Many RTS campaigns have a meta-map which dictates the flow of the campaign, such as the one found in Dawn of War: Dark Crusade. Meta campaigns aren't inherently a good or bad approach compared to a traditional, linear mission structure, it depends on how well it's executed. Meta campaigns are easy to mess up and be an anti-fun grind, so they should not be tacked on as an extra feature. A meta-campaign should be the entire focus for the single player campaign or a substantial DLC like Company of Heroes 2's Ardennes Assault. So what is it that makes a good meta-campaign?
Not all types of RTS games are going to work equally as well with a meta-campaign which is crucial to identify. Meta campaigns tend to consist of procedurally generated skirmish missions with certain perks such as different win conditions. Short 1v1 skirmishes with constraints don't make sense for an RTS game like Supreme Commander, which instead shines in lengthy sandbox style missions on massive maps. With an emphasis on skirmish style missions, the skirmish AI has to be really fun to verse for a meta-campaign to not just feel like a grind. While Rise of Nations has minimal variation in its missions, the skirmish AI is so fun to verse that it hardly matters. Let's take a look at some of the reasons an RTS might want to do a meta-campaign over a traditional linear campaign.
Meta-Strategy & Progression
Meta-campaigns give the player more agency which ties into the power fantasy that RTS games can deliver. It's one thing to be in control of a big army, but it's another to coordinate an entire planetary conquest! Good meta campaigns deliver a sense of progression as territory lines shift and grow. Meta campaigns give players the choice of where and when they attack, do you go straight for the enemy stronghold or first build up by capturing neutral territory? Do you finish off the crippled empire or do you prioritize harassing the biggest threat? Do you fortify areas as you advance or focus full aggression? These types of decisions are fun and more meaningful than lame story choice cliche's.
Meta-campaigns typically feature a meta-resource gained from conquests and is used for upgrades or temporary bonuses. They also generally include progression of unlocking of new content, hero upgrades, global perks, and the pacing of starting forces. Progression is typical in linear campaigns too, but it's a lot more fun when it's the consequence of the player's decisions such as being tied to sector bonuses. Meta-campaigns generate tension by being an arms race of escalating power between you and the other warring factions, as your territory lines grow and forces strengthen, so too does the enemies which is viewable at all times.
Meta-maps also serve as a great form of contextualization without needing excessive exposition and introductory dialog. Some RTS games have excellent story-driven campaigns, but those are the exception and not the rule. RTS games make it hard to do narrative because of the scale and perspective. There's no face-to-face character interaction or dialog choices while the top-down view detaches the player. The new player thoughts of "Why do I care and why do my actions matter?" is easy to answer when you're a small army sitting on the island of Japan surrounded by barbarian hordes and the empires of Korea and China! Later on, looking out on the map and seeing you now own the entire continent of Asia feels awesome and is innately a motivation to continue the conquest for the rest of the world.
With minimal character dialog and narrative, an RTS can instead focus on lore and backstory, the context around the game. Who are the factions and characters fighting over this world? What are their personalities? Why do I care? I don't consider that narrative because context doesn't need character development or interesting plot twists. Dawn of War: Dark Crusade is a perfect example of lore without story. Meta-campaigns also have the potential for interesting show-don't-tell through gameplay on the meta-map. What if the Chaos in Dark Crusade were overly aggressive, while the Orks rapidly regenerate forces and Space Marines could summon reinforcements from their Imperial Guard allies?
Video games should always try to communicate through mechanics instead of cutscenes and dialog. The Global Conquest mode in C&C3: Kane's Wrath did this by having unique victory conditions for each faction. The global government of GDI win by securing control of enough of the globe, the chaotic forces of Nod win by bringing enough cities into full unrest and the alien invaders of Scrin win by constructing 9 Threshold Towers. (Planetary Gateways). Company of Heroes 2's Ardennes Assault uses excellent show-don't-tell through its representation of the battered US companies holding out in the Battle of the Bulge. Manpower is finite, so every loss you take on the battlefield is permanent across your global company strength that can result in a campaign loss if each company is depleted. Watching as your Airborne company is removed from the meta-map due to your failed mission is far more emotionally impactful than some corny dialog in a cutscene.
Playing Past Losses
Another benefit of meta-campaigns is the ability to make failing a mission acceptable. In a traditional RTS campaign, you need to win every mission to progress else it doesn't make sense narratively. Being forced to replay a mission in a traditional campaign isn't fun because it's heavily scripted. Meta-campaigns, on the other hand, aren't limited by story so losses can just delay progress or incur a penalty. Because meta-campaigns make it okay to fail an individual mission, they can be made more challenging without the loss being overly frustrating. When failing a mission is more of a possibility, then each more tension is generated and you can have wider fluctuations in difficulty to keep players on their toes. (Which is an inevitable consequence of semi-random missions or force strengths.)
Since it's okay for players to lose individual missions, the game can be designed around not being able to save-scum. Without the ability to restart a mission or load old saves, every decision is permanent and tension is much higher. Save-scumming is not fun, yet players will do it anyway out of the desire to optimize and be efficient. To be safe, you probably want to leave in traditional save functionality either as part of the lower difficulties or as an option to disable such as Ironmode in XCOM2. If a campaign is intended to be played without save-scumming, then the overall campaign should be short. Having to restart a 5 hour playthrough is very different to having to restart 20 hours of progress. (I found Ardennes Assault was the perfect length). With a short campaign play time, there needs to be an emphasis on replay value, which is a huge potential and benefit of meta campaigns.
