What I've found with campaigns is that the requirements of the game must come first, with as much lore being crammed in as possible without compromising those requirements.

So for instance, you need a rigid structure to the campaign to prevent it drawing out for too long (and players losing interest as it goes).

No one wants to play one-sided slaughter-fests, so you don't want to engineer situations where one player has to come in, deploy, play a game (and give up his afternoon off to do so) just to play a match where he is outnumbered four to one, or otherwise so outclassed that the match is a foregone conclusion.

You want to avoid situations where one player has to play lots of games, while others don't play any. Studio Tomahawk recently released their Age of the Wolf campaign book, and one of the things I don't like about it is that players choose who they want to fight at the start of each round, with no limitations. Which might mean that all players decide to attack Fred, while Jim and Tony choose to defend and don't get to play any games at all this turn. You don't want someone to feel victimised, nor do you want someone to feel ignored.

Some things you could do for Armada might be to have your Admiral attached to a particular ship in a particular fleet. If another element of your fleet is engaged, then your admiral can't take part (and you will have no admiral for that battle) because he is elsewhere. Ships could start with no upgrades, and only be awarded them after fighting battles, and there could be an element of resource management to repair damaged ships. Titles might be one-use-only, with the title being lost after that ship is destroyed.

1. So for instance, you need a rigid structure to the campaign to prevent it drawing out for too long (and players losing interest as it goes).
2. No one wants to play one-sided slaughter-fests, ...
3. You want to avoid situations where one player has to play lots of games, while others don't play any.

I agree with pretty much all of that, Chucknuckle.

The biggest hurdle in my opinion is finding enough relatively like-minded players, and that like-mindedness can happen on a number of different axes, where you have competitive playability on one end and lore-realism on the other (e.g. the ones you imply: scheduled contests vs. a map with freedom of movement; dictating the players' distribution of forces vs. giving them the freedom to distribute as they think makes sense).

I know that I'm close to one end of the spectrum: I find my greatest jollies in the modeling of the Galactic Civil War. I'm also secure enough in myself that I don't feel too badly if the odds are stacked against me and I lose, just so long as it nicely models what it's supposed to represent. I know other players for whom it is supremely important that they have a fair shot at winning. I understand that I'm an outlier, and I'll have to compromise many of the things I would like in order to achieve playabiity. However, I think I would need to warn people for whom the competitive balance is so important that their enjoyment depends on it, that a campaign may just not be for them.

Another aspect of it is to what degree the outcome of one game influences the balance of the following game. If there's too much, then early winners will have such a structural edge over those who lost those early matches. In that respect, I do think there's a role for the fog of war, and a disconnect between tactical and strategic victories, which can keep the campaign and the individual games exciting.

It will be really interesting to learn what sort of balance FFG struck with Corellian Conflict.