplexity chose to cite.
- Compare them with your own list of best pages. Is Perplexity picking pages you would not expect?
- If you can, link visitors to sales and see what those visitors bought.
Step 4 is the most valuable. The pages being cited tell you what kind of question your site is considered an answer to.
Example (made-up numbers for illustration)
A made-up small site, 90 days of Perplexity visits by landing page:
| Landing page | Visits | Sales |
|---|---|---|
| /blog/what-is-x-explained | 41 | 0 |
| /compare/x-vs-y | 18 | 2 |
| /pricing | 7 | 1 |
| / | 4 | 0 |
The invented pattern here: the most-cited page is an explainer that never sells, while the comparison page converts a few. The practical action would be to add a clear next step to the explainer and check that the comparison page is accurate and current. With 70 visits and 3 sales, you would not redesign the site on this evidence.
What are the limits of tracking it?
- Hidden clicks. A link opened from an app can arrive with no referrer, and a URL copied from an answer into a browser never has one. Those visits look like any other unlabeled visit and will be counted as Direct.
- No-click value. Someone may read your name in an answer and search for you later. That shows up under Organic Search or Direct.
- Moving target. How Perplexity chooses and shows sources changes over time. A page cited this month may not be cited next month.
So treat the number as a floor and a trend. For the broader question of whether such visitors buy, see do AI-referred visitors convert.
| Source | Visitors | Revenue |
|---|---|---|
| chatgpt.com | 1,411 | $196 |
| perplexity.ai | 620 | $98 |
| claude.ai | 388 | $49 |
Fig.AI assistants as their own channel in the source table. Sample data, not a real site.
What helps pages get cited?
No one has a formula, but a few commonly suggested habits cost nothing, though none is proven:
- Put the direct answer in the first two sentences of the page.
- Add a small table of facts a reader could check.
- Keep dates and numbers current, and say when something was last updated.
- Be mentioned on other sites. Assistants often repeat what third-party pages say about you, so off-site mentions can matter as much as your own pages. We go through the list in our generative engine optimization basics.
None of this guarantees a citation. Do it because the pages become better for people too.
When should you ignore Perplexity?
If your audience is local, non-technical, or mostly arrives from social media, Perplexity may be a rounding error for you. Look at the report once a quarter. If it is empty, move on without guilt.
At a glance
| Item | What we know |
|---|---|
| Usual referrer | perplexity.ai |
| Tagged links | utm_source=perplexity or ref=perplexity map to the same source in Ownlytics |
| PerplexityBot | Indexes for Perplexity search (per Perplexity; check its current docs) |
| Perplexity-User | Fetches pages for a user's request |
| Missing clicks | App links and copy-paste arrive without a referrer |
| Guaranteed citations | None possible |
To see AI sources next to search and social, with revenue per channel on a sample site, there is a live demo. The docs on how sources are decided explain the rules.
FAQ
Does Perplexity send a referrer?
Clicks on its citations usually arrive with perplexity.ai as the referrer, so they appear as that source. Clicks from apps or from copied URLs may arrive with none, and those will show up as Direct.
Is PerplexityBot the same as a visitor?
No. PerplexityBot is a crawler that indexes pages and follows robots.txt. A normal analytics script does not count it as a visit. Only a person clicking a link produces a visit with a perplexity.ai referrer.
How much traffic should I expect from Perplexity?
It varies widely by topic and site, and we will not guess. Start by measuring 90 days of visits, then judge whether it is worth more effort.
Will blocking Perplexity's crawlers hide me from its answers?
It may reduce your chance of being used as a source, but we cannot promise exactly how any assistant behaves. Check Perplexity's current documentation before making that call.