What Perplexity is, in practice
Perplexity is a synthesis-led search product. Each answer is a generated response with numbered citations to the sources the model treated as load-bearing. Users see, in line, which source supplied which claim. The system therefore exerts unusual editorial discipline: a source that does not earn a citation is not surfaced.
For our purposes, Perplexity behaves as a particularly demanding reader. The system's citation set is consistently smaller than ChatGPT's, more academically inclined, and more sensitive to signal quality than to signal volume.
How its citation logic works
Three patterns recur in the institution's observation across hundreds of subject prompts.
- Editorial sources outweigh commercial sources. Where editorial coverage exists, it tends to be cited ahead of the subject's own page on the same topic.
- Specificity outweighs breadth. Pages that answer one question precisely outperform pages that answer many questions partially.
- Attribution outweighs volume. A page authored by a named, credentialed practitioner is cited more frequently than a page authored by “Team” or unattributed.
Perplexity is the system on which adding a credentialed, named author to existing pages produces the largest single observable lift. The page need not change in substance; it needs to acquire a person standing behind it.
The three moves that work
One — attribution architecture
Each substantive page on the subject's site carries an explicit named author with credentials and a linked authority profile. The author is real, the credentials are verifiable, and the relationship between author and page is consistent across the site.
Two — editorial standing
The subject earns mention in editorial sources that Perplexity's reranker treats as authoritative — trade press, professional bodies, recognised academic adjacents. This is a slow move, and the move that compounds.
Three — specific-question pages
For each substantive question buyers ask, the subject publishes a page that answers it directly, with evidence, in compact form. Pages of this kind are surfaced by Perplexity disproportionately to their search volume.
What does not work
- Cluttering pages with disclaimers, related links, and CTAs that dilute the citable passage.
- Reproducing high-volume content at superficial depth across many pages.
- Unattributed “team-authored” content even where the team is credible.
- Submitting content directly to Perplexity in the belief that submission is part of the system.
How the institution measures it
The Index measures Perplexity citation density across a stable prompt set per subject. Movement on Perplexity tends to track movement on the Authority and Citation Source Quality signals more tightly than on any other system, which makes Perplexity a useful early indicator: improvement on Perplexity typically precedes improvement on ChatGPT and Google AI by one to two quarters.
A closing observation
Perplexity rewards the editorial discipline classic SEO let businesses unlearn. It is the system that makes a credentialed author worth more than a thousand backlinks.
The institutions that succeed on Perplexity are, with striking consistency, the institutions that would have done well in print: clear arguments, named authors, evidence on display. The system reads the way an editor reads.