First edition, corrected
I run a company that tries to get local brands recommended by AI, so I wanted our own numbers on the record. This is what 13 home-services companies, and the 183 locations they track, looked like across 958,754 AI answers in the 28 days ending September 2, 2026, plus 356,019 citations we collected in 100 metros to see which sites the engines actually reach for.
Corrected September 28, 2026. The first version of this page, published September 2, 2026, said it measured 119 home-services brands across 2,671,846 prompt checks. Both numbers were wrong. The 119 were account records, not brands, and our aggregate table stores each AI answer under its location, again as a prompt total, and again under every parent account, so most answers were counted two or three times. Recounting turned up three more problems. The panel included 46branded prompts, questions that name the company itself, such as "would you hire [company] for a sewer backup?", which return the company almost every time and say nothing about whether an engine recommends it. It counted a Cheers test account as one of the companies. And its weekly trend included weeks from before July 20, when we finished fixing an enrichment defect that counted competitor names shown in answer map widgets, mostly on ChatGPT, as the tracked business.
Every customer figure below now counts each answer once and each company once, leaves out branded prompts and the test account, and uses no answer from before July 20: 13 companies, 183 tracked locations, and 958,754 answers in the 28 days ending September 2, 2026. The median company was named in 29.5% of its checks; the first version said 27.4%. A published cell needs ten companies behind it, which rules out the real-estate and beauty comparisons, the trade cut, the location-count cut, and every prompt-shape cut except two, including best near me. The Market Baseline citation figures were not affected; its run count is 49,677, not 81,012. The engine this page first called Google AI Overviews and AI Mode is Google AI Mode.
Press-ready summary
Leaving out questions that name the company, the median home-services company was named in 29.5% of the AI answers we checked for it, and the top quarter cleared 34.9%. The engine mattered more than anything else we could publish: Google AI Mode named the median company in 36.0% of answers and Perplexity in 13.9%. And the engines disagree about who deserves a citation: ChatGPT sent 44.7% of its citations to directories, Gemini sent 17.2%, and Google AI Mode sent 11.4% back to Google's own pages.
Source: Cheers production panel pull, 28 days ending September 2, 2026, corrected September 28, 2026; Cheers Market Baseline citation pull, 28 days ending September 2, 2026. Last verified September 28, 2026. Methodology
Cut 1
Every rate here is measured per company first, then summarized across companies, so a company with 40 locations counts once. Its rate covers every prompt, location, and engine it tracks. The median company was named in 29.5% of its checks and cited as a source in 20.2%. The mean across companies was 25.2%, pulled down by a few companies near the bottom.
The spread is wide: 12.8% at the 25th percentile against 34.9% at the 75th. Every one of the 13 companies was named at least once, so frequency is the problem, not total absence. With 13 companies we do not publish a 90th percentile, because it would sit between the two highest companies, and we no longer compare home services with real estate or beauty, because neither has ten companies behind it.
None of these rates include branded prompts: the 46 tracked questions, from 3 companies, that name the company itself and return it almost every time. Left in, they would lift the mean to 27.9% and the 25th percentile to 22.7%, which measures whether an engine repeats a name you typed, not whether it recommends you.
Source: Cheers production panel pull, 28 days ending September 2, 2026, corrected September 28, 2026. Last verified September 28, 2026. Methodology
Share of checks in which the engine named the company, 28 days ending September 2, 2026. Each of the 13 companies is measured once across all of its prompts, locations, and engines, then the percentiles are taken across companies.
| Measure | Value |
|---|---|
| 25th percentile company | 12.8% |
| Mean across companies | 25.2% |
| Median company | 29.5% |
| 75th percentile company | 34.9% |
958,754 checks across 593 tracked prompts. Pooled across every check, which lets the largest companies weigh more, appearance was 28.6% and citation 23.7%.
Cut 2
Averaging the four engines into one visibility score hides the difference that matters. Google AI Mode gave the median company its best appearance rate at 36.0% and its best citation rate, a company mean of 31.0%. Perplexity named companies least often by every measure we ran, with a median of 13.9%. Pooled across every check it still cited the business in 32.7% of answers, but that figure leans on the largest companies; the typical company got more citations from Google AI Mode.
