Vacation rental SEO now has two audiences: Google and Bing, and the AI engines guests ask where to stay. This is the method page for our November 2026 study of how those AI engines recommend vacation rentals. We asked ChatGPT, Google Gemini, Google AI Overviews, Perplexity and Claude 1,330 traveler questions across 50 US vacation markets, on three separate days, and reviewed every cited website one at a time. The findings are in AI search for vacation rentals and its six deep dives. This page explains how we got them, how firm each one is, and how this study relates to vacation rental SEO.
This is a snapshot from November 2026. AI engines change fast, so treat these numbers as a read on the moment, not a permanent rule.
We wrote the questions the way guests ask them, at every stage of a trip, from picking a destination to booking specific dates. Here is one question from each stage, with Gatlinburg as the example market:
Each market got the same 26 questions with its own place names and landmarks, so every market can be compared with every other. Here is who the engines cited at each stage, across all 50 markets:
Each stage adds to 100% of the sources the engines cited, each engine weighted equally.
| Stage | Property managers | OTAs | National rental brands and travel search sites | Hotels, resorts, inns | Travel blogs and media | Tourism boards and government | Attractions and local businesses | Social sites and forums | Other |
|---|---|---|---|---|---|---|---|---|---|
| Where to go | 5% | 12% | 8% | 3% | 49% | 5% | 4% | 9% | 5% |
| Things to do and trip planning | 10% | 6% | 4% | 4% | 28% | 29% | 6% | 10% | 2% |
| Which area to stay in | 9% | 18% | 7% | 5% | 29% | 14% | 4% | 14% | 1% |
| Which company or kind of place | 36% | 21% | 7% | 3% | 7% | 6% | 2% | 11% | 7% |
| A home for my group | 31% | 30% | 13% | 5% | 6% | 4% | 2% | 7% | 1% |
| Several needs stacked in one ask | 19% | 50% | 13% | 4% | 2% | 3% | 1% | 5% | 2% |
| Dates or a budget | 41% | 34% | 14% | 2% | 2% | 3% | 1% | 2% | 0% |
| Long, the way guests talk | 18% | 44% | 10% | 8% | 5% | 4% | 1% | 8% | 1% |
Base: 123,695 citations from 19,950 answers, five engines, 50 markets. November 14 to 16, 2026. A snapshot; AI answers change.
Property managers are a small slice of the answer while guests plan, and one of the two biggest sources once guests shop for a company, a home for their group or dates. The full story is in where property managers beat OTAs, stage by stage.
Every website the AI engines cited, returned in their own searches, or ranked in Google's or Bing's results for these questions was reviewed one at a time, not by a rule of thumb, and put in a category: property manager, OTA, national rental brand or travel search site, hotel, travel blog or media, tourism board, attraction or local business, social site or forum. More than 21,000 websites were reviewed this way, including the 8,715 the engines cited.
To check the sorting, a second, independent review of 241 of the same sites agreed with the first on 96% of them, and on 98% for the question that matters most here: property manager or not.
When we compare engines, each engine counts equally, so Perplexity's 10 sources per answer do not outweigh ChatGPT's 3. The five engines cite very different mixes of websites:
Each bar adds to 100% of that engine's citations.
| Engine | Property managers | OTAs | National rental brands and travel search sites | Hotels, resorts, inns | Travel blogs and media | Tourism boards and government | Attractions and local businesses | Social sites and forums | Other | Answers citing a property manager | Answers citing an OTA | Sources per answer |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ChatGPT | 30% | 26% | 6% | 4% | 10% | 17% | 2% | 1% | 3% | 39% | 38% | 3.1 |
| Gemini | 29% | 28% | 12% | 5% | 11% | 8% | 3% | 3% | 2% | 73% | 69% | 7.6 |
| Perplexity | 10% | 46% | 11% | 1% | 11% | 9% | 2% | 6% | 3% | 40% | 91% | 10.0 |
| AI Overviews | 25% | 12% | 7% | 9% | 8% | 6% | 5% | 27% | 2% | 58% | 45% | 6.8 |
| Claude | 24% | 23% | 14% | 3% | 21% | 5% | 3% | 5% | 2% | 41% | 46% | 3.5 |
Base: 19,950 answers, 123,695 citations. November 14 to 16, 2026. A snapshot; AI answers change.
