What are some ideas for possible combat outcomes other than, "You win!" and TPK? Like, you want to give your party challenges and have there be consequences for their choices, but you don't want to end the whole campaign with an anticlimactic minion battle.

• Ticking clock. You'll almost certainly win the combat, but if you don't do well, it will take more rounds, and you might fail to achieve some objective. E.g., the villain gets away with a hostage.
• Non-lethal outcome. Even if the enemies get all party members unconscious, they will not strike the killing blow and might even stabilize you after putting you in manacles. One PC might die if they go unconscious early and fail all saving throws before the battle is over. But the rest might wake up in a dungeon cell and have an escape adventure next session. But the delay causes a setback in the main quest line. And you will almost certainly lose possession of any McGuffins.
• Party unwillingly split. For example, the enemies might be targeting a particular member of your party, and will let the rest go free. Now you have a rescue adventure. The down-side is that player will not be very involved for at least one session, but it might be a good idea if a player has a scheduling conflict next session.
• War of attrition. You have a limited number of rounds before the reinforcements arrive. Then you'll have to retreat, but the more of the enemy you slay, the better the chances your main army will have when you reach the redoubt. Or the longer you hold your ground before you're overwhelmed, the more time your side has to set up defensive structures.

What are your ideas?

Less concrete, but the party could suffer a loss of reputation in the area. People don't trust them as much, or think they're as capable as they did before. Might effect who's willing to talk to them, what kind of jobs they can get, or even where they're allowed if they beef it hard enough. 

Broadly speaking, you need objectives other than "kill all the members of the other team". If the villains are trying to kill someone, or kidnap someone, or steal something, or destroy something, and they achieve their objective, they've succeeded even if the PCs take out all active combatants, and they've failed if don't achieve that goal. Much the same can apply to PCs, though "you accomplished the goal at the cost of all of your lives" is generally only the sort of things you do for a final battle.

Reward tiers

• Hostages get killed, enemies escape possibly with loot regardless of the mechanism the reward decreases over time 

remember hostages themselves can be rewards in the form of contacts, patrons, stores and followers

Earlier is easier

• This is basically the inverse  where the longer the players take to deal with a problem the harder it gets. It may take the form of repairs or other side activities they need to do which reduces down time ect..

For example when dealing with a plague the longer they take to find a cure/ the less they do to slow the spread the more cure needs to be produced so the longer it takes to get back to normal or maybe the more resources they need to spend to cure it. 

The way is shut 

• Some areas may just be inaccessible or become inaccessible if they fail a task. These can be area for side quests, additional loot, short cuts ect... 

You could have a situation where many innocents are going to die  and hte parties job is to minimise it. Dozens of zombies are going through a village a couple of fireballs may easily take them all out but also kill a number of villages, taking out the zombies will result in the zombies killing possible more villagers before they are all taken care of.

Another option is to have a bad guy trying to cast a ritual and the rest trying to protect him. If the caster is able to survive a number of rounds without losing concentration the ritual is successful even if the party "win" the fight. The ritual can be anything from casting Geas on the king (probably want he PCs not the know the spell and itts cure just to find the king starts making somme dubious decisions and they then have to find out why) to destroying one of a number of anchors which protects the world from a terrible evil, if a certain number of these anchors are destroyed this thing is let lose.

In general though I far too many D&D combats are fights to the death. Most people would rather escape than die in combat so if things are looking bad they should be looking for some way to run away, (simply flee, cast teleport / cast dimension door, swig a potion of haste, surrender and ask for mercy....) this should apply to the party as well as the monsters.

Quote from Jegpeg >>

 

In general though I far too many D&D combats are fights to the death. Most people would rather escape than die in combat so if things are looking bad they should be looking for some way to run away, (simply flee, cast teleport / cast dimension door, swig a potion of haste, surrender and ask for mercy....) this should apply to the party as well as the monsters.

 Perhaps a recurring monster that gets stronger every time it flees. So the trick becomes to deliver a big final punch round so you can get it to 0 hp before it triggers its < 25% hp escape action.

I've always been partial to "you win the combat only to discover you were fighting the wrong enemy."

I like combats where the target is something (an object or NPC) that you're trying to protect, or even a specific PC, and once the bad guys destroy/kill the target, they flee. For example, a dragon might not want to kill all the party members, only obtain a specific magic item from them, using force as is necessary. Similarly, a pack of starving wolves will relent after it has killed one PC (or a mount or other creature), or is given fresh meat.

Speaking of fleeing, let you PCs run away from a fight if they are doing poorly. Most monsters will not attack retreating enemies. Letting PCs run away lets them lose a fight without all dying.

Quote from pavilionaire >>

War of attrition. You have a limited number of rounds before the reinforcements arrive. Then you'll have to retreat, but the more of the enemy you slay, the better the chances your main army will have when you reach the redoubt. Or the longer you hold your ground before you're overwhelmed, the more time your side has to set up defensive structures.

I actually do this with dungeons. Let's say the players need to clear out a bandit hideout, they could head inside and kill as many as they can before they need to retreat out of the dungeon and take a rest. At this point the surviving bandits are also allowed a rest and may have time to prepare for the players' return. I also have monsters do death saving throws and if stabilized they regain 1 hp after 1d4 hours. This is why I keep track of in-game time. It also wouldn't be a bandit hideout if there wasn't an armory, a kitchen, a restroom, etc. The bandits follow a routine so depending on when the players enter the dungeon, a lot of things may change.

Another example of war of attrition is how I handle literal war. If there's a war going on in the campaign, the players may choose to burn down the weapon and food supplies in an enemy camp. Killing all enemies may be impossible for the players, but they can use stealth to destroy those supplies during the night. This also works wonders during a siege. If the players are on the side of the besieging party, they can help burn down bridges and cut off supply routes, and they can infiltrate the besieged city to destroy its supplies. If the players instead want to defend a besieged city, they can infiltrate the enemy and assassinate the leaders to drop morale and cause the enemy to become disorganized.

Should the players fail their stealth checks and alert the enemy, it can count as a failure that isn't a TPK. Suddenly, their objective becomes extremely difficult if not impossible. Good luck burning down those supplies or assassinating that leader in its sleep when you now have dozens of soldiers after you.