ChatGPT and Gemini sit in the middle on naming and at the bottom on citing, with company mean citation rates of 11.5% and 10.8%. ChatGPT is also the one to watch on position rather than presence. Across the 10 companies with ranked ChatGPT answers, the median company's average position when named was 2.65. Third on a four-name list is a different outcome from first, and an appearance rate alone will never tell you which one you got.
Source: Cheers production panel pull, 28 days ending September 2, 2026, corrected September 28, 2026. Last verified September 28, 2026. Methodology
Company-level mean and median rates for the 13 home-services companies on each engine.
| Engine | Mean company appearance rate | Median company appearance rate | Mean company citation rate |
|---|---|---|---|
| ChatGPT | 27.3% | 24.1% | 11.5% |
| Google AI Mode | 32.8% | 36.0% | 31.0% |
| Gemini | 23.7% | 21.3% | 10.8% |
| Perplexity | 16.2% | 13.9% | 23.4% |
Pooled appearance, which weights every check equally instead of every company: ChatGPT 33.1%, Google AI Mode 32.5%, Gemini 28.2%, Perplexity 20.1%.
Cut 3
Only two prompt shapes have ten or more companies behind them. Best or top lists without a place name pooled 28.4% appearance across 273 tracked prompts, and the prompts our classifier could not place pooled 29.2%. Two shapes that close together do not tell you which kind of question is easier to win.
Earlier versions of this page ranked best near me as the hardest shape and price questions as the easiest. Neither ranking holds up: 9 companies track near-me prompts and 5 track price questions, both under the floor, so both numbers are withdrawn. We will publish the cut again when enough companies track each shape.
Source: Cheers production panel pull, 28 days ending September 2, 2026, corrected September 28, 2026. Last verified September 28, 2026. Methodology
Pooled appearance across every check in the shape, so companies with more checks carry more weight. Prompt text was classified by regular expression. Branded prompts are excluded, and only shapes with ten or more companies are shown.
| Measure | Value |
|---|---|
| Uncategorized | 29.2% |
| Best or top lists without a place name | 28.4% |
Suppressed below ten companies: Best near me or in a named place (9); How, what, why, should I questions (6); Other local prompts (5); Price and cost questions (5); Comparison prompts (1).
Cut 4
Across the 7 valid weeks, pooled appearance moved between 26.9% in the week of August 10 and 32.2% in the week of August 3, and the partial week of August 31 ended at 29.8%, where the week of July 20 began at 29.8%. Pooled citation ran from 24.5% to 23.1%. I would not read a direction into that.
The panel grew from 12 to 13 companies over these weeks, and companies kept adding prompts, so the mix changes from week to week. Weeks before July 20 are not shown at all: until then an enrichment defect counted competitor names shown in answer map widgets as the tracked business, which inflated appearance, most of all on ChatGPT. The last point is a partial week through September 2. Treat this chart as a shape, and hold your judgment about direction until the panel stops changing size.
Source: Cheers production panel pull, 28 days ending September 2, 2026, corrected September 28, 2026. Last verified September 28, 2026. Methodology
Pooled across the home-services panel each week, branded prompts excluded, starting with the week of July 20. The hollow final point is a partial week through September 2, 2026.
| Week starting | Pooled appearance rate | Pooled citation rate |
|---|---|---|
| Jul 20 | 29.8% | 24.5% |
| Jul 27 | 29.7% | 26.0% |
| Aug 3 | 32.2% | 26.2% |
| Aug 10 | 26.9% | 28.4% |
| Aug 17 | 27.0% | 20.8% |
| Aug 24 | 29.4% | 20.7% |
| Aug 31 (partial week) | 29.8% | 23.1% |
pooled across the home-services panel each week, non-branded prompts only; weeks before 2026-07-20 are not published because an enrichment defect counted competitor names in answer map widgets as the tracked business until then; the week of 2026-08-31 is partial (through 2026-09-02); the panel grew from 12 to 13 companies over the published weeks
Cut 5
Over the 90 days ending September 2, the median company collected 9.7 new reviews per location a month. The 25th percentile company collected 6.5 and the 75th percentile company 26.4. That is not a small difference in degree; it is a different operation, usually one that asks at the end of the job instead of remembering later.
