AI Search Is Breaking Hotel Booking Attribution

Hotel marketers are losing the ability to explain how a booking happened, even though demand for rooms is not shrinking. That is the argument in a Hospitality Net opinion piece by Martin Soler of Soler & Associates, who says zero-click search and AI answers that never send travelers to a hotel website are erasing the trail that tracking once followed.

The scale of the blind spot is still small, but the direction is clear. Lighthouse puts AI tools at just under 1 percent of hotel website visits, according to a Hospitality Net piece on the Shortlist Effect. A Yext survey of 3,848 adults in five countries found about 5 percent acted on an AI recommendation without checking further, while roughly 49 percent went to the business’s own website. That second group is the one a last-click report mislabels: the guest discovered the hotel in an AI answer, then arrived as a direct or branded-search visit.

For an operator, the P&L risk is misallocated spend. If direct bookings that began in an AI assistant are credited to branded search or to nothing, the channel that created them looks worthless and gets no budget, while the channels that merely closed the sale look efficient. Direct booking cost per acquisition then gets measured against the wrong denominator.

Soler’s fix is low-tech: ask guests how they heard about the property. He also says marketers should get comfortable with incomplete data and use pattern recognition and correlation to fill the gaps. His piece contains no statistics, so treat it as a framing, not a measurement.

The practical version is a “how did you find us” question at booking or check-in, with AI assistants as an answer option, reviewed monthly next to the PMS source codes. Then pull last quarter’s direct bookings and count how many you could trace to a source. Whatever is left over is the number to watch.