While a consolidated condition phase would be both convenient and a time saver, I think the better way is to have timing occur normally (conditions are applied immediately following the attack that causes them). This allows players (and monsters) to play tactically without having to try to think an entire round ahead (which is part of why the current system takes as long as it does to resolve, IMO).
That is an issue. I don't have so much of the analysis paralysis in our group (because most of them aren't that tactically savvy) as process and handling overload. Fortunately, your solution dovetails with mine.
I think RW's applying conditions only at the end of the turn will cause more slowdown than it helps. You have to record what happened. And then at the end of the turn, you have to apply it. This is exactly like "picking up the same piece of paper multiple times"--which, as any efficiency expert can tell you, should be avoided when possible.
This is not true, however, of doing the saves, ending conditions, applying stuff from ongoing effects, etc. By definition, it has to be handled "later". So handling it all at one time in the round is convenient.
As for the everyone taking their turn problem, having spent many combats running with group initiatives, I can say that the DM and players pretty quick to pick up on an optimal flow for that. I've found with us that I can have two or three players going at once, because the natural variations in decision making, rolling the dice, and communicating the results provide windows to handle it. (That is, one player is nearly always ready before the others.)
Having sometimes nine players, I'll occasionally get more than 3 going at once. In that case, I'll divvy them up by some ad hoc termination that makes sense in context and resolve it that way. "OK, you two are going after the lizardman. So we'll do that first, and then come back to you other two holding off the shaman." By the time I get over to the latter group, they've resolved everything and are ready to report/narrate.
I do think it is important in a system like this to not try to exactly mirror the likely results. But you do want to be in the ballpark.