For years I have relied on an adventure-design framework I call clear goals, fluid obstacles.
1. "Clear goals" means the whole party shares one (or several) very specific and clear goals so they know what they are supposed to be achieving. E.g., kill this dude, rescue that dude, solve this mystery, retrieve this MacGuffin, etc. This is very important to keeping the party on track, preventing them from wandering off or splitting the party or not knowing what they are supposed to be doing (that sort of waffling sometimes comes up in a sandbox game and can really kill momentum).
2. You can have a patron assign the goal, or for a longer-term goal you could ask the players to come up with reasons why it's important to them. For example, if you want a sandbox campaign that is about killing the local evil wizard, ask each player to explain why their PC hates that wizard and wants him dead.
3. "Fluid obstacles" means you think up a bunch of stuff that is in the way, preventing the players from completing their goals -- but you don't tell the players what those obstacles are. Well, maybe you tell them the initial and obvious obstacles (e.g. "The local evil wizard is a powerful wizard who lives in a tower and probably won't want you to kill him.") but for the most part you let the players discover what it is they need to do.
4. The obstacles are fluid because you can add them at more or less any time. This gives you good control over pacing. If the players are racing towards the goal in a way that is boring, add a couple of obstacles (encounters, puzzles, complications). If the adventure is dragging on... well, maybe skip the last obstacle or two. I think it's handy to come up with these obstacles ahead of time, and make sure all the obstacles are plausible and make sense. That way the players won't feel like you are just messing with them.
Anyway, my point is, if you want to run a sandbox game with a "story" this is one good way to do it. As long as the PCs advance towards their goal and eventually face a climactic encounter, it should work out OK.
1. "Clear goals" means the whole party shares one (or several) very specific and clear goals so they know what they are supposed to be achieving. E.g., kill this dude, rescue that dude, solve this mystery, retrieve this MacGuffin, etc. This is very important to keeping the party on track, preventing them from wandering off or splitting the party or not knowing what they are supposed to be doing (that sort of waffling sometimes comes up in a sandbox game and can really kill momentum).
2. You can have a patron assign the goal, or for a longer-term goal you could ask the players to come up with reasons why it's important to them. For example, if you want a sandbox campaign that is about killing the local evil wizard, ask each player to explain why their PC hates that wizard and wants him dead.
3. "Fluid obstacles" means you think up a bunch of stuff that is in the way, preventing the players from completing their goals -- but you don't tell the players what those obstacles are. Well, maybe you tell them the initial and obvious obstacles (e.g. "The local evil wizard is a powerful wizard who lives in a tower and probably won't want you to kill him.") but for the most part you let the players discover what it is they need to do.
4. The obstacles are fluid because you can add them at more or less any time. This gives you good control over pacing. If the players are racing towards the goal in a way that is boring, add a couple of obstacles (encounters, puzzles, complications). If the adventure is dragging on... well, maybe skip the last obstacle or two. I think it's handy to come up with these obstacles ahead of time, and make sure all the obstacles are plausible and make sense. That way the players won't feel like you are just messing with them.
Anyway, my point is, if you want to run a sandbox game with a "story" this is one good way to do it. As long as the PCs advance towards their goal and eventually face a climactic encounter, it should work out OK.