The number nobody can source
There is a figure circulating in hospitality at the moment. It says that improving your visibility in AI search will lift your direct bookings by somewhere between 12 and 22 percent.
We tried to trace it. We could not find a primary source.
That does not make it wrong. It makes it unusable. Nobody should be signing off a budget on a statistic with no methodology behind it, and we are not going to be the ones repeating it.
So we asked a different question. Instead of guessing how much AI visibility lifts direct bookings, we asked the question you can actually answer with arithmetic:
If it lifts them at all, what is that worth?
That turned out to be the interesting part. Because almost every estimate we looked at gets the arithmetic wrong, and they all get it wrong in the same direction.
The mistake everyone makes
Here is the error. It is easy to make and it inflates everything downstream.
Say a lodge picks up 20 extra direct bookings in a month. The obvious move is to multiply 20 by the room rate and call that the value created.
That is wrong, and it is wrong by a lot.
Most of those bookings are not new business. They are bookings the lodge was going to receive anyway, arriving through its own website instead of through an OTA. The guest was always coming. The booking simply took a different route.
So the lodge does not gain the room rate. It gains the commission it no longer pays.
On a room selling at 700 dollars with an 18 percent commission, that is 126 dollars. Not 700.
Only the genuinely new bookings, the ones that would not have happened at all, are worth the full rate.
Count every booking at the full rate and you overstate the value by roughly two and a half times. Make the second common error as well, applying the uplift to all of your bookings rather than only your direct ones, and you can be out by a factor of seven.
We have seen both errors in the same calculation more than once.
The model
Here is ours. Every assumption is visible, so you can disagree with any of them and rerun it yourself.
| Input | Value we used | Why |
|---|---|---|
| Direct booking share | 30% | Typical for independent properties |
| Split of uplift: channel shift | 70% | Bookings that would have happened anyway |
| Split of uplift: new demand | 30% | Bookings that would not have happened at all |
| OTA commission | 18% | Mid-point of the common 15 to 25 percent range |
| Occupancy | 60% | Used only to estimate annual revenue |
And the arithmetic:
Annual room revenue = keys x nightly rate x 365 x occupancy Direct bookings = annual revenue x direct share Uplift = direct bookings x scenario Value from shift = uplift x 70% x commission rate Value from new demand = uplift x 30% x full rate Total annual value = shift + new demand
Which collapses to something you can carry in your head:
Every 10 percent increase in direct bookings is worth roughly 1.3 percent of a property's annual room revenue.
What that looks like in money
We have not claimed an uplift figure, because we cannot source one. Instead, here is what each scenario would be worth if it happened.
Small tented camp
12 keys at 1,200 dollars a night. Roughly 3.2 million dollars in annual room revenue.
| Uplift in direct bookings | Annual value |
|---|---|
| 5% | about $20,000 |
| 10% | about $40,000 |
| 15% | about $60,000 |
| 20% | about $81,000 |
Mid sized safari lodge
25 keys at 700 dollars. Roughly 3.8 million in annual room revenue.
| Uplift in direct bookings | Annual value |
|---|---|
| 5% | about $24,000 |
| 10% | about $49,000 |
| 15% | about $73,000 |
| 20% | about $98,000 |
Larger boutique property
45 keys at 550 dollars. Roughly 5.4 million in annual room revenue.
| Uplift in direct bookings | Annual value |
|---|---|
| 5% | about $35,000 |
| 10% | about $69,000 |
| 15% | about $104,000 |
| 20% | about $139,000 |
Notice that the small expensive camp and the larger cheaper property land in broadly similar territory. In luxury hospitality, exclusivity and rate move in opposite directions. Keys multiplied by rate is what matters, not either one on its own.
Work it out for your own property
| Uplift in direct bookings | Annual value |
|---|---|
| 5% | $24,490 |
| 10% | $48,979 |
| 15% | $73,469 |
| 20% | $97,959 |
At a 10 percent uplift, that is worth roughly $48,979 a year, or 1.3 percent of your room revenue.
Change any assumption you disagree with. If you think your direct share is higher than 30 percent, or your commission rate is worse than 18, put your own numbers in. The model does not care whose numbers they are.
What we cannot tell you
We want to be straight about the limits of this.
This is a model, not a measured result. We have not published outcome data from live engagements, because we do not have enough of it yet. When we do, we will publish that instead and you should trust it more than this.
We are not claiming any specific uplift. That is why the table runs on scenarios. If somebody tells you AI visibility will lift your direct bookings by a precise figure, ask them where the number came from.
The 70 to 30 split between channel shift and new demand is an assumption. It is the weakest part of the model and we have not found a good source for it. At a more conservative 90 to 10 split, the value drops by around a third. At 50 to 50 it rises by about half. The calculator lets you test both.
Seasonality is not in here. Safari markets are heavily seasonal and an annual average hides a lot.
So what does it mean
Three things follow from this, and none of them require you to hire anybody.
One. The value is real but it is not enormous, and that is fine. A mid sized lodge is looking at somewhere between 24,000 and 98,000 dollars a year depending on how well the work goes. That is a meaningful number. It is not a transformation. Be suspicious of anyone describing it as one.
Two. Your rate and your size decide whether this is worth doing at all. Below roughly two million dollars in annual room revenue, the arithmetic stops supporting an ongoing engagement with anyone. Find out where you sit before you spend anything.
Three. The defensible number is the commission you stop paying. It is the smaller of the two figures, and it is the one that comes straight off your own books. Revenue growth is contestable. A commission line is not.
If you want to know where your property currently stands in AI search, we run a free Snapshot. It takes about five days and there is nothing attached to it.
Study 01. Published 20 September 2026. Written by Nigel Amoro. Veenoa publishes its workings. If a study is worth citing, it should be possible to check.