I worked for a sister company. These features were A/B tested, their effect on bookings measured, and the numbers affected the bonuses of POs. They were thus incentivized to try these kinds of annoying tricks, because ultimately they work, or at least they do when their effect is observed only in terms of bookings.

I asked several times if we somehow measured customer retention, as we kept adding stuff I thought would have a negative effect in the long term (a customer would go through the whole booking process, only to never return because the experience was too bad.) We didn't. I guess it's difficult, I have no idea how this kind of metric would be computed.

The checkbox is a nice idea, it could help retain customers on the verge of leaving. It wouldn't work for customers who visit once and get so annoyed they never come back.

Also: a good portion of the traffic was affiliate traffic, like from Google Maps. I guess people booking from a Google Map listing care less since they land on the hotel page ? They won't be looking for hotels through the site UI, so they get fewer annoying pop-ups and messages.