That’s a hard one. I’ve had several good, successful campaigns, and one phenomenal one that ended this year. What made them good or great? Honestly, in some ways it’s hard to say — just like good and bad art can be the matter of a bit of shitty editing, there’s alot of things that can go right and wrong at the same time.
First, and most importantly, you have to have buy-in from the the game master and the players. They all have to be interested in the game for longer than a one-off, and they have to be committed to the idea of playing regularly.
Second, play regularly. I can’t stress this enough — if you let a game dangle for a few weeks, the momentum is gone, and often the group won’t even hold together. People are busy and there’s always something that can take you away from the table — work, travel, kids, sickness, school…there’s always something. So if you’re going to commit, commit to a schedule that you can maintain, even if it means that every once in a while, someone has to play someone else’s character for a session or two.
Third: Have a consistent vision of the universe and the story you are telling. It’s going to change due to character action, losing or gaining players, and the longer it goes, the more chance that you will deviate from the original concept or story. This isn’t necessarily bad. The first few sessions really are more like a pilot episode of the TV series, to set the tone, world, and get people interested. Then comes the hard work.
You can half-ass the metaplot, like X Files or Lost, or you can have a consistent view of the story that might need to wiggle about to finish close to the mark, like Babylon 5.
Tied to that is the fourth point: Consistent characters, and this is where the players come in. Good characters might need a bit of tweaking until you settle into them. It’s rare you have a game where the characters are “right” at the start, much like TV shows. (The most wonderful example of this not being the case is probably Firefly, where the characters — even when they were changing and growing — were amazingly consistent and well-rendered.)
Another point about the characters — they should have some kind of connection to each other. It doesn’t have to be direct, but there should be web of why characters A, B and C are working together. Some campaigns lend themselves to this. A military, police, or espionage-themed game gives you the ability to throw people together because their skill sets jibe, or they simply were the guys that drew the short straw.
For example on a non-military campaign: our current Hollow Earth Expedition game features several characters that might not ordinarily work together or move in the same circles. Lady Zara is the money — she needed help finding her uncle and hired Gus Hassenfeldt to be her African guide. Simple. Dr. Gould came in a session later. He was a doctor with the Spanish that were harassing the White Apes Zara’s uncle found, and I tied him to the Atlantean background the city the apes inhabited. Now he’s a plot device and driver of the story, but still in her employ after they escaped Africa. Later, we added Hunter, a Terra Arcanum overseer/agent, who was sent to protect the Atlantean blooded doctor and prevent the secrets of the Inner World from exposure to the public (and more importantly, the power-mad men running Germany and the Soviet Union.)
So Zara binds Gus and Gould through employment, as well as other concerns, and Gould binds Hunter to the group through the Atlantean angle.
Your characters should have some kind of connection. Maybe they were old service buddies, maybe they’re related, maybe they work for the same people, or their goals are similar enough to pull them together. There should be something besides meeting in a tavern to “adventure together” to pull the group together.
With characters that have a connection beyond “we want to play”, a consistent vision for the world, and a commitment to play regularly, I think you’re halfway there.
Now you just need to catch lightning in a bottle.