Perplexity shows its sources.
Are any of them yours?

Perplexity searches, then writes. That means there are two ways to lose. The first is retrieval: the search layer never surfaces your page for the question, so nothing you wrote matters. The second is selection: your page is retrieved but the model quotes someone else, because their page answers the question in a cleaner, more liftable way. These have different fixes, and most advice conflates them. We report them separately.

Across the checks we've run, the pages Perplexity lifts from share a shape. The answer appears in the opening paragraph, in plain declarative sentences. Facts carry numbers, dates and named entities rather than adjectives. Headings are literally the questions people ask. There is a visible published or updated date. And the page is not 70% shared boilerplate with forty other pages on the same site. None of that is exotic. It is just rarely done on purpose.

When Perplexity cites a competitor for one of your questions, we fetch that exact URL and your closest equivalent page and compare them on the signals engines use: structured data, brand identity, trust signals, specificity, and duplication. You get a side-by-side of what their page has that yours doesn't: a specific list, on a specific URL, for a specific question you lost.

Don't test vanity questions. Test the ones with buying intent: comparisons, "best X for Y", "who does X in Z", and the objection your sales calls always open with. You can edit every question before we run it, so the report reflects your funnel rather than a generic keyword list.