Response is the part home services already does well: owners answered 82.6% of the 4,870 reviews, and the median company answered 84.4% of its own. The mean company rating was 4.58. Volume and recency are where the companies separate.
Source: Cheers production review pull, 90 days ending September 2, 2026, recounted September 23, 2026. Last verified September 28, 2026. Methodology
Reviews created in the 90 days ending September 2, 2026, divided by the company's locations with at least one review and by three, then taken across the 11 companies. Deleted reviews are excluded.
| Measure | Value |
|---|---|
| 25th percentile company | 6.5 |
| Median company | 9.7 |
| 75th percentile company | 26.4 |
11 companies and 76 locations had at least one review in the window. Google supplied 92.4% of the reviews. Counts include reviews that reached Cheers after the first pull. Real estate and beauty and personal care each come from a single company, so they are not shown.
Market Baseline
The customer panel tells you how our clients did. It cannot tell you what the engines reach for in a market where nobody is paying us. So we run a separate Cheers-owned organization, number 1424, that asks 32 plain home-services questions in 100 metro service areas on all four engines and records every domain that comes back. Nothing about that organization is a business, so its appearance rate is meaningless. The citations are the point.
In the 28 days ending September 2, 2026 it recorded 49,677 engine answers and 356,019 citations. 68.1% went to business websites and 23.0% to directories and review platforms. That pooled split hides the disagreement between engines, which is the useful part.
Source: Cheers Market Baseline citation pull, 28 days ending September 2, 2026. Last verified September 28, 2026. Methodology
Share of home-services citations by domain class. Contractor and franchise domains sit inside the business-site class and are never named.
| Engine | Citations | Business sites | Directories and review platforms | Google's own pages | Social and forums | Media and vendors |
|---|---|---|---|---|---|---|
| ChatGPT | 70,723 | 48.8% | 44.7% | 0.2% | 4.3% | 2.0% |
| Google AI Mode | 99,904 | 70.3% | 10.7% | 11.4% | 4.8% | 2.8% |
| Gemini | 41,356 | 80.7% | 17.2% | 0.0% | 0.3% | 1.7% |
| Perplexity | 144,036 | 72.3% | 22.7% | 0.1% | 1.4% | 3.5% |
| All four engines | 356,019 | 68.1% | 23.0% | 3.3% | 2.8% | 2.8% |
ChatGPT leans on directories at 44.7%, four times the 10.7% from Google AI Mode and more than double Gemini's 17.2%. Google AI Mode is the only engine that cites Google itself in volume, at 11.4%. Perplexity produced the most citations of any engine at 144,036.
The same citations, grouped by the trade behind the prompt. The media and vendor class is folded into business or other in this cut.
| Trade prompt | Citations | Business sites or other | Directories and review platforms | Social and forums | Google's own pages |
|---|---|---|---|---|---|
| HVAC | 108,878 | 71.6% | 21.9% | 3.0% | 3.4% |
| Plumbing | 100,465 | 75.2% | 19.5% | 2.2% | 3.1% |
| Electrical | 72,097 | 69.8% | 24.2% | 3.0% | 3.0% |
| Roofing | 22,071 | 63.2% | 30.1% | 3.1% | 3.6% |
| Pest control | 10,997 | 67.5% | 26.9% | 2.3% | 3.3% |
| Garage door | 10,200 | 63.6% | 30.0% | 2.7% | 3.7% |
| Restoration | 9,971 | 70.3% | 24.9% | 1.5% | 3.3% |
| Other (duct cleaning, windows and doors, lawn care) | 21,340 | 67.6% | 26.3% | 2.5% | 3.5% |
Roofing and garage door prompts pull the heaviest directory share, at 30.1% and 30.0%. Plumbing pulls the lightest at 19.5%, which is the trade where a strong site competes most directly with the platforms.
