The Shortlist Effect: She found you through AI. She won't book through it. Yet.
Only 5% act on an AI recommendation without checking first. Half go straight to the website, and until the agent takes her card, that booking is yours to win.
by Markus Busch
In 2009, Chris Anderson of Cornell's Center for Hospitality Research listed four JHM hotels on Expedia, took them off, then listed them again. In the weeks they were visible, reservations through the hotels' own channels rose by 7.5 to 26 percent, on top of whatever Expedia sold itself. He called it the billboard effect. When he and Saram Han came back to it in 2017, they found it weaker but still alive, for one reason: many travelers still look at an OTA before they book. The billboard was a visit.
That visit now starts somewhere else. A traveler asks ChatGPT for a quiet hotel by the sea and gets three names. Then she goes to check, and one of the places she goes is the hotel's own website. For an independent, it is the one page on her whole search where no other hotel sits next to yours. When that visit ends in a direct booking, we call it the Shortlist Effect.
The billboard effect was how the industry made peace with the commission. The OTA billed for the bookings it took and showed your hotel to everyone else for free. The Shortlist Effect comes with no bill at all.
It does come with an end date. The window closes when she trusts the AI enough to let it pay.
Only 5% act on the AI's answer alone
Yext surveyed 3,848 adults in the US, the UK, France, Germany and Italy this spring about what they do after an AI recommends a local business. Only about 5% acted without checking anything. Fifty-three percent searched Google or Bing. Forty-nine percent went straight to the business's website. Trust barely moved the number: in the US sample, people who rated their trust in AI five out of five checked Google at 62%, neutral users at 63%.
Yext sells listings and AI visibility, so it has a stake in finding that the website still matters. The survey covers local businesses in general, not hotels.
Adobe's traffic data points the same way for travel. In July, visits from AI tools to US travel sites were up 119% on a year earlier. Those visitors were 21% more engaged than other traffic and converted only 1% less, where a year earlier the gap had been 47%. They arrive ready to read the page closely.
A brand vouches for itself. An independent has a website.
Anderson's studies looked at guests of major hotel brands. A guest looking at a Courtyard knows roughly what she will get before she clicks. The brand has done most of the checking for her.
An independent has no such promise working on its behalf. The AI described you in a sentence: sea view, good breakfast, quiet at night. Your website is where she finds out whether that sentence is true, and what it costs.
Here the independent holds an advantage it rarely gets. On an OTA, she compares you against twenty other hotels in a grid sorted by price, with a Genius badge on some of them. On your own site, she sees only you. Your rooms. Your rate. If the page confirms what the AI said and the price makes sense, she has little reason to go back and look again.
A lost visit comes back at 15 percent or more
If the page doesn't convince her, she doesn't disappear. She goes back to the AI, or to Booking.com, and books the same room there.
Same guest. Same dates. Same room she was looking at on your site ten minutes earlier. Now the booking carries a commission that on Booking.com starts at 15 percent and climbs from there for hotels in its visibility programs. The AI made the introduction for free, and the hotel now pays commission on it.
It lasts until she lets the agent pay
In July, Lighthouse, which sells hotel software, put AI tools at just under 1% of visits to hotel websites, after a jump of more than 50% in a few weeks. Small, and growing fast.
The visit survives because paying inside the AI is still rare and still distrusted. That is being built away. In late August, Google switched on hotel booking inside AI Mode in the US, with Google Pay at the end, and every step that lets her pay inside the answer makes the trip to your website optional. We have written about what happens on the other side of that line: she accepts a hotel instead of choosing one, and the commission goes to whoever held her card.
Until then, she comes to you.
Your first screen sits next to Booking.com's
Ask ChatGPT for a hotel in your town the way a guest would, with dates, a budget and one thing she cares about. If your hotel is named, write down the sentence the AI used to describe it. Tap through on your phone and read your first screen against that sentence.
Then open the same dates on Booking.com and put the two screens side by side. Whichever one makes the case faster is the one she books on.
Nobody can count the Shortlist Effect yet
Anderson needed Expedia's cooperation to switch listings on and off before the original effect showed up in the numbers. No one can switch off ChatGPT, and no property can yet say how many of these visits end in a booking. The Yext and Adobe figures come from companies with commercial interests and cover more than hotels. Booking.com commission varies by contract and program; the invoice is the number that counts.
Read also: Travelers trust AI enough to start. Not enough to book · She didn't choose your hotel. She accepted it. · AI recommends the hotel it knows best, not the best one
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 →