Meta-campaigns are inherently repetitive, so the repetition has to be fun and not feel like a grind. Meta campaigns don't need scripted missions, but it also needs to be more than just a set of regular skirmishes strung together. To avoid repetition, missions should have random (or semi-random) properties such as win conditions, mutators, map types, AI personalities, and other modifiers. Mix all these qualities and you end up with a large number of semi-unique random missions which you could describe as procedural generation. The extent of the required variation depends on how fun the skirmish AI is to play against, which is more a consequence of the design and mechanics of the game than the technical complexity of the AI. (Although smart AI obviously helps.) Dawn of War: Dark Crusade gets away with very little mission variety because its skirmish AI is fun, but that's not the case for Company of Heroes 2. Ardennes Assault navigates around its poor skirmish AI by having massive mission variety and more scripted missions and mutators.
Procedural Generation can also apply to the strategic level, the meta map. Factors such as spawn locations, territory lines, mutators, resources, bonuses, and progression can vary, which will make repeat play through feel much more unique. The great thing about meta-game design is it isn't limited to RTS gameplay, XCOM2 is one of my favorite games and a solid implementation of the meta-campaign. XCOM2 has different global build orders that have massive implications for your in-game performance such as weaponry and abilities. As XCOM2 is an RPG it makes sense to have your tactical gameplay heavily emphasized by your global progression, more than an RTS. The balance of how much impact meta-strategy has opposed to RTS gameplay is a delicate line, as a general rule for RTS games, the tactical side should matter much more. It feels awful to be thrown into utterly unwinnable battles, which is why I think the emphasis of Strike Forces made the Kane's Wrath Global Conquest mode unpopular. Try to find ways to add variation to the meta-game without overly impacting the RTS gameplay.
Adjustable Campaign Options
While the metagame should have randomized variation, it should also have settings that the player can tweak to configure to their preferences and to add more variety to repeat playthroughs. Difficulty is one factor that can be very flexible, such as separate difficulty sliders for the meta element and gameplay element. Specific settings may make a game more easy or difficult by removing or tweaking a meta-mechanic, or it may just be a personal preference that someone finds a particular mechanic annoying. The campaign could be tweaked to go twice as long, or all the battles are 4v4 AI battles instead of 1v1. The meta-campaign could only contain one powerful enemy instead of 6 weaker ones. There's so much potential for options that are just minor tweaks but can give enormous replay value.
The majority of missions in meta-campaigns will randomized skirmish missions, but not all of them need to be. There's room in meta-campaigns to have scripted missions for extra tension, climax or inversely a more guided introduction. It's typical to see a scripted introductory mission that serves as a tutorial then opens up the meta-map to the player. The final mission can also be an epic battle, or each rival faction can have a scripted stronghold mission such as in Dark Crusade. Parts of the meta-screen could also be scripted, such as the Alexander the Great campaigns in Rise of Nations. Linear campaigns, as opposed to meta campaigns, don't need to be opposites, they need not be all-or-nothing. The StarCraft II campaigns are mostly linear but give the player a choice of which missions to pursue first to determines the unlock order of units.
Even with varied missions, it's still possible to get a feeling of a grind if you have major scripted missions like base sector assaults broken up between randomized skirmish missions. To avoid that feeling of grind in the late game, missions should have auto-resolve. The lack of auto-resolve in XCOM2 made unimportant missions in the late stages of the game frustrating to encounter as they ended up just being a time sink. For auto-resolve to best function without being overly random, there needs to be a system of force strength quantification which are compared and determines win chances. Force strengths in the RTS gameplay typically manifests as the size of the starting army and base.
Mission Resolve States
Missions can have more than simply win or lose states depending on the game mode. If missions can be resolved with grey areas such as stalemates or pyrrhic victories (Where you win but suffer massive losses), then there's much more tension with each mission. Even if victory is inevitable in an Ardennes Assault mission, you're still trying hard to minimize losses as the manpower losses diminish the company strength which bleeds over to subsequent missions. In XCOM2, suffering wounds on your operatives leave them unable to participate in future missions while they recover, forcing you to rely on inexperienced recruits. Non-linear success states also adds more depth and less RNG to an auto-resolve mechanic.
RTS games are already very confusing and complicated, so adding a layer of metagame on top of that can further alienate new players. Try not to front-load all of the complexity to a new player. The first missions in a campaign could be purely linear without introducing the meta-game or showing it without giving the player control until the 3rd mission. Elements of the meta-game can be introduced gradually, such as choosing where to attack but being locked out from upgrades, abilities and supporting armies until later in the campaign. The meta-screen should avoid clutter and be clean to prevent overwhelming new players and use submenus to separate information.
Meta-campaigns aren't inherently better than a traditional linear campaign and are suited to particular RTS styles more than others, but they have many benefits. Story-driven campaigns are challenging to execute well in RTS due to the perspective and scale, whereas meta-campaigns contextualize and immerse the player without needing story. They also have the potential to communicate through show-don't-tell gameplay which resonates more with players than fluffy dialog. Meta-campaigns enhance the power fantasy of RTS games and give the player choices about their global strategy while hooking them with progression. Meta-campaigns are less scripted, so individual missions and the wider campaign should have procedural generation of maps, win conditions, game modes, mutators, and AI personalities. Advanced options further add to the replay value by allowing players to customize their experience and make repeat playthroughs differ. Meta campaigns aren't limited by a narrative, so losing individual missions can be acceptable. Higher mission failure rates create more tension and opens up the removal of save-scumming. As RTS games are very complicated, the introduction to the meta-game should be delayed or gradual to avoid overwhelming new players. To prevent a feeling of grind, battles should have the option for auto-resolve and preferably with non-binary resolutions.
Cheers, and happy RTS'in'!