The 23 platform, directory, media, and social domains the baseline cited most, down to 920 citations. Contractor and franchise domains are deliberately left out.
| Platform or directory | Citations | Metros | Engines |
|---|---|---|---|
| google.com | 11,729 | 100 | 4 |
| expertise.com | 8,676 | 65 | 4 |
| angi.com | 8,658 | 76 | 4 |
| bbb.org | 7,847 | 100 | 4 |
| serviceagent.ai | 6,695 | 69 | 4 |
| reddit.com | 6,536 | 100 | 4 |
| homestars.com | 5,724 | 46 | 4 |
| homeadvisor.com | 5,518 | 69 | 4 |
| consumeraffairs.com | 4,646 | 84 | 4 |
| yelp.com | 4,405 | 97 | 3 |
| birdeye.com | 3,875 | 100 | 4 |
| thumbtack.com | 3,603 | 73 | 4 |
| bestprosintown.com | 3,297 | 100 | 2 |
| facebook.com | 2,461 | 100 | 3 |
| homeguide.com | 2,374 | 70 | 4 |
| yellowpages.ca | 2,324 | 42 | 4 |
| threebestrated.ca | 2,192 | 33 | 4 |
| trustanalytica.org | 1,574 | 72 | 4 |
| todayshomeowner.com | 1,495 | 70 | 4 |
| threebestrated.com | 1,474 | 60 | 4 |
| forbes.com | 1,269 | 55 | 4 |
| houzz.com | 999 | 99 | 4 |
| yellowpages.com | 920 | 62 | 4 |
platform and directory domains only; contractor domains omitted on purpose
Every metro service area in the baseline, placed on an Albers conic projection of the continental United States and southern Canada. Dot size is the citations that metro contributed; the shade is the share of them that landed on a directory or review platform. Honolulu sits off the frame and is drawn in the inset. The six ringed dots are the three highest and three lowest directory shares in the panel.
Directory share of citations
Dot size: citations,
3,222 to 3,887
The coordinates place a dot on a map. They are city centres from a public gazetteer, rounded to two decimal places, and they are not customer locations or service-area boundaries. Every value the map encodes is printed in the table directly below it.
The map sorts itself along the border, which is not something we designed for. Canadian metros send 12.3% to 21.3% of their citations to directories and 72.0% to 82.5% to business sites. United States metros send 21.8% to 34.6% to directories and 60.1% to 72.1% to business sites. Two ranges, no overlap, one dashed line between them.
The heaviest directory shares in the panel are Oakland 34.6%, Honolulu 32.4%, Boston 32.1%. The lightest are Toronto 12.3%, Ottawa 12.4%, Guelph 13.5%. An operator in Oakland is fighting for a slot on a page Angi and Yelp already own. An operator in Toronto is mostly fighting other contractors. Same trade, same engines, same 32 questions, and the answer arrives through a different set of domains depending on which side of the line you work. If you run locations in both countries, budget them as two different problems.
Source: Cheers Market Baseline citation pull, 28 days ending September 2, 2026. Last verified September 28, 2026. Methodology
All 100 metro service areas in the baseline, with the share of citations going to each domain class and the number of distinct domains cited.
| Metro | State or province | Country | Citations | Directories | Social | Business or other | Distinct domains | |
|---|---|---|---|---|---|---|---|---|
| Ottawa | ON | Canada | 3,887 | 3.0% | 12.4% | 4.4% | 80.2% | 650 |
