Vacation rental marketing has run on one rule for fifteen years: earn links, build domain authority, climb the rankings. We went looking for proof that the rule still holds in AI search. It does not.
We looked at thousands of AI answers across the major engines, covering 20 vacation rental markets. We wanted to know why an AI names one property manager and skips the next. We expected backlinks to matter less here. We did not expect them to point the wrong way. The sites AI cited had fewer backlinks than the sites it skipped, about 20 percent fewer. Their domain rank was no better either. What separated them is something the SEO industry barely measures.
The short answer: backlinks do not get vacation rentals cited in AI search
Every link signal we tested came out flat or slightly negative. Cited sites had about 20 percent fewer linking domains, and about 20 percent fewer backlinks in total. Domain rank, schema markup, and homepage length were identical between the groups.
One thing did separate them, and it was local content, in volume. Sites that got cited had 13 times as many area guides as the sites that did not. They also had 3.5 times as many guest pages, meaning pages built for families, for groups, for luxury travelers, and for each type of property.
We repeated the entire test on a much bigger group of sites and got the same answer. Guest pages, 9 times more. Area guides, 5 times more. Total pages, 2.6 times more. Schema was identical again.
AI search is not a tougher version of SEO. It is a different game with a different scorecard. As far as we are concerned, that should end the backlink conversation for AEO.
AI search is already picking property managers
This is not a problem you can schedule for next year. AI Overviews now show up on about 48 percent of Google searches, up from 31 percent in February 2025. That comes from BrightEdge data, reported by Search Engine Journal. Travel is moving faster than almost any other category. In a single year, travel’s share of the pages cited in AI Overviews went from 5.7 percent to 17.7 percent.
The cost of being left out shows up immediately in clicks. Seer Interactive tracked 3,119 searches across 42 companies. When an AI Overview appeared, clicks to the normal results fell 61 percent. But the brands named inside that answer got 35 percent more clicks than the brands left out.
Here is the part the industry keeps missing. Property managers are not locked out of these answers at all. In our data they show up in about 9 of every 10 AI travel answers, and they take about 42 percent of all citations. That is property managers as a group, though, not any single company. In a typical market, 80 to 150 different operators split those citations, so any one of them lands in only a few percent of answers. AI already recommends property managers by name. The only open question is which ones.
What we found: the cited vacation rental sites had fewer backlinks
We started with a set of property manager sites where we already knew the outcome. Some got cited by AI and some did not. We pulled link data on all of them and compared the two groups.
Then we ran it again on a much larger set, this time 500 cited sites and 681 that were never cited. Four in ten sites in that group got cited, so it was a fair fight.
Here is how the link signals scored:
- Linking domains: cited sites had about 20 percent fewer. No help.
- Total backlinks: same story, about 20 percent fewer.
- Domain rank: identical. No measurable gap between the groups.
- Schema markup: identical. Both groups carry it at the same rate.
- Homepage word count: identical. Depth on a single page is not the lever.
Here is where it gets uncomfortable for the national brands. We ran the 20 largest US property management firms through the same local questions, and only 7 showed up at all. Most of those showed up as booking marketplaces rather than as operators.
The firm with the most linking domains in our entire data set was cited nowhere. Another firm runs tens of thousands of units and publishes more than a thousand guest pages, and it was cited nowhere either. Meanwhile a regional operator with a tiny fraction of that authority was cited.
All that authority, and they still lose local answers to operators a fraction of their size. Money cannot buy the one signal that matters here.
What actually gets vacation rental sites cited: local content
The useful news is on the opposite side of that finding. Sites that get cited run a deep library of local pages built around what guests actually want. Not a five page brochure with a booking button, but real coverage of the market.
We looked at the pages AI actually cited. Here is what they were:
- Property type pages, like cabins, cottages, and villas: 29 percent
- Area guides and things to do: 14 percent
- Group and large group pages: 13 percent
- Location and landmark pages: 13 percent
- Pricing and deals pages: 0 percent
The more local guides a site had, the more often AI cited it
We sorted the sites by how many area guides they run, then checked how often each group got cited. The line climbs at every step:
- 0 to 1 guides: 32 percent got cited
- 1 to 5 guides: 38 percent
- 5 to 15 guides: 48 percent
- 15 to 40 guides: 52 percent
- 40 or more guides: 60 percent
Now run the same test on backlinks. The line does not climb at all. It sits flat and then it drops, because the sites with the most links are the national brands, and they get cited least in local answers. A real driver climbs the way the guide line climbs. Links do not. That is not a weak signal. That is the wrong signal.
