Clixera Research · Ottawa & Toronto 2026
Big Brands vs. Main Street
ChatGPT and Google Maps are recommending different local businesses. And many independents aren’t appearing on either.
61%
Different picks
Google Maps’ #1 missing from ChatGPT’s top three.
51%
Invisible independents
Not named by either platform in tested searches.
4,760 local questions · 106 business districts · 12 categories
Provisional figures, version v4 · 2026-10-20. The collection date is still being confirmed.
Ottawa and Toronto
Most answers came from Toronto (11,537 of 14,203). Ottawa on its own: 46%. Toronto: 64%. The 51% counts 5,135 listed independents: exact name matches, plus looser matches at the rate our hand check found real.
Ottawa
46%
ChatGPT answers that left out Maps’ #1. Out of 2,666.
53%
Local independents never named by either. Out of 1,145.
Toronto
64%
ChatGPT answers that left out Maps’ #1. Out of 11,537.
50%
Local independents never named by either. Out of 3,990.
Google Maps’ #1 left out of ChatGPT’s top 3, by city
Each bar counts that city’s answers only. The cities are not weighted equally in the combined 61%. Toronto has far more answers, so the combined figure is closer to Toronto’s. The thin line shows the likely range.
Grocery chains
2 times
Chains are 30% of grocery stores on these streets. They got 61% of ChatGPT’s top three grocery picks.
Tim Hortons vs local cafés
17% → 1%
Tim Hortons is 17% of coffee shops here. It made about 1 in 100 ChatGPT coffee top threes.
Different big brands
Home Depot: 30 vs 8
Rexall: 13 vs 34
Top three appearances per 100 ChatGPT answers vs per 100 Google Maps lists in that category.
Part 1
Being #1 on Google Maps doesn’t carry over
For each question, we checked if Google Maps’ top businesses also made ChatGPT’s top 3.
Ranking first on Google Maps did not mean ChatGPT would name you. Most of the time, it didn’t. ChatGPT’s first pick matched Maps’ first pick in 23% of answers. We asked ChatGPT each question three times. Maps’ #1 was missing from all three answers for 52% of questions.
A written answer and a ranked map are different things. A business needs to show up in both. This study compares what each one shows. It doesn’t measure how many people use them.
Part 2
Most local independents never come up
We checked all 5,135 independent businesses on our list. Did any answer or map list name them, even once?
51%
Never named by either
30%
Named by ChatGPT at least once
29%
Shown by Maps at least once
Named at least once, by category
About half of local independents were never named. That matches what Clixera found before. For most of Main Street, the problem isn’t the chains. It’s not being seen at all.
Hair salons and real estate agents were named least. Optical shops and pizza places were named most. The chart shows a strict count, so the true share named is a little higher.
Part 3
Where chains get more attention than you’d expect
We compared the share of chains on the street with the share of chains in each platform’s top 3.
| Category | Chains on the street | Chains in ChatGPT’s top 3 | Chains in Maps’ top 3 | Compared with the street |
|---|---|---|---|---|
| Dentists | 1% | 3%3.7 times | 6%7.1 times | |
| Grocery | 30% | 61%2.1 times | 51%1.7 times | |
| Hair salons | 2% | 4%1.9 times | 10%5.0 times | |
| Gyms | 25% | 38%1.6 times | 40%1.6 times | |
| Optical | 23% | 22%1.0 times | 32%1.4 times | |
| Pharmacy | 54% | 52%1.0 times | 51%0.9 times | |
| Hardware | 64% | 58%0.9 times | 40%0.6 times | |
| Real estate | 37% | 15%0.4 times | 22%0.6 times | |
| Pizza | 48% | 17%0.4 times | 19%0.4 times | |
| Burgers | 60% | 21%0.3 times | 18%0.3 times | |
| Coffee | 37% | 11%0.3 times | 9%0.2 times | |
| Sandwiches & bakeries | 39% | 5%0.1 times | 9%0.2 times |
The grey line marks “same as the street.” Bars past it mean more chains than you’d expect.
Overall, chains got about the share you’d expect from what’s on the street. Grocery is the big exception. ChatGPT picked chains about twice as often as the street mix. Google Maps did too, a little less. Gyms also got more chain picks than expected. Coffee, burgers, pizza and bakeries went the other way. Both platforms picked local places far more often.
Tim Hortons is a good example. It is 17% of the coffee shops on our list. It made ChatGPT’s top 3 in about 1 of every 100 coffee answers. It made none of the 421 Google Maps coffee lists. Starbucks and McDonald’s were similar.