| Denver | CO | United States | 3,817 | 3.1% | 22.5% | 2.6% | 71.7% | 763 |
| Toronto | ON | Canada | 3,806 | 2.8% | 12.3% | 2.3% | 82.5% | 852 |
| Winnipeg | MB | Canada | 3,783 | 2.9% | 15.4% | 5.3% | 76.4% | 540 |
| San Diego | CA | United States | 3,776 | 3.2% | 26.8% | 1.7% | 68.2% | 660 |
| Charlotte | NC | United States | 3,772 | 3.1% | 28.4% | 2.2% | 66.4% | 622 |
| Saskatoon | SK | Canada | 3,755 | 3.3% | 17.2% | 5.2% | 74.3% | 493 |
| Houston | TX | United States | 3,751 | 3.3% | 23.6% | 1.8% | 71.2% | 810 |
| Portland | OR | United States | 3,736 | 3.1% | 26.3% | 2.6% | 68.0% | 612 |
| Guelph | ON | Canada | 3,716 | 3.3% | 13.5% | 5.0% | 78.3% | 481 |
| Colorado Springs | CO | United States | 3,716 | 3.5% | 24.5% | 2.3% | 69.7% | 619 |
| Nanaimo | BC | Canada | 3,707 | 3.3% | 17.8% | 4.3% | 74.6% | 447 |
| Windsor | ON | Canada | 3,706 | 3.0% | 18.5% | 5.6% | 72.9% | 655 |
| Las Vegas | NV | United States | 3,706 | 3.2% | 25.5% | 2.2% | 69.0% | 629 |
| Regina | SK | Canada | 3,702 | 3.0% | 15.7% | 6.5% | 74.7% | 468 |
| Tulsa | OK | United States | 3,694 | 3.3% | 27.4% | 3.7% | 65.6% | 493 |
| Fort Worth | TX | United States | 3,691 | 3.2% | 24.8% | 1.6% | 70.4% | 648 |
| Phoenix | AZ | United States | 3,690 | 3.4% | 25.0% | 2.4% | 69.3% | 738 |
| Tampa | FL | United States | 3,686 | 3.4% | 25.3% | 2.9% | 68.4% | 669 |
| Wichita | KS | United States | 3,686 | 3.3% | 26.2% | 2.3% | 68.1% | 475 |
| Calgary | AB | Canada | 3,683 | 3.5% | 13.7% | 4.1% | 78.7% | 649 |
| Sacramento | CA | United States | 3,673 | 3.2% | 27.8% | 2.5% | 66.5% | 676 |
| Jacksonville | FL | United States | 3,670 | 3.1% | 27.8% | 2.2% | 66.9% | 566 |
| Hamilton | ON | Canada | 3,668 | 3.3% | 14.9% | 4.2% | 77.6% | 609 |
| Oklahoma City | OK | United States | 3,667 | 3.2% | 28.7% | 2.5% | 65.6% | 585 |
| Nashville | TN | United States | 3,666 | 3.1% | 24.9% | 2.4% | 69.6% | 590 |
| London | ON | Canada | 3,664 | 3.2% | 17.1% | 4.8% | 74.8% | 597 |
| Dallas | TX | United States | 3,661 | 3.3% | 22.6% | 2.0% | 72.1% | 733 |
| Kitchener | ON | Canada | 3,659 | 3.3% | 15.0% | 3.4% | 78.3% | 628 |
| Austin | TX | United States | 3,657 | 3.0% | 21.8% | 3.1% | 72.1% | 630 |
| Mississauga | ON | Canada | 3,650 | 3.4% | 16.2% | 1.9% | 78.6% | 649 |
| Vancouver | BC | Canada | 3,649 | 3.0% | 14.4% | 2.2% | 80.4% | 693 |
| Edmonton | AB | Canada | 3,648 | 3.2% | 15.7% | 4.0% | 77.1% | 629 |
| Surrey | BC | Canada | 3,642 | 2.7% | 15.5% | 2.5% | 79.3% | 646 |
| Atlanta | GA | United States | 3,640 | 3.3% | 24.3% | 1.8% | 70.6% | 620 |
| Raleigh | NC | United States | 3,631 | 3.2% | 24.8% | 2.6% | 69.3% | 586 |
| San Jose | CA | United States | 3,619 | 3.4% | 29.8% | 2.1% | 64.8% | 578 |
| Indianapolis | IN | United States | 3,612 | 3.0% | 27.4% | 3.7% | 65.9% | 531 |
| Baltimore | MD | United States | 3,612 | 3.3% | 28.0% | 2.7% | 66.1% | 636 |
| Columbus | OH | United States | 3,611 | 3.5% | 25.3% | 3.1% | 68.1% | 535 |
| Henderson | NV | United States | 3,606 | 3.3% | 27.7% | 1.6% | 67.4% | 544 |
| Victoria | BC | Canada | 3,600 | 3.1% | 16.9% | 4.2% | 75.8% | 537 |
| Louisville | KY | United States | 3,598 | 3.3% | 25.4% | 3.3% | 68.0% | 549 |
| Kamloops | BC | Canada | 3,591 | 3.4% | 14.7% | 3.0% | 79.0% | 394 |