It is the content, not just a bigger website
Here is the obvious pushback. Maybe guides just mean a bigger site, and bigger sites win anyway. So we split the sites by size and compared them inside each size group:
- Small sites: 34 percent got cited without guides, 43 percent with them.
- Mid-size sites: 30 percent without, 52 percent with.
- Large sites: 26 percent without, 57 percent with.
Local content roughly doubles the odds inside every size group, and it helps the biggest sites most. This is the result we tried hardest to break, and it held. Size alone does not get you cited. Local depth does.
Content gets you in the door, then stops paying
One finding here will save you money. Look only at the sites that already get cited, and content volume stops predicting anything. For that group, the link between how much they publish and how often they get named is basically zero.
So content carries you over the line from invisible to visible. After that, how often you get named comes down to how crowded your market is. Build a real local library, then stop counting pages and start making them better.
Other researchers found the same thing
This is not one company’s hot take. Three separate research groups have independently landed in the same place.
- Ahrefs studied about 75,000 brands. Backlinks barely tracked with AI citations, scoring 0.218 on a scale where 1.0 is a perfect match. Brand mentions around the web scored 0.664, three times higher. Ahrefs also found that only 37.9 percent of the pages cited in AI Overviews rank in Google’s top ten, so most citations skip the ranking game entirely.
- The Princeton GEO paper (KDD 2024) coined the term generative engine optimization. The team tested about 10,000 questions. Adding statistics to a page lifted its odds of making the AI answer by about 40 percent, and citing solid sources did the same. Keyword stuffing, the old SEO trick, did nothing or actively hurt.
- A controlled arXiv study kept the words on a page identical and changed only the layout. Same claims, same facts, new structure. Citations rose 17.3 percent across six AI engines. That is cause and effect rather than a guess, which makes it the strongest outside proof that page shape drives citations.
A pattern is forming across all of this work. Getting mentioned around the web tracks with AI visibility, while getting linked to does not, or barely does. Our research adds the vacation rental version of that, plus the part most of these studies leave out: what to build instead.
The best argument against us, and our answer
A sharp SEO will push back here, and the pushback is a good one. It deserves a straight answer rather than a dodge.
Some AI tools do not read the open web on their own. Perplexity and Google’s AI Overviews pull candidate pages from a search index first, then write the answer. That index does count backlinks, so authority can sneak in through the back door. One widely shared study using Otterly.AI data puts the figure at 78 percent, meaning the share of AI sources that come from sites with a Domain Rating above 60. That is the sharpest counterpoint out there.
Two things keep it from changing the verdict:
- Most citations do not come from the top of Google. Ahrefs put that number at 37.9 percent. If backlinks were the gate, it would be far higher.
- Our content effect holds when authority is held steady. This is the key test. Inside every size group, and so across similar link profiles, local content still doubled the odds. If borrowed authority were the real driver, that gap would disappear once you compare like with like. It gets wider instead.
So the honest claim is not that links never matter anywhere. It is narrower than that. To get named inside an AI answer in your market, links are not the lever. Local content is.
Backlinks still matter for Google. That is a different scoreboard.
We are not saying link building is dead. Studies of millions of search results still show the top Google result carrying far more links than the ones below it. Classic ranking still rewards authority.
The point is that you now play two games at once, and they score differently:
- Classic search: ten blue links in a row. Links and authority move you up, and position is the prize.
- AI search: one written answer that names a few sources. Local coverage and clear page structure move you up. There is no second page to settle for. No Page 2.
Give AI search its own budget and its own scorecard instead of bolting it onto SEO. A property manager who ports the link playbook straight across can spend a year buying links and never get named once.
What to build instead of buying backlinks
Redirect the link budget here. This priority order comes straight from what separated the cited sites:
- Guest type landing pages. Cabins for families, homes for big groups, luxury places with a view, all in your market. These showed the widest gap of all at 9 times more on cited sites, and they win the questions guests really ask.
- Property type pages. The biggest slice of cited pages at 29 percent. Build one page per type per market, with real homes on it.
- Area guides. The strongest site-wide signal we found. The guide itself rarely gets cited, but a site with a real guide library becomes the local source AI pulls from.
- Location and landmark pages. Near the lake, near the park, walk to downtown. These made up 13 percent of cited pages.
- Review markup, not just review stars. AggregateRating code showed up on 16 percent of cited pages and 0 percent of the rest. Pages that never got cited never had it.
