I can't speak for Jeremiah, but this is how supporting a game post launch typically works. I'll modify some of my knowledge to the Reach / 3 situation.
At the start of each release cycle, you have to determine what your priorities are. New gametype? New map? Update to existing map? Update to existing gametype? An entire playlist?
Now add in all the known exploits sitting out there that you need to fix. Boneyard exploit, broken weapon on Hemorrhage, etc.
Now, you've got to sort the priority of the updates you want, because you have to maximize your sustain dollar. You have to maximize TIME x MONEY x AMOUNT OF PLAYERS BENEFITTING or AFFECTED
So you sort all of these outstanding needs and fixes for your update priorities.
Now, let's take Sword Base as an example. If you remember, the map had a changed weapon layout and other gametype support added to it. While this update was pending to be released, pushing out an update that blocks the ledge - which mostly affects Infection and wasn't a CRITICAL bug - would have taken time away from more critical things. You have a limited amount of time to build, test, and cert the map updates.
So they obviously made the decision to include the softkilled Ledge as part of the Sword Base Update. That way the soft kill being added gets included with the testing for that variant, instead of a bunch of testing for the previous (and now defunct) Sword Base just for a kill zone.
Then you package up all the high priority updates you could get to, and push them out in the next monthly update. You save mid-month updates for actual emergency problems, since each time you update the lists, you're pushing out a new playlist data file to millions and millions of players. If you did microupdates every 2 days, your bandwidth bill gets huge.
It was honestly kind of painful to watch some people here while I was stuck in registration limbo mock Shishka for "taking a month to delete a fusion coil" from Valhalla. Besides the fact that the map had other updates made to it along the way, and that it had to wait for the next update cycle to be pushed up anyway, since it wasn't an emergency-level fix. I hope a few specific people are never allowed near a code repo nor a test department, because some people's idea of how game support actually works is a galaxy away from reality.
It's easy to say "just go delete ___ in Forge and upload it!" when you don't have to consider cert time or budget implications of that action. Especially when the way they have to rig up maps and gametypes is not the way we build maps and gametypes.