Yep, I know what you're thinking. This guy's going to talk about splitting the group up again, and to be perfectly honest, I am. Today though, I want to look at it from the practical side of the coin. See, movies and tv shows often break up the group mid-action sequence in order to up the tension. The group is suddenly faced with a situation where in order to meet their objectives they have to split up. The only problem is, as they break off, this makes the situation rougher on those who are left behind to deal with the original problem. Now, some groups will fight against this - and that is fine - but some will dive into it head first too, so let's take a bigger look at it.
At the very basic level, this works like so: During a tense action sequence, you present the PCs with two objectives that need to be done simultaneously, or one will fail. This leaves them with the choice to either focus on one, split up to try to do them all, or to go for some other middle of the ground type scenario. The point here is though, a choice is coming into play, and a choice always brings tension with it. The trick to this is simple too: you have to be willing to declare an objective lost if they don't go for it in time, or at least much worse/harder to do.
I have seen, and I personally have, this sort of set up ruined on numerous occasions because for whatever reason the GM wasn't willing to go with what the player's chose and just call it a loss. It is one of the rough parts of being a GM sometimes, especially since you also have to be fair, but just keep this in mind: If you are clear that the choice is there, and the potential consequences of the choice, then you haven't done anything wrong.
This might go easier with an example, so here is a fairly basic one. The PCs are trying to stop a gang from robbing an armored train car. They find where the heist is going down, show up, and begin to engage the bad guys. The fight is going back and forth, made all the harder by the fact that you have cars and the train moving around. However, when it becomes clear the PCs might win, the gang leader and his second jump back onto one of their trucks and begin to take off. Now, the PCs are left with a choice: they can stay and make sure the majority of the gang is stopped from robbing the train, they can split up and have some go after the gang leader while others stay to finish up the train part, or they can all go for the boss and leave the train alone.
Whatever way the PCs choose to go, is how the situation will play out. If they split up, then both groups have a tougher fight as they don't have all the PCs with them. However, if they abandon one side or the other, then those people are going to get away. After all, no one is there to stop them. Now, it is possible that the PCs can find whomever they let go later, but that should be harder to do. For one thing, they have to find them again, and for another, the NPC now knows the PCs are coming for him, and coming hard. It doesn't have to be an automatic loss, but it should be harder. Respect the PCs choice, and let the consequences flow from that.
There Shouldn't Be a 'Right' Answer
This is where a lot of people mess up on choices, and one of the hardest things to really do when you're GMing. However, when you are presenting these choices to the PCs, it is vitally important that their not be a right answer. The game shouldn't come crashing to a stop because the PCs chose to approach a problem a certain way. Become harder? Sure. Stop? No. See, by putting a right choice in, you aren't adding tension, you are just adding anxiety. The PCs then worry about choosing wrong and screwing the whole thing up.
Instead, you want your choices to reflect how the PCs approach the problem. In the above example, the 'right' answer could appear to split up to catch the boss and stop the train robbbery, but that isn't necessarily the case. All three ways can lead to a PC victory, all three ways can consequences for good and ill on the group. All this means that the choice just lets you know how the story goes, but not whether or not something wrong has been done.
Don't believe me? Well, let's play it out.
If the group ignores the fleeing boss, they can still stop the train robbery. That was their initial goal, and they succeeded at it. Sure, the boss got away, but most of his crew has been caught/killed and anyone still breathing can be made to flip on the boss to get the PCs back on his trail. Sure, it'll be harder than when he was right there, but the game is still afoot.
If the group splits up, they have the chance of getting a "total victory" right off the bat. Stopping the train robbery and grabbing the boss all in one go. Of course, they also increase the risk of losing both at the same time. If they pull it off, then they get all the reward and awesome of that total victory, but they also can make themselves out into bigger targets for later work. After all, people tend to not discredit these kinds of actions as just being 'lucky'.
Finally, if the group focuses on the boss, then anyone still on the train can get away with the loot. Sure the PCs stop the boss, the ring leader is down, and that'll put a huge dent in the underworld as a planner is down. At the same time though, the PCs still have to catch those guys that got away.
The point is, all three leaves the game plenty of ways to still go and none is ultimately the right decision. Good and bad can come from all three, and in a way should come from all three. It is what keeps the story going, the game interesting, and the choices meaningful.
So, try to have fun with the choices you give your group, but also try to make sure that it isn't a choice between "right" and "wrong". "Hard" vs. "Easy" is ok. "Right" vs. "Wrong" isn't really a choice.