The first model is the chase. Either the PCs are chasing someone/something, someone is pursuing the PCs, or both. A campaign where the PCs are chasing someone gives the GM a certain measure of control concerning where they go next, but that can limit the PCs if they hear about somewhere else interesting that they want to see. Being chased can be bad for keeping the group together. Eventually some PC will decide that tracking two/three/however many separate will be harder for the pursuer and odds are the group will splinter. A combined model of the chase solves the problem of the party splitting, but still has the problem of the party feeling led by the nose.
Second is a quest campaign. For quests either the PCs are going somewhere known and distant, going somewhere unknown and assembling clues, or doing a rod of seven parts where they need to visit multiple locations. Going to a known, but distant location allows the PCs to chart their own course with stopovers of their own choosing. It also means that they could avoid interesting parts of the setting, because neither PC or player know that their interesting. A quest where clues need to be assembled requires a certain amount of existing setting knowledge and runs the same risks as any game that relies on solving mysteries (too many clues, too few clues, bad rolls, strange logical leaps, etc.). Collecting pieces from multiple locations will get the PCs around the setting and lets them pick the order of travel, but cuts down on their ability to take side trips to places that sound interesting.
The last model is one where all PCs are part of one group and have a superior (or superior force) that just sends them places. This will let the GM send the PCs to what the GM considers the high points of the setting, but it limits the range of logical PC types and/or prevents almost all opportunity for side trips. (Sliders had a range of character types, but none of them had control on where they were going.)
Balancing player agency with player ignorance in a new setting is a difficult thing. Does anyone have a technique or model that as worked consistently when showing players a completely new (to them) secondary world?