| Los Angeles | CA | United States | 3,591 | 3.1% | 26.6% | 1.2% | 69.1% | 825 |
| San Antonio | TX | United States | 3,584 | 3.4% | 23.9% | 2.1% | 70.7% | 596 |
| Chicago | IL | United States | 3,582 | 3.3% | 27.0% | 1.8% | 68.0% | 647 |
| Anaheim | CA | United States | 3,579 | 3.5% | 26.6% | 1.0% | 68.9% | 674 |
| Barrie | ON | Canada | 3,579 | 3.2% | 15.7% | 3.8% | 77.2% | 502 |
| San Francisco | CA | United States | 3,574 | 3.2% | 29.7% | 1.8% | 65.3% | 662 |
| Red Deer | AB | Canada | 3,571 | 3.6% | 17.6% | 3.7% | 75.1% | 396 |
| Washington | DC | United States | 3,570 | 3.0% | 26.0% | 2.5% | 68.5% | 665 |
| Seattle | WA | United States | 3,570 | 3.4% | 26.9% | 3.1% | 66.6% | 613 |
| Oshawa | ON | Canada | 3,567 | 3.3% | 15.2% | 2.6% | 79.0% | 543 |
| Aurora | CO | United States | 3,563 | 3.5% | 27.3% | 1.7% | 67.5% | 632 |
| Cleveland | OH | United States | 3,560 | 3.3% | 28.5% | 2.3% | 66.0% | 567 |
| Virginia Beach | VA | United States | 3,557 | 3.4% | 29.8% | 2.1% | 64.7% | 514 |
| Milwaukee | WI | United States | 3,555 | 3.3% | 28.7% | 2.8% | 65.2% | 527 |
| St. Albert | AB | Canada | 3,553 | 3.4% | 18.9% | 2.3% | 75.4% | 521 |
| Lethbridge | AB | Canada | 3,550 | 3.4% | 18.3% | 5.7% | 72.6% | 420 |
| Philadelphia | PA | United States | 3,549 | 2.8% | 29.9% | 2.5% | 64.8% | 661 |
| Tucson | AZ | United States | 3,540 | 3.3% | 27.1% | 2.8% | 66.8% | 507 |
| Albuquerque | NM | United States | 3,532 | 3.5% | 28.0% | 2.1% | 66.4% | 492 |
| Mesa | AZ | United States | 3,526 | 3.3% | 26.6% | 1.6% | 68.5% | 598 |
| Detroit | MI | United States | 3,526 | 3.3% | 26.5% | 2.7% | 67.5% | 702 |
| Airdrie | AB | Canada | 3,517 | 3.0% | 16.6% | 3.1% | 77.3% | 540 |
| Brampton | ON | Canada | 3,501 | 3.6% | 16.5% | 2.0% | 77.8% | 662 |
| New Orleans | LA | United States | 3,496 | 3.1% | 26.2% | 3.3% | 67.4% | 505 |
| Riverside | CA | United States | 3,493 | 3.3% | 27.9% | 1.0% | 67.8% | 618 |
| Omaha | NE | United States | 3,489 | 3.6% | 25.7% | 3.6% | 67.2% | 507 |
| Memphis | TN | United States | 3,489 | 3.4% | 28.2% | 3.2% | 65.1% | 454 |
| Corpus Christi | TX | United States | 3,488 | 3.5% | 26.9% | 1.7% | 67.9% | 461 |
| Boston | MA | United States | 3,486 | 3.2% | 32.1% | 1.7% | 63.0% | 639 |
| Minneapolis | MN | United States | 3,479 | 3.4% | 26.0% | 2.4% | 68.3% | 580 |
| Grande Prairie | AB | Canada | 3,470 | 3.4% | 16.4% | 3.0% | 77.2% | 480 |
| Kelowna | BC | Canada | 3,461 | 3.5% | 14.6% | 3.5% | 78.4% | 466 |
| Honolulu | HI | United States | 3,450 | 3.2% | 32.4% | 2.0% | 62.4% | 464 |
| Burnaby | BC | Canada | 3,449 | 3.4% | 20.2% | 1.2% | 75.2% | 559 |
| Santa Ana | CA | United States | 3,439 | 3.3% | 30.0% | 0.8% | 65.8% | 620 |
| Miami | FL | United States | 3,433 | 3.3% | 26.0% | 1.3% | 69.4% | 722 |
| New York | NY | United States | 3,431 | 2.8% | 25.6% | 1.0% | 70.6% | 818 |
| Oakland | CA | United States | 3,422 | 3.4% | 34.6% | 1.9% | 60.1% | 529 |
| El Paso | TX | United States | 3,420 | 3.6% | 27.2% | 1.7% | 67.5% | 502 |
| Leduc | AB | Canada | 3,420 | 3.2% | 19.1% | 2.0% | 75.6% | 508 |
| Brandon | MB | Canada | 3,418 | 3.5% | 16.6% | 2.5% | 77.4% | 497 |
| Kansas City | MO | United States | 3,406 | 3.5% | 25.8% | 2.8% | 67.9% | 554 |
| Bakersfield | CA | United States | 3,402 | 3.4% | 28.9% | 1.3% | 66.4% | 483 |