For every question, we recorded the first 50 results on Google and on Bing on November 14 (Google's first 10 on the later days), and the Google Maps results for each market. Then we built a panel of 4,953 local property managers found on Google Maps, in Google and Bing results, or in the AI answers. To compare what goes with being cited, we used the ones found on Maps or in search results (5,564 manager-market pairs), whether or not any AI engine cited them. For each one we recorded where it ranks for each question, its Google Maps position and Google review count, links from other websites, and the pages on its site.
Building the panel this way matters. If you only study the sites AI engines already cite, you can only compare winners with winners. Starting from every manager we could find shows what the cited ones have that the others do not.
We asked the same 1,330 questions on three days. The big picture held; the individual names did not.
Bars are drawn to a 50% scale so small day-to-day changes are visible. Google changed the Gemini model on November 16; Gemini's overall share did not move.
Base: the same 1,330 questions asked on three days; 19,950 answers. November 14 to 16, 2026. A snapshot; AI answers change.
What it means for a property manager: AI visibility is a rate, not a yes or no. Checking ChatGPT once and seeing your name, or not seeing it, tells you little. Ask the same questions several times and track how often you appear.
We label every finding by how much weight it can carry:
Four limits apply to everything:
Yes. Search ranking was the strongest signal we measured: where a property manager's site ranks for the exact question a guest asks, on Google and on Bing. A property manager in Google's top three for a question was cited for that question by at least one AI engine 68% of the time; in Bing's top three, 66%. For four of the five engines, a property manager on Bing's first page for a question was more likely to be cited than one on Google's first page; Google AI Overviews follow Google.
What changes with AI search is the scoreboard. There are five engines instead of one, they do not agree with each other, and the answer depends on the stage of the trip.
Each bar is the share of manager-question pairs at that position that at least one of the five engines cited for that same question.
Base: the November 14 property-manager panel × their market's 26 questions (210,050 manager-question pairs). Search results and AI answers from November 14, 2026. A snapshot; AI answers change.
Read is AI search different from SEO? for the full numbers.
These come from patterns in the data, not experiments, so treat them as good bets rather than guarantees.
If you are starting from scratch, our guide on how to build a vacation rental website covers the site itself.
This is our fourth study of AI search for vacation rentals, and the largest. Each one asked a different kind of question:
This one is the largest and the most precise. Where they differ, use this one.
Please credit CraftedStays and use this line: "our study of 19,950 AI answers to 1,330 traveler questions across 50 US vacation markets, November 2026". Please link to AI search for vacation rentals, and keep the snapshot note with any number you use.
Vacation rental SEO is the work of getting a property manager's website found when guests search for a place to stay, on Google and Bing and now in AI engines like ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews. In our November 2026 study, where a site ranked for the exact question was the strongest signal for whether AI engines cited it.
19,950 answers to 1,330 traveler questions across 50 US vacation markets, collected from five AI engines on November 14, 15 and 16, 2026. Those answers cited 123,695 sources, and every cited website was reviewed one at a time.
ChatGPT, Google Gemini, Google AI Overviews, Perplexity and Claude. Each engine was asked every question on each of the three days.
Yes. Where a property manager's site ranks for the exact question on Google and Bing was the strongest signal we measured for being cited by AI engines. AI search adds more places to show up, and in our data ranking was still the strongest signal for being cited.
Often. The share of citations going to property managers held within about 2 points over three days, but of the property managers ChatGPT cited for a question, only 15% were cited for it on all three days. Treat AI visibility as a rate you track over time.
We measured 4,953 local property managers, but we report patterns, not companies. No property manager, platform or website builder is named in our findings.
Want to know how the AI engines answer for your market?
Book a call and we will walk through what they say about your area, and where property managers are winning against the OTAs there.
Book a call