Generative engine optimization · Market size
In Chicago, Miami, and New York, ChatGPT puts directories up front. In Tulsa, that pattern starts to disappear.
We've already shown that for local questions, ChatGPT tends to list directories before the actual business, grabbing that top spot about 41% of the time. But we hadn't dug into whether that's true everywhere, or if city size changes the game.
There's a difference between a business getting cited at all and a business getting cited early enough that a customer actually reads it before they stop scrolling the answer. Most of what we've measured so far has been about the first kind of visibility. This time we wanted to know whether the second kind, where you land in the list once you're already in it, depends on the size of the market you're in.
So we went back to our probe database from September 28, a set of 1,423 citations spanning ten service categories and five U.S. cities. This time, we grouped cities by population instead of by niche. Chicago, Miami, and New York were our large metros; Denver was mid-size; and Tulsa was our small market. We asked two things. How often does a directory get cited at all? And when it is, where does it show up in the rankings?
The first question has a pretty straightforward answer. Directories are 28.6% of citations in the large metros, 26.2% in Denver, and 27.1% in Tulsa. If this were the only stat we looked at, we'd say market size is irrelevant here and call it a day.
But position tells another story entirely. In the large metros, directories average position 3.82 versus 4.71 for individual business sites, shoving businesses back by almost a full spot. In Denver, that gap vanishes and flips: directories average 4.53, businesses 4.24, so the business actually lands higher. In Tulsa it's basically a tie, 4.39 versus 4.32. You see the same split at the very top spot: a directory takes position one 40% of the time in the large metros, 25% in Denver, and 27.5% in Tulsa.
So the odds of being cited are about the same everywhere. What changes by city size is whether you get stuck behind a directory once you are cited.
Why the gap might open up where it does
We can't see the model's thought process, so this is an educated guess rather than a proven cause. Still, a few clues point in the same direction. Big metros have way denser directory ecosystems: more reviews, more aggregators built specifically for Chicago or Miami or New York, all heavily optimized for the exact queries we ran. A smaller market like Tulsa has fewer businesses fighting for the same spot and thinner directory coverage to start with. The kind of listing that normally outranks a business page there often just doesn't have much detail. When the directory option is weaker, the model seems to default to whatever source is actually current and specific, which tends to be the business's own site.
Picture the same question asked twice, once about a plumber in New York and once about a plumber in Tulsa. In New York, there's a stack of plumbing directories with thousands of reviews each, built by people who treat New York as a market worth optimizing for directly. In Tulsa, a lot of that infrastructure simply doesn't exist at the same depth, so there's less for the model to lean on instead of the plumber's own page.
What we can't claim from this
Our large-metro figure is an average of three cities. Denver and Tulsa are each single cities representing their whole category. That's a real limitation: we're looking at one data point standing in for "mid-size markets" and another for "small markets," not an average like the large-city number. A different mid-size or small city could land somewhere else entirely, and we'd want more cities in both buckets before we call this a confirmed market-size effect instead of a first look.
We also checked whether the pattern holds within specific niches, auto repair in Tulsa versus auto repair in Chicago, for example. The sample sizes get small quickly, some down to five or six citations per niche per city size. That's not enough to split honestly, so we're reporting the pattern at the city-size level instead of breaking it down niche by niche.
What this means depending on where you operate
If you're running a business in a place like Tulsa or Denver, this is mildly good news: when ChatGPT cites you, it isn't automatically slotting you behind a directory the way it does in the biggest metros. The fight is still to get cited in the first place, and our directory share numbers say that part doesn't change by city size. But once you're in the answer, you aren't starting from a worse position just because the market is smaller.
But if you're in a market the size of Chicago, Miami, or New York, the story flips. Being cited at all isn't the finish line. Our data says you're still likely to land behind a directory even when your own site made the cut, and that matters if you're counting on being the first name a potential customer reads before they stop scrolling.
Source: askchat.studio's GEO probe database, citations across ten service categories (auto repair, optometrist, personal injury lawyer, plumber, gym, med spa, dentist, chiropractor, veterinarian, hair salon), five U.S. cities split by market size (Chicago, Miami, New York as large; Denver as mid-size; Tulsa as small), collected September 28, 2026, through the OpenAI Responses API with web search grounding.