If I'm going to do a declaration phase (I don't currently, but I have in the past), I want declarations which are more specific than that, because the main thing I want out of a declaration is to get rid of the "optimal play" pattern of everyone focus-firing one enemy into oblivion and then instantaneously changing to the next target as soon as the previous one drops. I want declarations so that players will spread out their attacks to avoid the risk that the first hit drops the enemy, and then everyone else's attacks are wasted on beating the corpse to a bloody pulp.
I guess I'm also a fan of re-reolling initiative each round, and making declarations - especially for spells or withdrawal from melee - before initiative is rolled. So it seems the main thing I want from declarations is that they be specific enough to have a risk that, by the time you actually get to act, your declared action may no longer be relevant, or no longer possible (e.g., you declared to cast a spell, but got thumped in the noggin before getting it off, so it fizzles). In theory, at least, that encourages actual planning and strategy around the declarations, rather than every turn's action being "what is the optimal way to react to the exact situation at this instant?"
I'm getting that by how I handle the flow of combat within the "group" actions. Let's say I've got 9 players, and 4 of them win initiative and thus go before the monsters this round. Two of the players want to keep wailing on the same monster as last round. So they don't need to say anything, just start rolling.
But, I don't take the results yet. Meanwhile, the other two need to clarify something with me, which takes a few seconds. I get them sorted and they start rolling. By now, I can start taking the results from the first two. And so forth. It's a little chaotic, but I rather like the chaos because it gets 80% of the effect you discuss, which is good enough for me. Every now and then, someone at the table overhears excitement from a good roll and can guess that another player has pounded some monster good, but its incidental enough that I don't mind, the same way that players can guess AC after enough close hits and misses.
New players all eventually ask me if they can wait and see what happens before they act. Sure, if you don't mind losing initiative. Or if already lost initiative, giving up your action to automatically get initiative next round. That usually puts an end to that thought. Plus, general declares work both ways. If they say they are firing at the group of goblins, I'm usually fairly generous about letting additional hits spread over once a monster is downed. OTOH, I'm the one choosing the exact targets, and I spread them out when multiple people attack. If there are 8 goblins and 2 ogres, and all 4 initiative winners said they were concentrating on the first ogre, then that's what they do, waste or not. Their reward for an overkill is an over the top description of how the ogre goes down (or reels but stays up or whatever). Or the 4 players can all say the ogres, in which case I choose the exact targets. More often, a player will sense that a monster is reeling, and that player will declare to lock on to the target until resolved, but other players will be more general.
There is a certain element of training the players for all of this to work, due to the chaos. I'm usually running without a grid. So in return for less control over the exact targets, I'm also fairly generous with movement. The scene is imagined more as a confused melee over here and skirmishing over there, and the ranged opponents trying to hang back. A character may not get to pick their exact target, but a successful attack has a target somewhere, even if that sometimes means the target was already going down.
Mainly, it lets me run my work game which has 1:15 sessions, such that I can do a fight with 10 players and up to twice that in opponents, and resolve the whole thing in 20 minutes. We've had exactly one giant fight that took a session and a half, but that was when they let themselves get trapped on top of a rugged hill (admittedly a great defensive position) by 4 encounters at once.
Edit: I don't get the spell declaration effect as strong as you describe in my D&D games. In my own system, I'm approaching that from a different angle, but it is dependent on how magic works in that system (usually multiple actions to cast all but the simplest of spells, and skill rolls needed).