We watch every short-term rental calendar in Memphis and record it nightly. When a night flips from open to taken, we see it happen. That means we are not estimating when guests book. We are watching them do it.

Here is what the last 17,920 of those bookings looked like.

HOW LATE PEOPLE ACTUALLY BOOK

Of the bookings that land within three weeks of the stay:

• - 21% are made the same day the guest arrives
• - 38% within two days
• - 50% within four days
• - 65% within a week

The median is four days. Not four weeks. Four days.

If you have ever stared at a soft calendar ten days out and started cutting your rate, this is the number you were missing. Most of that business had not been booked yet by anyone, for anyone.

THE PART THAT ACTUALLY CHANGES DECISIONS

The average hides something better. Break the same data out by which night of the week the guest is staying, and the market splits in two.

Share of nights booked the same day the guest arrives:

• - Monday 26.4%
• - Tuesday 24.8%
• - Sunday 23.3%
• - Wednesday 22.1%
• - Thursday 21.8%
• - Friday 16.0%
• - Saturday 12.3%

A Monday night in Memphis is more than twice as likely to be booked same-day as a Saturday night.

That makes sense once you say it out loud. Saturdays get planned. Somebody decides in advance that they are coming for a wedding, a game, a weekend downtown, and they book it. Mondays get needed. A flight got cancelled, a job ran long, a crew is in town for the week, a closing slipped. Nobody plans a Monday in Memphis three weeks out.

So the same empty square on your calendar means two completely different things depending on where it sits in the week.

An empty Saturday two weeks out is a warning. Your weekend demand already came through and picked somebody else.

An empty Monday two days out is Tuesday's problem, not today's. It is exactly on schedule.

WHAT I WOULD DO WITH IT

Stop reading your whole calendar with one set of eyes. A soft midweek two days out is normal and does not justify a price cut. A soft weekend two weeks out is real information and it is worth acting on while you still have time.

And if you are going to discount, discount the night that has already missed its window, not the one whose window has not opened yet. Most people do the opposite, because they look at the calendar as a single block and panic at the total.

THE OTHER PART

The guest who books sixty days out is not in this data. The late window is where operating decisions get made, which is why it is the one we can speak to, but the number is 21% of late bookings, not 21% of all bookings.

Two other things worth stating. We count a booking when a night goes from open to taken on a calendar we checked the night before, so a night that was blocked by the host rather than sold looks the same from outside. And we only used stay dates where we had watched the entire run-up, because counting partial windows would make late bookings look even more dominant than they are.

Happy to run the same split for a specific submarket or bedroom count. Post the ask in the thread and I will put the numbers up.