Part 4
Neighbourhoods: Kanata, Barrhaven, Kensington Market and more
Kanata Central
- Chains on the street
- 42%
- Chains in ChatGPT’s top 3
- 45%
- Chains in Maps’ top 3
- 31%
Kanata North
- Chains on the street
- 37%
- Chains in ChatGPT’s top 3
- 35%
- Chains in Maps’ top 3
- 28%
Barrhaven
- Chains on the street
- 44%
- Chains in ChatGPT’s top 3
- 42%
- Chains in Maps’ top 3
- 40%
Kensington Market
- Chains on the street
- 15%
- Chains in ChatGPT’s top 3
- 13%
- Chains in Maps’ top 3
- 18%
Bloorcourt Village
- Chains on the street
- 19%
- Chains in ChatGPT’s top 3
- 14%
- Chains in Maps’ top 3
- 21%
In Kanata Central, ChatGPT named many more chains than Maps did. In Barrhaven, the two were close. In Kensington Market and Bloorcourt, independents held most spots on both.
Fewer chains than the street · ChatGPT
- The Waterfrontstreet 44% · ChatGPT 23%
- DuKe Heightsstreet 61% · ChatGPT 35%
- Downtown Yongestreet 53% · ChatGPT 34%
- Old Town Torontostreet 41% · ChatGPT 26%
- Bloor-Yorkvillestreet 39% · ChatGPT 26%
- Toronto Downtown Weststreet 38% · ChatGPT 27%
More chains than the street · ChatGPT
- Weston Villagestreet 9% · ChatGPT 26%
- Albion Islington Squarestreet 13% · ChatGPT 31%
- Village of Islingtonstreet 16% · ChatGPT 34%
- Uptown Yongestreet 15% · ChatGPT 32%
- Leslievillestreet 12% · ChatGPT 25%
- Rosedale Main Streetstreet 12% · ChatGPT 24%
We looked at 71 BIAs with enough businesses to compare. In 20 of them, ChatGPT recommended chains more than you’d expect from what’s on the street. In 8, it recommended them less. Google Maps was similar, at 17 and 7. Chains got extra attention on main streets that have few chains. Downtown, where chains are everywhere, both platforms favoured independents.
Part 5
Same category, different brands
How often each brand made the top 3, out of every 100 questions in its own category.
The Home Depot
ChatGPT: 30 in 100.
Google Maps: 8 in 100.
RONA
ChatGPT: 21 in 100.
Google Maps: 10 in 100.
Rexall
ChatGPT: 13 in 100.
Google Maps: 34 in 100.
GoodLife Fitness
ChatGPT: 27 in 100.
Google Maps: 42 in 100.
ChatGPT named Home Depot and RONA far more often than Maps did. Maps showed Rexall and GoodLife far more often than ChatGPT did.
| # | Brand | Top 3 spots, out of every 100 questions in its category |
|---|---|---|
| 1 |  Home Hardware Hardware | ChatGPT 69.2 Maps 64.3 |
| 2 |  Shoppers Drug Mart Pharmacy | ChatGPT 52.0 Maps 32.5 |
| 3 | GoodLife Fitness Gyms | ChatGPT 26.7 Maps 41.5 |
| 4 |  No Frills Grocery | ChatGPT 31.6 Maps 21.7 |
| 5 | Rexall Pharmacy | ChatGPT 13.2 Maps 34.4 |
| 6 | Guardian / I.D.A. Pharmacy | ChatGPT 19.9 Maps 23.5 |
| 7 |  Metro Grocery | ChatGPT 26.8 Maps 15.8 |
| 8 |  Anytime Fitness Gyms | ChatGPT 19.1 Maps 18.5 |
| 9 | The Home Depot Hardware | ChatGPT 30.1 Maps 7.6 |
| 10 |  Loblaws Grocery | ChatGPT 16.7 Maps 14.9 |
| 11 | RONA Hardware | ChatGPT 21.2 Maps 10.4 |
| 12 |  Pharmasave Pharmacy | ChatGPT 15.8 Maps 15.7 |
Part 6
Which platform shows more chains, by category
Dark bars: ChatGPT showed more chains. Blue bars: Maps showed more. Faded bars: no clear difference.
ChatGPT showed more chains in hardware and grocery. Maps showed more chains in optical, real estate and hair. For pharmacy, gyms and pizza, the two were about the same.
Part 7
Chains, independents and names we couldn’t place
We sorted every name into chain, independent or not sure. “Not sure” means the name wasn’t on any of our lists.
ChatGPT (OpenAI)
Best estimate once unknown names are sorted: about 27% chains.
Best estimate once unknown names are sorted: about 26% chains.
Google Maps
Best estimate once unknown names are sorted: about 26% chains.
