In Brief: New research analyzing 695 hotel searches across three European markets reveals that AI assistants like ChatGPT and Google’s AI Mode recommend hotels based on factors beyond traditional search rankings, with nearly half of AI-recommended hotels not appearing in Google’s top results for the same queries.
-
Image Credit Tharro
The process of finding hotels online is shifting as travelers increasingly turn to AI-powered assistants instead of traditional search engines. Rather than scrolling through long lists of search results, users now ask AI tools such as ChatGPT or Google’s AI Mode for hotel recommendations in plain language. These assistants typically provide a short list of four or five hotels, and most users do not look beyond these initial suggestions. This change is making “AI hotel visibility”—whether a hotel appears in AI-generated answers—an important new metric for the hospitality industry.
Understanding AI Hotel Visibility: Mentions vs. Citations
The study distinguishes between two concepts: mentions and citations.
– A mention occurs when a hotel is named in an AI-generated answer, representing a direct recommendation to travelers.
– A citation refers to a source link that the AI assistant includes as the basis for its recommendation.
A hotel can be mentioned without citing its own website, and vice versa. The distinction is crucial for understanding how AI recommendations work and how hotels can influence their visibility.
Study Methodology
The research, conducted by Tharro, analyzed 695 unbranded hotel searches across the Algarve, Mallorca, and Rhodes. Each search was phrased as a question a traveler might ask an AI assistant, such as “family hotels in Mallorca” or “where to stay in Rhodes.” The searches were run through both ChatGPT and Google’s AI Mode, and the results were compared to 293,925 ranked Google and Bing search results for the same queries. The analysis generated nearly 10,000 hotel recommendations, naming 3,626 distinct hotels, and included 14,129 citations across 1,604 domains.
Does Google Ranking Predict AI Recommendations?
The study found that while a strong Google ranking can increase the likelihood of being mentioned by AI, it is not a guarantee. Specifically:
– Hotels with their own website in Google’s top 3 results were named by AI about 66% of the time.
– For hotels in the top 10, the figure was about 61%.
– Hotels ranked 11–20 were named 45% of the time, and those below position 20 were named 27%.
However, nearly half (48–56%) of the hotels recommended by AI did not appear in Google’s organic results for the same search. For example, in Albufeira, only two of the dozen hotels ranking on Google’s first two pages were named by AI. This indicates that AI-generated recommendations often feature hotels that are not visible in traditional search rankings.
Mentions vs. Citations: The Role of Third-Party Sources
The study also found that being highly ranked in Google does not necessarily lead to being cited as a source in AI answers. When AI named a hotel with a top Google ranking, it cited the hotel’s own website as the source only about 20% of the time, and just 37% for hotels in the top 10. Across all recommendations, a hotel’s own domain was cited as the source less than 10% of the time.
Most citations instead went to third-party sources, such as online travel agencies (OTAs) like Booking.com and Expedia, review platforms like Tripadvisor, and editorial guides. This suggests that AI assistants rely more on intermediaries than on hotels’ own websites when assembling recommendations.
Differences Between ChatGPT and Google AI Mode
The two AI engines studied produced notably different results.
– ChatGPT named 3,151 distinct hotels.
– Google’s AI Mode named 1,286.
– Only 691 hotels appeared on both lists, representing just 18% overlap.
Google’s AI Mode tends to use a Maps-driven approach based on prominence and proximity, while ChatGPT relies more on editorial lists and travel guides. As a result, being recommended by one AI engine does not predict being recommended by the other.
Factors Influencing AI Recommendations
Contrary to expectations, guest rating was not the most influential factor in AI recommendations. Instead, review volume and star class were more significant, followed by price. Hotels with a large number of reviews and higher star ratings were more likely to be recommended, regardless of their average guest rating. The data showed that established, upper-tier hotels with moderate scores and many reviews were favored over smaller, highly rated properties with fewer reviews.
Practical Steps to Improve AI Hotel Visibility
Based on the findings, the study suggests several actions for hotels aiming to improve their AI visibility: – Measure AI visibility separately from SEO: Use AI assistants to check if your hotel appears in their recommendations for relevant searches.
– Optimize presence on third-party sources: Ensure accurate and comprehensive listings on OTAs, review platforms, and editorial guides, as these are frequently cited by AI.
– Increase review volume: Focus on generating more guest reviews across multiple platforms, as volume is weighted more heavily than rating.
– Treat each AI engine as a separate channel: Recognize the differing criteria and sources used by ChatGPT and Google’s AI Mode.
– Do not rely solely on your own website: Most AI citations come from third-party sources, not hotel websites.
Conclusion
The research highlights a growing divide between traditional search rankings and AI-generated hotel recommendations. AI assistants use different sources and prioritize factors such as review volume and star class over search ranking and guest rating. Hotels seeking to improve their visibility in AI recommendations should focus on third-party platforms and review generation, and treat AI engines as distinct channels requiring separate strategies. The study provides a framework for measuring and improving AI hotel visibility as this new mode of travel search continues to evolve.
