Hey there, Comfr.
To create a new campaign is to bring ambition to fruition. A good campaign is engaging in a variety of ways, but I will tell you of my design philosophy for gameplay. I see Revolta has already covered stories. Bear in mind that this is my own approach; you may find that there are other ideas you associate with more closely than the ones I will present.
To make an engaging map, you have to balance familiarity with original concepts. Resources with taxes. Consider the ways anything you add will impact the gameplay. If you give the player two barracks at the start, will they be inclined to play more defensively and build more bunkers? If this mineral field is only lightly guarded, should there be some trigger-spawned enemy forces to counter its benefit? If there is an optional objective during the map, what will it ultimately accomplish? Does the benefit warrant the effort?
A good map or mod takes these factors into account. To engage the player, their survival must be uncertain. Make it too easy, and it'll more likely be boring. Make it too difficult, and frustration will drive players away. Think of it like this, if you will:
Minerals and Vespene are your resources in Starcraft. They are required to build up your forces, construct your base, and therefore survive in Starcraft.
Enemy forces - and their tech, how their AI affects your objective, like being aggressive or defensive, etc - will generally have the most impact in taxing those resources.
Difficulty settings (those exist in SC2, right?) and tech counter that tax.
Resources, tax, and temper. An ideal setup with these things in mind would be giving the player enough tech to survive what the enemy throws at them, and enough resources somehow to fund the production of that tech. Give the player something to work towards, something to think about, very often.
I'll give you a few examples from my own redone SC1 campaigns (so far) to give you an idea. They may be imperfectly implemented at this juncture, but the idea is there!
In Biting the Bullet, there are three phases to the map. The first involves saving four Terran bases full of civilians from the Zerg, and then attacking a hive. The second phase is the shortest, you simply hold the line against the Zerg until the Protoss unexpectedly arrive. The third involves taking out the rest of the Zerg.
Not only are resources somewhat scattered around the map, but rescuing Terran bases gives you minerals and vespene to spend on repairs, upgrades, or if you want to activate the Zerg AI early, the construction of your own base. You capture buildings like academies or engineering bays in a few rescuable bases, enabling you to make those upgrades. This keeps a player engaged and gives them something to look forwards to - and even in the latter stages of the map, the fact that resources are spread around in general enables the Zerg to strike where your forces are spread thin.
In Egression, the Protoss are split into two - one with an "insane" AI, and another with an "easy" AI. The easy AI trickles units at the player while the insane AI builds up and causes trouble. This creates the need for a constant defense while challenging the player here or there with larger waves. You do have an ally, though, and resources are pretty much okay from the get-go. So, the variation in Protoss attack patterns makes this map more interesting.
There are barely enough resources on this map to wage a fight against one Zerg base. However, there are multiple potential expansion locations that are guarded by Zerg. This forces the Protoss player to secure intersections across the map and invade enemy territory if they want to keep up the fight, using resources to build units to get more resources.
Hopefully this perspective has been helpful to you. Good luck!