Perplexity SEO

Retrieval first, then re-ranking

Perplexity is a retrieval-augmented system. A question becomes one or more search queries; those return a candidate pool; a ranking model then scores that pool and only a handful of pages make it into the answer as citations. Two separate failures are possible and they need different fixes. If you never enter the candidate pool, the problem is discovery — you are not indexed, not linked, or not topically matched to the query. If you enter the pool but never get cited, the problem is extraction quality — the model could not find a clean, attributable claim on your page. We diagnose which of the two is happening before changing anything.

PerplexityBot and Perplexity-User

Perplexity operates distinct agents: PerplexityBot builds the index, while Perplexity-User fetches a page in real time because a user's question pointed at it. Robots.txt rules apply per agent, so a blanket AI block, a Cloudflare bot-fight setting or an aggressive WAF rule can quietly remove you from consideration even though your robots.txt looks permissive. We verify from the logs, not from the config file, because the two disagree more often than people expect. We also check for soft blocks: interstitials, cookie walls, JavaScript-only content and login gates all reduce what any retrieval agent can actually read from the page.

What makes a passage survive extraction

Perplexity has to attribute a specific claim to a specific source, so it favours text where the claim and its evidence sit adjacent. Pages that do well tend to name the entity explicitly rather than saying "we", answer in the opening lines, carry visible dates on time-sensitive facts, hold to one topic instead of five, and put figures in tables. Pages that do badly bury the answer under a scene-setting introduction, spread one topic across a dozen thin service pages, or state claims with no attribution at all. We rewrite for the first shape. It is unglamorous editing work and it is the part that actually moves.

Being a source it already trusts

Re-ranking leans on trust signals, and trust is largely earned off your own site. Named authors with real credentials, a consistent Organization and Person entity, citations to primary sources, and mentions on directories, industry sites and publications that Perplexity's retrieval already surfaces all feed this. There is no shortcut and no paid placement to buy. For a small Indian firm the realistic route is niche authority: be the most precise, most current source on a narrow set of questions rather than a vague generalist on a broad set. Off-page work is ₹20,000/month where authority is genuinely the bottleneck; often it is not, and we will say so.

Measurement and honest limits

Perplexity does not publish a webmaster console, so measurement is empirical. We define a prompt set covering the questions your buyers ask, record the cited domains for each, and re-run the set on a fixed schedule so movement is visible over months rather than argued about. Analytics referral data from Perplexity hosts gives a second, weaker signal. What we will not do is promise citations, quote a "citation share" as if it were a stable metric, or invent statistics about how often AI answers convert. Outputs vary by session and change with every model update; eligibility is what we can actually engineer.

Perplexity SEO — FAQs