Friday is not Sunday.

In SiteMinder’s 2025 figures for US hotels, one-night stays were 72.85 percent of bookings. The year before, they were 72.45 percent. The average stay was 1.61 nights. The usual booking is therefore a single night, and that share rose a little. It did not fall.

That first fact changes how the second should be read. If nearly three bookings in four are for one night, nearly three guests in four pay for one particular night. They do not pay for “the week.” They pay for a Friday, or a Sunday, or whichever day they chose.

Those nights do not cost the same. In that set, Friday’s average rate was $286.57. Sunday’s was $223.54. They are two nights in the same week, at the same hotels, at different rates. The gap does not need to be turned into another percentage. It is enough to see that they are not the same price.

A weekly average joins those nights into one figure. The Friday guest and the Sunday guest did not pay that figure. Nobody booked it. The average looks steadier on a report precisely because it has erased the night. It describes a day that does not exist.

That is why the reason beside the rate has to name the night. “Friday, and this is why” can be checked against what is known about that Friday: demand, events, what was left unsold. “The week, and this is why” cannot be checked against anything specific, because the week is not the unit the guest bought.

The average stay of 1.61 nights says the same thing by another route. It is close to one night and far from a week. Using the week as the unit of price mixes two different decisions and presents the result as if it were one.

For that reason, it is a poor idea to set or explain the rate from a weekly average. The usual booking is one night, and Friday and Sunday in that set did not cost the same. Name the day next to the price. If a report only shows “the week,” ask for the breakdown by night before you decide.