- Show up where locals talk. Reddit and forum presence was a smaller but real signal, and some engines lean hard on community sources.
Skip these: single amenity pages, pricing pages, VacationRental schema crusades, and generic national content that never names your market.
What a citable vacation rental page looks like
Breadth gets your site in the running, and page structure gets you quoted. We compared pages that got cited against pages on the same sites that did not. Four things stood out:
- Lots of sections. Cited pages had 13 headings and the pages that missed had 1. This was the biggest difference we measured at the page level. A long wall of text loses, even when it is well written.
- An FAQ block. On 28 percent of cited pages, against 2 percent of the rest.
- Review markup. On 16 percent of cited pages, against 0 percent of the rest.
- The place name in the headings. On 22 percent against 10 percent, and on guides it was 44 percent against 4 percent.
The cited pages ran about 1,800 words with about 10 headings, and nine in ten of them used lists. They averaged roughly one concrete fact every 100 words. Write each section so it completely answers one question and can be lifted out of the page on its own. That is the entire mechanic.
The next wave of SEO
Read this research one way and it stings, because ten years of link building buys a property manager very little in AI search. Read it the other way and it is the best news in a decade. The old game rewarded whoever had the biggest budget and the longest head start. This one rewards whoever knows their market best.
You cannot conjure up 2,000 linking domains this quarter. You can publish a deep library of local pages, built for real guests and shaped to be quoted. That is a fair fight, and most property managers are better set up to win it than the national brands are. We will defend that finding anywhere. The engines are already naming property managers by the thousands, so build the pages that make yours the obvious one to name.
The next wave of SEO. Same goal, new rules. Get discovered. Convert the right guests.
How CraftedStays approaches this
CraftedStays builds direct booking sites for property managers, with Aria as the AI inside the platform. The Sightline Report is the playbook Aria builds for your market, and it is built on the research above: which guest types you can win, which pages to build first, and how to shape them so AI can pull answers out. You can have a website that degrades over time, or one that evolves and matures over time.
For the full playbook that turns this research into pages you can build, see the property manager’s guide to AI search for vacation rentals.
FAQ
Do backlinks help my vacation rental site get cited in AI search?
Not really. We analyzed thousands of AI answers across 20 vacation rental markets. The sites AI cited had about 20 percent fewer linking domains and backlinks than the sites it skipped, and domain rank was identical. Ahrefs found the same pattern at a much bigger scale. Backlinks tracked with AI citations at just 0.218, while brand mentions around the web scored 0.664.
Is AEO the same thing as SEO for vacation rentals?
No. Both aim to get you found, but they score differently. Classic SEO rewards links and rank position. AI search rewards local content and page structure that is easy to quote. Only 37.9 percent of the pages cited in AI Overviews rank in Google’s top ten, so ranking well is not the same as getting named.
Should I stop doing SEO for my vacation rental website?
No. Normal search still drives real bookings, and it still rewards links. Treat AI search as a second scoreboard with its own plays. The mistake is to assume your existing SEO work buys AI visibility for free. It does not.
How much local content does a property manager need before AI starts citing the site?
Sites with almost no guides got cited 32 percent of the time, and sites with 40 or more got cited 60 percent of the time. The gain gets real in the 5 to 15 range, where the rate hit 48 percent. Content works like a door rather than a dial. Among sites that already get cited, adding more pages did not raise how often they were named.
Does VacationRental schema get me cited in AI search?
It did not separate the winners from the rest in our research, since both groups carried schema at identical rates. The markup that did matter was AggregateRating on your reviews. It showed up on 16 percent of cited pages and 0 percent of the rest. Location markup like GeoCoordinates helped too.
Can a small property manager beat a national brand in local AI search?
Yes, and it happens all the time. Only 7 of the 20 largest US property management firms showed up at all in local AI answers, and the firm with the biggest link profile in our data was cited nowhere. National brands publish generic content at scale. AI answers about one market reward local detail, and that is where a focused operator wins.
How do I know whether AI engines are citing my vacation rental site?
Ask the engines the questions your guests ask, in your market, and write down who gets named. Run each question a few times and on more than one engine. Results shift between runs, and engines favor different sources. Track how often you show up, on which questions, and which of your pages they link to.
What should I build first if I only have time for one thing?
Build guest type landing pages for the travelers you already serve best: families, big groups, luxury guests, in your market. These showed the widest gap between cited and skipped sites at 9 times more on cited sites. They win the simple one and two requirement questions where property managers beat the big booking sites, and they convert better once a guest lands.