The AI booking scare runs on a number that isn't about hotels
The conversion figure quoted to prove AI is rewiring hotel bookings was measured across every industry except hotels
Every panel deck this year carries the same line: AI referral traffic converts at 11.4%, more than double organic search at 5.3%. It gets quoted to prove the booking journey has already moved into the chat window, and that your hotel had better pay to be visible there before it's too late. The number is real. It just isn't a hotel number.
The number that isn't yours. That 11.4-against-5.3 figure is Similarweb's aggregate across the open web — a blend dominated by e-commerce, by Amazon and Reddit and the kind of one-tap checkout a hotel booking never is. Look for the travel cut and it sits near the bottom of the range: travel referral traffic converts well below the all-industry average, because a stay is researched over weeks, not grabbed in a session. Quoting the blended number to a room full of hoteliers isn't evidence the funnel collapsed. It's a category error with a decimal point.
The pattern predates AI. The baseline everyone says AI shattered was already made of folklore. The "38 sites before booking" stat is not Google's — it's an Expedia clickstream study from 2013, and it measured a narrow slice: travelers who'd hit a tourism-board site in the 45 days before buying a package. The "7 hours, 11 touchpoints, 4 locations" rule quoted as gospel has no verifiable Google research under it at all; even the marketers who repeat it can't point to the origin. The funnel AI supposedly broke was drawn in numbers nobody had checked.
What the number is for. A borrowed statistic does a job. Scale it up, drop the caveat, attach a credible research name, and a conversion multiple stops being data and becomes a reason to buy — an AI-visibility audit, a subscription, a seat in whatever auction the discovery layer is turning into. The scare has a supply chain, and misattribution is a link in it. The louder the number, the better the odds it was measured somewhere other than your business.
The stake. AI is genuinely changing how hotels get found; the research mix really is shifting. But that case keeps getting built on the wrong evidence, and wrong evidence sends money in the wrong direction. Before a number moves your budget, check the denominator: what industry it was measured in, who actually published it, and whether a hotel booking behaves anything like the average it's folded into. The honest questions here are smaller and duller than the headline — whether the booking engine is fast, the listing accurate, the direct rate worth the click. They don't sell subscriptions. They happen to be true.
Enjoying this analysis? Hospitality.today delivers daily insights on hotel distribution, AI trends, and travel commerce — straight to your inbox. Subscribe for free at Hospitality.today →