By and large, I've been talking about the big picture for KWORC so far, with occasional forays into fine details for purposes of illustration. This is because for the most part, the model works equally well for a campaign as it does for a single session, but the small points of divergence can be a problem. And the most noticeable of those is the endgame. It is easy for a campaign to have a high-level, slightly abstract core goal, but a given mission tends to be very specific in its needs.
See, if your goal is something simple, like killing the bad guy, then there's really nothing to worry about. You work your way through the problems, get to the end and grab the golden ring. Victory! However, the goal is not always simple. A nuanced goal like "bring the villain to justice" may actually have multiple components including "Stop his scheme" "Provide compensation for his victims" and "Keep him from ever doing it again" and those are not necessarily resolved in the same way. Similarly, a mission like "Steal the Golden Egg and swap in this duplicate without anyone realizing it" actually splits into two or three subgoals right there on the surface of it.
Now, the good news is that structurally, this is not that different from a normal KWORC construct. Let's use our 'Justice' example to illustrate - in practice, if we accomplish all three of the subgoals (Stop him, provide compensation and keep him from doing it again) then we have reasonably accomplished our core goal of bringing him to justice. Structurally, those three goals are similar to problems that must be overcome to achieve the goal, so we can just branch out from them, figure out the KWORC needed for each and we're good to go.
Now, as much as I can point to these subgoals all falling under the same umbrella and being part of a larger whole, that's not always true. Sometimes these goals have differing levels of priority, or only some goals are important to some characters. And once you've opened this door, then you also invite in other goals entirely. And this is where it gets scary. One goal with lots of branching, interesting problems looks like a beautiful, complex jellyfish with long, spiny tentacles. It's a little tricky to handle, but manageable. But multiple goals? Holy crap, how are you going to manage the complexity of a pool full of those spiny critters!?!?
It may seem like a contradiction, but the solution is simplicity. See, it's entirely possible to build a very long, complex KWORC construct around a single goal if you're so inclined. Doing so is desirable if you want to build an entire campaign around that goal, but if you want something less grandiose, the simple reality is you rarely need to go more than two layers deep around any goal, and often a single layer of problems is all you need.
This may sound like simplification makes for less boring play, but that overlooks the power of multiple goals. Goals are much more interesting drivers of play than problems are. In a broad sense, when measuring interest, adding a problem adds to interest, but adding a goal multiplies it. To illustrate this, look at show that have complex, multi-problem/goal plots. The path to any particular goal is usually fairly short, but enough things need to get done to keep busy.
Now, once you've opened up the floor to the idea of multiple goals, your bag of tricks gets much bigger. In addition to the aforementioned introduction of secondary goals, you are now armed to handle things like changing goals midstream. And because you stick close to the obstacles surrounding the goals, you're not making any more work for yourself. Players may create interesting threads as they work their way to those core problems, and that's awesome, but you don't need to stress the bookeeping because you still know where the endgame is.
1 - In this context, a "layer" is how many iterations of problems you need to reach the goal. A goal that requires knowledge (as password) and Opportunity (access to a computer) has only one layer provided you can get that knowledge and opportunity. If you need to steal the password and keycard from a guy to get those things, that's a second layer.
2 - Another, unrelated trick. Capability obstacles are a fantastic basis for a hiring montage. Consider the beginning of Ocean's 11 or any Mission: Impossible through a KWORC lens. The core goal has been identified, the problems have been branched, and several of them are capability ones. At this point we know we need someone who can crack a safe, someone who can pass as a Ukrainian housekeeper and someone who can take down 3 security guards without giving them a chance to raise an alarm. BAM. Recruitment montage follows.