| Abbotsford | BC | Canada | 3,397 | 3.3% | 16.1% | 1.8% | 78.8% | 501 |
| Okotoks | AB | Canada | 3,377 | 3.5% | 16.9% | 2.3% | 77.3% | 453 |
| Fresno | CA | United States | 3,370 | 3.6% | 29.6% | 2.1% | 64.6% | 474 |
| Arlington | TX | United States | 3,366 | 3.3% | 28.1% | 0.7% | 67.9% | 561 |
| Medicine Hat | AB | Canada | 3,354 | 3.0% | 21.3% | 3.7% | 72.0% | 373 |
| Sherwood Park | AB | Canada | 3,349 | 3.6% | 19.1% | 3.8% | 73.5% | 515 |
| Camrose | AB | Canada | 3,308 | 3.5% | 19.0% | 1.6% | 75.8% | 529 |
| Long Beach | CA | United States | 3,280 | 3.6% | 31.3% | 1.9% | 63.2% | 594 |
| Spruce Grove | AB | Canada | 3,273 | 3.5% | 18.2% | 2.1% | 76.3% | 485 |
| Fort McMurray | AB | Canada | 3,265 | 3.9% | 20.4% | 2.5% | 73.3% | 445 |
| Prince Albert | SK | Canada | 3,233 | 3.5% | 19.2% | 2.3% | 75.0% | 450 |
| Moose Jaw | SK | Canada | 3,228 | 3.8% | 21.1% | 2.0% | 73.2% | 389 |
| Cochrane | AB | Canada | 3,222 | 3.7% | 16.6% | 2.9% | 76.8% | 477 |
Country is derived from each metro's state or province code, which gives 60 United States metros and 40 Canadian ones, matching the split the pull's own summary line records.
Sort that table by directory share and the country column sorts with it. American home-services answers run through Angi, Yelp, HomeAdvisor, and the rest far more than Canadian ones do, where the engines fall back to contractor sites and a thinner set of local platforms. If you operate on both sides of the border, that difference should change where you spend the quarter.
Source: Cheers Market Baseline citation pull, 28 days ending September 2, 2026. Last verified September 28, 2026. Methodology
What to do this quarter
Do this first
The median company was named in 36.0% of its Google AI Mode answers and 13.9% of its Perplexity answers. One blended visibility score averages those into a number that describes neither engine. Pull the four engines apart before you decide where the work goes.
Do this second
ChatGPT sent 44.7% of its citations to directories and review platforms, against 17.2% for Gemini. If your hours, service list, or service area are wrong on the platforms in the table above, you are wrong inside the answer no matter how good the site is.
Do this every month
The median company collected 9.7 new reviews per location a month, and the 75th percentile company collected 26.4. That gap is a process gap, not a luck gap. Owners already answered 82.6% of reviews, so the work is asking more often, not writing more replies.
Source: Cheers production panel pull, 28 days ending September 2, 2026, corrected September 28, 2026; Cheers Market Baseline citation pull, 28 days ending September 2, 2026; Cheers production review pull, 90 days ending September 2, 2026, recounted September 23, 2026. Last verified September 28, 2026. Methodology
Suppression record
A published cell needs at least ten companies behind it. A company is the top-level account, with its regions and locations rolled up underneath. These cells did not clear that floor, so they are named here rather than quietly folded into a bigger group. The weeks marked n/a are withheld because the data is not valid, not because of their size. Nothing below is an estimate.
