How Perplexity picks local businesses

Short answer

Perplexity is the smallest of the four engines by usage and the most useful of the four to audit, because it tells you what it read.

Every answer carries numbered citations. When a competitor is named and you are not, you can open the sources it used and see exactly what it found about them and did not find about you. No other major engine gives you that.

What makes it different

It runs its own index. Not Bing, not Google. So presence elsewhere does not transfer automatically, and a practice can be well covered on the other three and thin here.

Community discussion is weighted unusually heavily. Reddit threads, forum posts and long-form discussion appear in its citations far more often than they do in ChatGPT or AI Overviews. For categories where people research publicly before buying — cosmetic procedures, GLP-1 treatment, anything expensive and elective — this matters more than it sounds.

It answers with sources attached. Which is the whole diagnostic advantage.

Why the Reddit weighting is awkward, and why that is the point

You cannot buy your way into it. Astroturfing a subreddit is against site rules, obvious to moderators, and one of the faster ways to acquire a permanent negative result attached to your business name. It is also the kind of thing that ends up screenshotted.

What does work is slower and less satisfying: being genuinely discussed. Patients who had a good experience mention you when someone asks. Practitioners who answer questions in their own name, without pitching, get associated with the topic. Local threads about “where do I go for X in this city” accumulate over years.

There is no shortcut here, which is precisely why the signal holds value. The engines weight it because it is hard to fake.

How to use it as a diagnostic

This is the practical part, and it takes about twenty minutes.

1. Open Perplexity in a signed-out window.
2. Ask it the five questions a patient would ask about your category and city.
3. For each answer, note who was named — and then open the citations.
4. Write down what each cited source says about the named business.

You will usually find one of three things:

• A directory you are absent from. Fixable in an afternoon.
• A review platform where they are strong and you are thin. Fixable over months.
• A community thread you have no presence in. Not directly fixable, and worth knowing before you spend anything trying.

That third case is the one that saves money. A practice can spend a year on listings and never touch the reason a competitor is being named.

What to actually fix

In rough order of return:

• Be in the directories its citations keep surfacing. They vary by category; read the citations rather than assuming.
• Get review depth on specific treatments, not just volume — cited review pages tend to be the ones with detail.
• Answer questions publicly under your own name where your category is discussed. Slowly. Without pitching.
• Make your own pages quotable. Perplexity cites pages that state things plainly. A page with a clear answer near the top gets pulled; a brochure does not.

What this will not tell you

Perplexity’s answer is not a proxy for the other three. Its index is its own, and being named here says nothing about ChatGPT — see how AI engines choose which business to name.

It is also the smallest of the four by query volume. Treat it as the surface that explains the others rather than the one that sends the most patients. The reason to run it first is that it is the cheapest way to find out what the engines are reading.

The caveat

Perplexity has changed its retrieval and its citation behaviour more than once, and it publishes no ranking documentation. This reflects observed behaviour rather than confirmed factors. Run it yourself before trusting anyone’s account of it, including this one — which, conveniently, is the one engine where doing so is easy.