Dungeon’s Master has a post regarding the use of skill challenges. The general guidelines seem to boil down to “If you’re in combat, you should use skill checks. If you’re not in combat, use skill challenges.”
I’m not sure I agree with that. The problem I see with that kind of distinction is that you’re going to limit yourself artificially, and you’re going to wind up using the wrong tool for the job from time to time. Sometimes a skill challenge does belong in combat, and sometimes a series of checks outside of combat is exactly what you want. To me, the distinction is one of dramatic weight, rather than relation to combat.
Let’s look at an example: an investigative scene. At first blush, it seems pretty obvious that you’d want to structure this scene as a skill challenge, and in many cases you’d be right. But it’s not always the case. The question you need to ask yourself is: how important is this investigation to the plot? Let’s break our example down into two sub-examples.
In our first example, the party is trying to find a killer before he strikes again. Someone’s been murdering people in Fallcrest, and the PCs have been hired to find and stop this guy. They can make various checks: to find clues, to analyze those clues, to question witnesses, to follow tracks, to get past the lock on the murderer’s door, etc. Each of these checks, if failed, contributes to a common outcome: it becomes harder to find the murderer, and he possibly kills again before he is found. This is a skill challenge.
But the stakes aren’t always so high. Let’s look at a different kind of investigation scene, one where the PCs are simply trying to gather information on a new town they’ve arrived at. The PCs have arrived at Tom’s Crossing, a small village ruled by an Elder who acts as the spiritual and political leader of the community. They can make various checks: to learn about how the villagers feel about this, to learn about how the Elder feels about this, to learn about individual characters’ true intentions, to learn about the surrounding area, to learn where to buy the best goods and get the best rates, and so on. There is no unifying set of stakes here, no sense of dramatic weight lending urgency to the scene; the PCs are simply gathering information that may or may not be useful later. Each check carries with it its own set of much smaller stakes: do they get correct information? This I would handle as a series of skill checks, assigning skills and DCs to particular bits of information, and awarding some experience for good role-playing at the end of the scene.
And that’s the distinction that I draw: how high are the stakes? If you can point to one set of stakes, one consequence for failure and/or reward for success for the entire scene, and everything is working toward that, then you want a skill challenge. This can happen both in and out of combat. The above example details an out-of-combat situation, but disarming a deadly trap that causes the room to slowly fill with poisoned gas is another example of a skill challenge, and that could easily happen while you’re trying to fend off some undead guardians.
If the scene does not have a unifying set of stakes, or if those stakes have a very low dramatic weight, then you’re better served using individual checks. Building a skill challenge–a good one, at least–takes time, and running one can take some time, too. Not only that, but using a skill challenge where none is warranted puts artificial constraints on the scene and the characters’ actions, and the players will notice that artificiality. They’re smart like that.