Direct Bookings Can Be More Profitable Than OTA Bookings
The usual argument for pushing direct bookings is the commission saved versus an OTA. Less discussed is what happens after the booking is made - specifically, how often it cancels. Two separate cancellation studies let us actually check whether direct bookings carry a meaningfully lower cancellation risk than OTA bookings, and whether that translates into anything useful about how reliable the demand on your calendar really is.
Two separate studies, the same ranking
One cancellation study tracking channels from 2024 to 2025 shows the direct booking engine improving from 9.7% to 9.2% cancelled, the lowest of any channel measured, while Booking.com stayed highest even after improving, from 27.0% down to 24.1% - still close to one in four reservations. A second, separate study puts direct at 14.4% of bookings cancelled and 17.1% of cancelled revenue, against 26.9% of bookings and 37.4% of revenue for the Booking Group, with Expedia Group sitting in between at 21.90% of bookings and 30.10% of revenue. The absolute percentages differ between the two studies, which is worth being honest about, but the ranking doesn't move - direct comes out lowest in both, and the largest OTAs come out highest in both.
The revenue lost to cancellations is worse than the booking count suggests
In both studies, the share of cancelled revenue runs higher than the share of cancelled bookings, meaning the bookings that cancel tend to be worth more than the ones that don't. That gap is small for direct - 14.4% of bookings versus 17.1% of revenue, a 2.7 point spread. For the Booking Group it's much wider - 26.9% of bookings versus 37.4% of revenue, a 10.5 point spread. So the commission an operator pays an OTA to acquire a booking isn't the only cost sitting on that channel - a disproportionate share of the revenue booked through it is also the revenue most likely to disappear.
What this means for demand reliability is only partly answered
The part of this question about more reliable demand signals is harder to settle with what's here. One finding shows that reservations made far in advance carry a disproportionately higher cancellation probability across all channels, which means the forward occupancy shown in a booking pace report is significantly inflated - but this finding isn't broken down by channel. It tells us that distant bookings are less trustworthy as a demand signal generally, not that direct bookings specifically are more trustworthy than OTA bookings at the same lead time. Given direct's lower overall cancellation rate, it's reasonable to expect its calendar is less distorted by this effect than an OTA-heavy one, but that's an inference on my part, not something directly measured here.
What this means at property level
For an operator looking at channel mix, the acquisition cost savings on direct bookings are the easier number to point to, but the lower cancellation rate is a real, separately measured advantage on top of that - fewer last-minute gaps to refill, and less of your booked revenue evaporating before arrival. How much that matters depends on how much of a property's volume actually runs through each channel and how far out those OTA bookings are typically made, since the occupancy-inflation issue applies to any channel with a long booking window, not just OTAs.
Roger's take
Two independent studies point in the same direction: direct bookings have the lowest cancellation rates, while the largest OTAs have the highest. The numbers differ between the studies, but the pattern is consistent. The booking-window data tells us something different. It shows how far in advance guests book, but it doesn't tell us whether direct demand is more reliable than OTA demand at the same lead time. So we can be confident about the cancellation difference, but we shouldn't claim that direct demand is inherently more reliable without more evidence.
Conclusion
The lower-cancellation finding holds up well: two separate studies show direct bookings consistently have the lowest cancellation rates, while the largest OTAs have the highest. The evidence also shows that OTA cancellations can be concentrated in higher-value bookings. The more-reliable-demand argument is less certain. The data suggests it, but neither study directly measures demand reliability by channel. So that part should be treated as an opportunity to investigate, not as a proven fact.