A question we are asked often
How to rank in ChatGPT
There is no ranking inside ChatGPT. There is a set of sources the engine draws on to answer one particular question, and that set changes with the question. We sampled 16 buying questions in one city and recorded what it used.
84 citations came from 63 different places.
We asked 16 questions a Cincinnati customer might type before choosing a contractor, a surgeon, a funeral home or a flight school. ChatGPT attached sources to every answer, 16 runs out of 16, with a median of 5 sources per answer.
Those 84 citations resolved to 63 distinct domains. The single most-cited source appeared 6 times.
There is no short list to get onto, because the engine did not use one.
Source. 16 questions across four industries, Cincinnati · sampled 24 and 25 August 2026 · ChatGPT through its API, model gpt-5.4 · counts derived from the stored answers and their cited URLs.
What it actually drew on
The sources in that sample included a city government page, the Better Business Bureau, a trade association's find-a-member directory, a manufacturer's dealer locator, Reddit threads and the businesses' own websites. No single category accounted for most of the citations.
This is the practical reason a business cannot buy its way in. The engine assembled a different set of references for each question, and a business named in one answer was often absent from the next.
The work that follows from it
If inclusion comes from many places rather than one, then the work is making a business easy to describe correctly wherever it is already mentioned, and making its own pages specific enough to be worth quoting. That means the facts a buyer needs, stated plainly, in the places a buyer and an engine both look.
It is ordinary work. Accurate listings, pages that answer a real question, and references from sources that already exist in the category. The reason to measure first is to find out which of those is missing for a particular business, rather than doing all of them.
Name the surface before you buy the work
Engines behave differently depending on what is asked. On questions that describe a purchase, the engines in our sample retrieved sources and cited them. On general how-does-it-work questions, we have measured answers that arrived with no sources at all, which means content aimed at those questions has a weaker path into a chat answer than into Google's AI Overviews or Perplexity.
So the useful question to ask any provider is which surface a recommendation targets, and what evidence says it reaches that surface. A plan that treats every engine and every question as one channel has not been measured.
What sampling cannot establish
Sampling records answers. It does not establish why a business was included, and we do not claim that a change caused a result. Counts describe the questions we asked under recorded conditions. They are not market share, not impressions and not buyers.
API results are observations under recorded conditions and are not the consumer apps. We do not guarantee citations, rankings, traffic or revenue.
How we measure · Showing up across engines · Being recommended, not just mentioned · Our scope and prices
Questions we get asked about this.
Can you rank number one in ChatGPT?
There is no ranked list inside ChatGPT to be first in. It assembles a set of sources for each question it is asked, and that set changes with the question. In our sample of 16 buying questions it used 63 distinct domains across 84 citations, with no single source accounting for more than 7 percent.
How does ChatGPT choose which businesses to name?
We do not know, and sampling cannot establish it. What we can record is what it drew on. In our sample that included a city government page, the Better Business Bureau, a trade association directory, a manufacturer's dealer locator, Reddit threads and businesses' own websites.
Is there a directory I can pay to be listed in?
No single source dominated the answers we measured, so there is no one listing that would have carried a business into them. The most-cited source in the whole sample appeared 6 times out of 84.
Does ChatGPT always cite its sources?
Not always. On the buying questions in our sample it attached sources to every answer. On general how-does-it-work questions we have recorded answers that arrived with no sources at all, which means content written for those questions has a weaker path into a chat answer than into Google's AI Overviews or Perplexity.
How long does it take to appear in ChatGPT answers?
We will not put a number on that. Answers change between runs and between model versions, and nobody controls a model's output. Any provider promising a timeline is describing something they cannot see.
We publish what we find.
We're measuring AI answers in Cincinnati and will publish what we find, including our own results, including when they're unflattering. No pitch attached.
One email a month at most, and you can leave with one click. We do not sell your address and we will not add you to anything else.