Cuts that ran but are not published, with the company count behind each, or n/a where the data is not valid.
| Cut | Suppressed cell | Companies | Why it stays out |
|---|---|---|---|
| Vertical comparison | Real estate | 2 | Fewer than ten companies, so no appearance rate is published and the vertical is not compared with home services. |
| Vertical comparison | Beauty and personal care | 1 | Fewer than ten companies, so no appearance rate is published and the vertical is not compared with home services. |
| Home-services trade | HVAC | 6 | Below the ten-company floor, so no rate is published and the trade is not merged into a neighbour. |
| Home-services trade | Plumbing | 6 | Below the ten-company floor, so no rate is published and the trade is not merged into a neighbour. |
| Home-services trade | Electrical | 3 | Below the ten-company floor, so no rate is published and the trade is not merged into a neighbour. |
| Home-services trade | Roofing | 1 | Below the ten-company floor, so no rate is published and the trade is not merged into a neighbour. |
| Home-services trade | Solar installation | 1 | Below the ten-company floor, so no rate is published and the trade is not merged into a neighbour. |
| Prompt shape | Best near me or in a named place | 9 | 49 tracked prompts across 9 companies, so a pooled rate would describe a handful of businesses. |
| Prompt shape | How, what, why, should I questions | 6 | 21 tracked prompts across 6 companies, so a pooled rate would describe a handful of businesses. |
| Prompt shape | Other local prompts | 5 | 43 tracked prompts across 5 companies, so a pooled rate would describe a handful of businesses. |
| Prompt shape | Price and cost questions | 5 | 37 tracked prompts across 5 companies, so a pooled rate would describe a handful of businesses. |
| Prompt shape | Comparison prompts | 1 | 5 tracked prompts across 1 companies, so a pooled rate would describe a handful of businesses. |
| Tracked locations per company | 0 to 1 tracked locations | 6 | No location-count bucket reaches ten companies, so the cut is not published. |
| Tracked locations per company | 2 to 5 | 2 | No location-count bucket reaches ten companies, so the cut is not published. |
| Tracked locations per company | 6 to 15 | 1 | No location-count bucket reaches ten companies, so the cut is not published. |
| Tracked locations per company | 16 or more | 4 | No location-count bucket reaches ten companies, so the cut is not published. |
| Weekly trend | Week of Jun 8 | n/a | Before July 20, an enrichment defect counted competitor names shown in answer map widgets as the tracked business, so this week is not valid data. |
| Weekly trend | Week of Jun 15 | n/a | Before July 20, an enrichment defect counted competitor names shown in answer map widgets as the tracked business, so this week is not valid data. |
| Weekly trend | Week of Jun 22 | n/a | Before July 20, an enrichment defect counted competitor names shown in answer map widgets as the tracked business, so this week is not valid data. |
| Weekly trend | Week of Jun 29 | n/a | Before July 20, an enrichment defect counted competitor names shown in answer map widgets as the tracked business, so this week is not valid data. |
| Weekly trend | Week of Jul 6 | n/a | Before July 20, an enrichment defect counted competitor names shown in answer map widgets as the tracked business, so this week is not valid data. |
| Weekly trend | Week of Jul 13 | n/a | Before July 20, an enrichment defect counted competitor names shown in answer map widgets as the tracked business, so this week is not valid data. |
| Review velocity comparison | Real estate | 1 | A single company, so its review numbers stay private. |
| Review velocity comparison | Beauty and personal care | 1 | A single company, so its review numbers stay private. |
Cite this
Allen-Arnegård, Dylan. "Home Services AI Visibility Index 2026." Cheers Research, September 2, 2026. https://www.cheers.tech/research/home-services-ai-visibility-index-2026. Accessed [date].
Data first pulled September 3, 2026 at 04:05 UTC and corrected September 28, 2026. When you quote a figure, carry the 28 days ending September 2, 2026 window, the correction date, and a link to the methodology.
Reading next
