You made my point. Facebook rolls out changes on a per-user basis. Microsoft often rolls out changes on a per-company basis. You don't roll out changes on a per-access basis. The original commenter was complaining that he had new maps and they would come and go depending on, apparently, the IP address he was using.
You don't want a new feature or UI that randomly disappears and reappears when you hit F5, or disconnect from Wi-Fi, or access the site again after a few hours.
You upgrade based on criteria, and a random probability that increases with time. You roll back because you have identified an issue, but not on a purely random basis.
The examples you gave (Microsoft, Facebook, etc.) do exactly that all of the time. Most of the changes though aren't as apparent as the visual change you are seeing with Maps. There are a ton of examples where Instagram in particular tried out features with limited users and then pulled them later before ether killing the "update" altogether or rolling out to all users at a later date to all users.
The only time we find out about them is when some very observant user notices the minute change and then sometimes (surprise, surprise) complains that they lost the feature as they thought it was a good idea. It happens ALL THE TIME.
The issue here is that this is a very public update and also this very article which we are commenting on incorrectly states that the new area was rolled out. I know they made a correction stating "some" users may not be seeing it, but the fact is that a very small percentage of people are actually seeing the update as has been the case with every other regional pre-release.
When Apple released the initial Northern CA update in the summer of 2018, it was during a beta period and only those on a specific developer beta got the update...yes, everyone on the update. Why they have decided to do it differently since the initial iOS12 beta release is unknown. Would it make sense to do the pre-release to all beta users so it is limited just to them and not everyone else? Maybe, but my guess is that they are doing it the way they are doing it (which is unknown to anyone on these forums) is to probably cover as many devices or software release as possible to make sure there are no issues once the layer is released.
Remember, it's not so much about POI's or building outlines, I think each pre-release is all about the anonymous traffic/routing data innovation they have added to confirm they're own road data that is part of the new Maps. Tracking route/road changes, road closures, etc.
They cannot possibly do an accurate regional test as they do not have access to that data due to privacy reasons (thank you, Apple), so they do a release to a variety of users across the entire country and pull the data they need from the smaller sample that is actually in the region being tested. This is probably done to tens of thousands of users, but some either do not use Maps and the data is still being sent since they didn't opt out of that tracking (most don't), some use Maps but are not aware enough to notice the detail changes in the graphics, and then there are the select few like us who are looking for every little change and see the difference immediately (or check 5 times a day to see if they have new Maps

).
It doesn't matter if someone is in Denver and has them but won't use them or if that same person goes to the new region and loses it. That is just a coincidence IMHO. There was most likely a very large group of people that had it and lost it during the same period. Traveling from one area to another had absolutely nothing to do with "losing it"...and I think that ssame person can confirm that they have the new Maps on other devices still.