Two costs, only one of them obvious
Run the default season — 90 nights, 26 requests, a length mix weighted towards two and three nights — at a three-night minimum and 66 nights book across 11 stays. Nine requests are refused for being too short, which is the cost everybody expects. The second cost is quieter: five empty nights sit in three separate holes that are shorter than three nights, so nobody is permitted to book them at any price. Those nights are not unsold because demand failed. They are unsold because of the rule.
| Minimum | Bookings | Nights | Dead gap nights | Refused, too short | Net |
|---|---|---|---|---|---|
| 1 | 15 | 72 | 0 | 0 | $9,584 |
| 2 | 13 | 70 | 3 | 2 | $9,372 |
| 3 | 11 | 66 | 5 | 9 | $8,880 |
| 4 | 10 | 63 | 11 | 15 | $8,493 |
| 5 | 8 | 55 | 5 | 18 | $7,440 |
| 6 | 5 | 40 | 1 | 21 | $5,439 |
| 7 | 4 | 34 | 0 | 22 | $4,632 |
Read the dead-gap column downwards: 0, 3, 5, 11, 5, 1, 0. It is not a line. A four-night minimum strands eleven nights on this calendar and a five-night minimum strands five, because whether a leftover hole is one night wide or four depends on where the bookings happened to land, not on how big the rule is. The column falls back to zero at seven only because so little books that the remaining gaps are all long.
Sometimes the minimum wins, and the page will tell you when
On the default numbers a one-night minimum nets $704 more over the season than the three-night rule, which is the opposite of the usual advice. Change the economics and it flips. Put the nightly rate at $40, charge no cleaning fee, make a turnover cost $75 and cut the requests to ten, and the three-night minimum nets $570 against $515 for taking everything — because at those numbers a one-night booking loses money and refusing it is the point. The crossover is real and it sits at a different place for every listing, which is why the page prints the whole column rather than a recommendation.
What the model does and does not do
Requests are placed on an empty calendar with start dates spread evenly across the period, taken first come first served, and refused if they are shorter than the minimum. Real demand clusters on weekends, and clustering makes the gap problem worse rather than better, so the dead-night count here is the optimistic end. The model also cannot know whether a guest refused at three nights books four instead or books somewhere else, and that single unknown is worth more than any of the arithmetic. A month of your own enquiry threads will answer it and nothing else will.
Questions people ask
What is a gap night?
A run of empty nights between two bookings that is shorter than your own minimum stay, so no guest is allowed to book it. With a three-night minimum a two-night hole earns nothing no matter how many people want it. On the default season the three-night rule leaves five such nights in three separate holes, worth $701 of nightly revenue if anybody had been permitted to take them.
Does a higher minimum stay always make more money?
No, and neither does a lower one. On the default numbers a one-night minimum nets $704 more over 90 nights than a three-night rule. Change the rate to $40, drop the cleaning fee, put a turnover at $75 and thin the demand, and the three-night minimum wins instead at $570 against $515. The direction depends on whether your turnover cost is large next to a night of revenue, which is why the page runs all seven minimums rather than recommending one.
Why do the dead gap nights go up and down instead of climbing?
Because a gap becomes unsellable the moment it is one night shorter than the minimum, and gap sizes are set by where bookings happen to fall. On the default calendar the count runs 0, 3, 5, 11, 5, 1, 0 as the minimum goes from one to seven. The jump at four is not a bigger rule doing more damage in proportion — it is three particular holes landing at three nights instead of four.
Should I read the length mix off my bookings?
Off your enquiries, not your bookings. Your current minimum has already deleted the short requests from your booking history, so reading the mix from bookings hides exactly the demand the page is trying to price. Count the enquiries you turned away as well as the ones you took, over a season rather than a month.