Best estimate once unknown names are sorted: about 26% chains.
Overall, the two platforms named about the same share of chains. The gap was wider on “What is the best…” questions. There, ChatGPT named chains more often than Maps did.
When the two first picks were different, ChatGPT chose the chain 788 times. Maps chose the chain 641 times.
Part 8
Do the results hold with just one answer?
We asked ChatGPT each question three times. To check, we redid the main numbers using only its first answer.
| Finding | All three answers | First answer only | Holds? |
|---|---|---|---|
| Maps' #1 missing from ChatGPT's top 3 | 60.6 | 60.4 | Yes |
| ChatGPT top-3 picks not in Maps top 3 | 68.4 | 68.3 | Yes |
| Listed independents never named (adjusted) | 50.6 | 53.1 | Yes |
| Estimated chain share, ChatGPT top 3 | 26.3 | 26.2 | Yes |
| Grocery chains compared with the street (ChatGPT) | 2.07 | 2.07 | Yes |
| Gyms chains compared with the street (ChatGPT) | 1.55 | 1.58 | Yes |
| Coffee chains compared with the street (ChatGPT) | 0.29 | 0.27 | Yes |
| Home Depot: ChatGPT vs Maps gap | 22.5 | 22.2 | Yes |
| RONA: ChatGPT vs Maps gap | 10.8 | 11.8 | Yes |
| Rexall: ChatGPT vs Maps gap | -21.3 | -20.9 | Yes |
| GoodLife: ChatGPT vs Maps gap | -14.8 | -14.9 | Yes |
Every main finding held. With only one answer, a few more independents go unnamed. That’s expected, since there are fewer chances to be named.
Downloads
Version v4 · 2026-10-20. Provisional until final checks are done.
Methodology and limitations
Questions. 5,088 questions: four question styles, 12 categories and 106 BIAs. Each went to ChatGPT (OpenAI) three times and to Google Maps once. 328 were dropped: 327 had no Maps results and 1 was missing a ChatGPT answer. That leaves 4,760 questions and 4,760 Maps lists. Of 14,280 ChatGPT answers, 14,203 named at least one business and are used. The collection date was not recorded in the dataset. Location settings were not recorded; every question named the neighbourhood and city.
Weighting. Each ChatGPT answer counts as one third of a question in shares, so a question weighs the same on both platforms. Brand rates are per 100 ChatGPT answers or Maps lists in the brand’s own category.
Recommendation gap. Unit: ChatGPT answers (n = 14,203). Maps’ #1 is “missing” if no name in that answer’s top 3 matches it. Matching: same name after removing branch or location words (exact, fuzzy token-set ≥ 90, or shared prefix of 6+ characters). A different branch of the same chain counts as a match. With strict matching (branch words must agree) the gap is 69.8%. Ottawa 46.1% (95% CI 40.5–51.8), Toronto 64.0% (95% CI 60.8–67.1). Overall 95% CI 57.7–63.5. First answer only 60.4%; one random answer 60.8%.
Never named. Unit: listed independent businesses (n = 5,135) in the 12 categories. Strict exact-name matching gives 62.8% never named (95% CI 60.1–65.4). Strict matching misses some real mentions, so it overstates “never named.” Loose fuzzy matching gives 41.3%, but a hand check of 60 loose-only matches found 34 real (57%). Adjusting for that gives 50.6% (range 48.1–53.3). Ottawa 53.2%, Toronto 49.9%. First answer only 53.1%.
Classification. Names were matched to a list of about 19,500 businesses, a national brand list, franchise and regional chain names, names at 3+ addresses in the data, and websites cited by ChatGPT. A chain is a brand with 3 or more locations. Unknown names: 37.4% of ChatGPT names and 40.2% of Maps names. A hand check of 200 random unknowns found 12 chains (6.0%, 95% CI 3.5–10.2) and 6 non-businesses. Chain shares spread unknowns using those rates.
Street comparison. Expected chain share = share of listed businesses in that category or BIA that are chains (7,053 listed in the 12 categories). Over-representation ratio = top 3 chain share ÷ expected share. 95% CIs use a two-stage cluster bootstrap (BIAs, then questions within BIAs; 1,000 draws), redrawing the unknown-sample rates each time. The listed share is held fixed. Answers can name businesses just outside a BIA.
Limits. Point-in-time snapshot. Answers change with wording and over time. Real estate agents count on their own unless their listing names a brokerage. The study does not measure how many people use either platform.
Does ChatGPT recommend your business?
See how you show up in ChatGPT and Google Maps. Compare yourself with the brands on your street.
Related
Prepared by Clixera, 2026.