I've spent my vacations planning my very first sandbox, however, I noticed that, being a hexcrawl, most of the time my players will be exploring the world and sometimes, they might spend a lot of time wandering into empty spaces. I know this can be solved with random encounters, however I don't want to make the random feel like bad filler on a TV series.
We're playing in a post apocalyptic fantasy Tokyo, which has pretty much became a continent. The game should take them from low heroic to low epic tiers of play, so I must take advantage of that region, giving each hex about 5 miles and placing myself a limit of "one important place/adventure per hex".
I want to avoid sessions where the only thing we do is exploring hexes and swinging blades or chattering with whatever comes, I want to make dungeon crawling, even when found randomly, serve a purpose beyond giving them experience and loot; I don't want, also, to make adventures feel forced. My players have long term goals that go beyond "I want this item!", so giving them a sense of "plot advancement" is important.
Best Answer
When I read this I thought: Just don't.
In "the real world" there is so much more random stuff that happens than meaningful stuff towards some goal that its the decision of the people on place to decide which dungeon to crawl and which to ignore.
If one wants to explore every single house in every street then in how many of that houses you will find something worthwhile? In the first hundred or so you find lots of useful stuff, good loot interesting challenges but then it starts to get boring and you start to skip houses you think are not worthwhile.
On the other hand there are houses that are interesting far longer because they are far rarer. An industrial complex vast enough to hold its own population that never needs to leave it holds buildings that are super interesting all the time.
As long as you give them enough information on how the hexes look like / what to expect from the NCSs (through rolls against social skills) the players will soon start to skip stuff and start to learn when something is "just random" and what stuff is "significant". (And think about all the interesting plot stuff that happens if they get it wrong.)
This learning process is one of the things that makes sandboxed games so interesting.
Also, sometimes people can't skip random stuff. That happens, really. The gang that wants to mop them up is going to pray on them, either because they carry around significant amounts of cool loot, or they happen to have the wrong shade of gray skin.
After about 10 sessions or so (your call) you should allow your players to make one roll if they think their current encounter is significant to their personal plot to fasten things up.
If you go this route you should tell your players so and recommend to them to keep notes, because it happens that they skip stuff they should not have skipped and they need to be able to figure this out afterwards.
As a sidenote: I found it very hard to come up with "good" randomness in such situations. I use the traveller world generation stuff for this, only I generate "countries" (in my Post-Apoc setting more like "tribes" or "warlords" but its all the same) with it. Works like a charm. I generate the countries by the Traveller rules and then write two to five lines on it what it means in the world. Stuff that just won't fit gets re-rolled.