Does it matter whether you're a roofer in Houston or Katy? It might. When we compared ChatGPT's answers across 23 Texas cities, the picture was a little different in the five biggest cities than in the smaller ones: the top choice was slightly less steady, and ChatGPT cited different websites.
This guide shows the numbers, offers some careful explanations, and explains what to do whether you work in a big city or a small town. It's one AI tool on one day, in Texas only, with 23 cities, so treat it as a hint, not a rule.
How we compared
We read the public snapshots behind our Who AI Recommends pages from November 10, 2026. To keep the comparison fair we used the 15 trades covered in all 23 cities (321 markets with six ChatGPT answers each, about 1,900 answers). We grouped cities by 2020 Census population:
- Large (500,000+): Houston, San Antonio, Dallas, Austin, Fort Worth (5 cities)
- Mid-size (150,000 to 500,000): Arlington, Plano, Irving, Garland, Frisco, Grand Prairie, McKinney, Mesquite (8 cities)
- Smaller (under 150,000): Denton, Carrollton, Pearland, Richardson, Round Rock, The Woodlands, Sugar Land, New Braunfels, Georgetown, Katy (10 cities)
What we found
| Large | Mid-size | Smaller | |
|---|---|---|---|
| Average share of answers naming the #1 business | 84.3% | 87.2% | 88.1% |
| Markets where #1 is named in all six answers | 43.3% | 50.0% | 50.7% |
| Markets with no business named in most answers | 10.4% | 5.8% | 2.7% |
| Markets citing Angi | 70.1% | 86.5% | 82.0% |
| Markets citing Thumbtack | 19.4% | 47.1% | 40.7% |
| Markets citing Birdeye review pages | 31.3% | 58.7% | 60.0% |
| Markets citing a government site* | 16.4% | 21 to 37% | 25 to 29% |
*Depends on how "government" is defined; large cities are 16.4% under every definition we tried.
Across all 23 cities, the link between city size (on a log scale) and the top business's share of answers was about -0.41: a moderate relationship from only 23 cities. Large-city markets were nearly four times as likely as small-city markets to have no business named in most answers.
What this might mean
We don't know the cause. These are possibilities, not findings:
- More competitors. A bigger city has more businesses per trade, which may spread ChatGPT's picks.
- More sources competing. In big cities there may be many more local articles, lists and directories, so ChatGPT's searches turn up a broader mix. We saw fewer Angi, Thumbtack and Birdeye citations in the largest cities.
- Fewer clear local standouts. A small town may have one or two businesses that stand out on the sites AI reads.
- Statistical noise. With five large cities, a couple of unusual markets can move the average.
What to do if you work in a big city
- Expect more rotation. You may be named in some answers and not others. Count across several runs. See why ChatGPT gives different answers.
- Look at your own city's sources. The sites ChatGPT read are listed under "Websites AI read to decide" on pages like roofing in Houston or plumbers in Dallas.
- Narrow your focus. In a large metro, "roofer in Dallas" is crowded. Be clear about your services and the neighborhoods and suburbs you really serve. See how to get recommended in nearby towns.
- Keep reviews fresh. See how many reviews and how recent.
What to do if you work in a small city
- A steadier field is not an easy one. If one competitor is named in nearly every answer, you're probably competing with their footprint on the sites AI reads. Compare theirs with yours, source by source.
- The basics can go a long way. With fewer competing sources, complete profiles on Angi, Thumbtack, BBB and the other sites in your market may stand out more. We haven't proven that.
- Local sources matter. City and chamber sites, local news and local lists showed up. A claimed listing on those sites is worth checking.
- Don't ignore the neighboring big city. Customers may search "roofer near [big city]."
If you work in several cities
Many contractors straddle a big city and its suburbs. Measure each market separately, because the sources and the favorite can differ even a few miles apart. Put your effort first where most of your jobs are, and check the neighboring markets next. A profile and listings that accurately list all the cities you really serve cover the basics for each of them.
For everyone
- Measure your own market with the five-minute test.
- Fix the sources it cites. See the directories ChatGPT reads.
- Make sure ChatGPT can read your site. See OAI-SearchBot.
- Compare by trade too. Painters and pest control were far steadier than general contractors, whatever the city. See which trades ChatGPT agrees on.
Limits of this comparison
- One AI engine (ChatGPT's search through OpenAI's API), one day, Texas only.
- 23 cities, and only five are large.
- Each market stored only the top 10 cited sites and top 12 named businesses.
- Population figures are 2020 Census numbers (The Woodlands is a census-designated place, about 114,000).
- Correlation isn't cause.
FAQ
Does it mean it's easier to get named in a small city? Not necessarily. The favorite there was a bit steadier. That doesn't mean it's easier for a newcomer.
Does it apply outside Texas? We don't know. We didn't study other states.
Should I change my plan based on city size? Probably not. Use it as a reason to measure your own market.
See your own market. Run a free scan at localpickai.com. You'll see whether ChatGPT names you in your city, who it names instead, which websites it read, and what to fix first. Free, no signup, about a minute.
