PerplexityBot indexes your pages so Perplexity can cite them in answers. That is a searcher, in Hardal's trainer / searcher / fetcher split. It is closer to OAI-SearchBot than to GPTBot.
When a person asks Perplexity a question and the product retrieves your URL live, you may also see Perplexity-User. That is the fetcher. Merge the two into "Perplexity traffic" and you will miss a live query on /pricing.
There is no Perplexity Ads conversion destination in Hardal. Do not install the OpenAI Measurement Pixel for this bot.
How it shows up
User-agent contains PerplexityBot. Playbook log snippets already include it next to GPTBot and ClaudeBot (AI Visibility Playbook). GA4 still will not record the crawl. Human click-throughs from Perplexity, if they happen, show up like any other referrer as long as the browser runs tags. The crawl that decided whether you were citable never ran a tag at all.
What to report
- Which URLs PerplexityBot actually reads. Zero hits on the comparison page you care about means you are not in that index, no matter how good the copy is.
- PerplexityBot vs Perplexity-User. Fetch spikes are the "someone asked just now" line. Same play as ChatGPT-User in AI crawler traffic for marketers.
- robots.txt: allow the indexer if you want citations; blocking it is a visibility choice, not an SEO-Google choice.
robots.txt
Stay eligible for citations:
User-agent: PerplexityBot
Allow: /
Block indexing (you will not show up as a cited source the same way):
User-agent: PerplexityBot
Disallow: /
User-initiated fetches are a different token. Perplexity has argued robots.txt may not bind those the way it binds the indexer. Measure them anyway.
How Hardal measures it
AI Visibility identifies PerplexityBot by name and attributes it to Perplexity. isvisible.ai checks whether Perplexity's agents can access the site at all. First-party server-side, no extra script.
If your actual paid channel is ChatGPT Ads, that guide is here.