Is it something you really want to happen?
This is the alpha and omega right here, the crux of any discussion about whether or not you should plan something out in advance. If you're going to plan on something taking place, you should want it with a passion, not just with a vague sense that it would probably be cool.
Keep in mind, I'm not talking about something as vague as "John and Sarah should meet." I'm talking about "John should employ Sarah as his new personal chef." These are developments that you feel are very important, probably too important to leave to random chances. You feel that this is in the best interests of one or both characters.
Keep in mind that just because you want something to happen doesn't mean that you should plan it ahead of time; that's just the first question to ask. Depending on your experience with writing and storytelling you might recognize that your character wants something that wouldn't actually be all that interesting. So figure out whether you do, in fact, really want this to happen, and then move on beyond that to determine whether it's worth the effort.
Is this scene difficult to make happen otherwise?
My wife and I spent most of our time in high school going to the exact same mall, but we never ran into one another despite having the same interests. We were both supposed to go to separate colleges. We both nearly didn't go to the club meeting in college where we first met. The first time we met in college, we didn't remember the other's name. The coincidences required for us to meet and spend nearly half of our lives together were pretty substantial.
Reality is unrealistic, and sometimes, you don't want to chance something that might not happen. You'd rather just take control and say that something's going to happen.
What you have to consider here is the possibility that something even better might happen if you left things to their own devices. Weigh the two against one another. If you think there's a realistic chance of something happening as a more off-the-cuff bit of roleplaying, you might not want to plan ahead. However, if something better happening would rely on someone going to a very specific spot or knowing things he or she couldn't possibly know in advance... yeah, planning ahead is probably the better option.
Are more people likely to enhance the scene?
Sometimes the random element turns out to be a great boon to the dramatic elements of a scene. You start interacting and some stranger (or even a friend of one character) arrives, throwing everything off its axis. That's sometimes just what you want, but sometimes it's actively counterproductive to what you're trying to accomplish, adding in distracting elements and unpleasant associations that muddy the waters.
Planning something out ahead of time doesn't eliminate that aspect, but it does mean that you're working toward a goal instead of just seeing what happens once the words start flying. If having other people around would ultimately be disruptive rather than beneficial, there's something to be said for having the end results planned out ahead of time. Even if you get disrupted, you still have a track to progress along.
Relying on this too heavily, of course, leads to every interaction taking place in a completely sterile environment under your direct control. Caution should be your watchword; if other people might enhance the scene or lead you in unexpected directions, you shouldn't stymie that. It's just that sometimes distractions are not a good thing.
Does this bit of directed plot lead to more dynamic plots?
Allow me to go off-topic for a moment and talk about one of my favorite open-world games of all time: Saint's Row the Third. It's a great game, as many people have pointed out, and it lets you do a whole lot of things in an open sandbox.
After the first hour or so, anyway. For that first hour, though, you are on rails so tight the game won't let you so much as glance to the left. It's a series of strictly separated vignettes that don't even give a hint of sandbox-style play, with each individual mission leading directly to the next.
Here's the trick: By putting you on very guided rails to start, the game makes sure you're ready when you get thrown into the wider world.
Final Fantasy XIV uses a similar trick to teach you about the value of the open world. You start in a limited space and later get the run of the house. And you can use the same trick in roleplaying, which is basically the ultimate upside to all of this discussion. Planning ahead of time should be something that's going to lead to even more dynamic roleplaying further on down the line.
It's not that you can't have more unplanned stuff otherwise but that the quality could improve, that there will be new relationships to deal with, new potentially uncomfortable confrontations and assertions, new challenges to overcome. In exchange for being slightly less flexible now, circumstances will become more flexible in the future.
That's something to chew on as you plan your events.
Feedback is welcome, as always, in the comments or by mail to email@example.com. If your interest in all of this is flagging a little, you'll be happy to know that next week's column is all about sustaining interest over the long term in roleplaying events. The week after that? Let's talk about prophecy.
Every Friday, Eliot Lefebvre fills a column up with excellent advice on investing money, writing award-winning novels, and being elected to public office. Then he removes all of that, and you're left with Storyboard, which focuses on roleplaying in MMOs. It won't help you get elected, but it will help you pretend you did. If you need a refresher, check out the Storyboard Library.