As with all things, there's the game-rules answer and the world-simulation answer.
If you just want to have there be an alternate goal in this one case and your game is otherwise constrained like a CRPG, just come out and list the potential end scenarios. State it. There's no reason to put lipstick on a pig. You'll just frustrate people who want to play a tactical skirmish game that there aren't clearly defined win states.
If you are running a realistic world/game, then think... How do people in the real world get missions where they just need to recon or skirmish or hold a strongpoint or retreat or guard people or defend a place or escort a convoy? Well, they can get orders if they're part of a larger organized force. Or they can be paying attention to the needs of the situation and their in-world goals and the goals of those around them. If the goal is "get the artifact in the chest to Galvin within the week," then wandering off to wipe out every vestige of the orc tribe they just met isn't a good use of time and resources (or heck, it may be way bigger than they can handle). And you have to be flexible as the DM, as they will come up with different ways to accomplish what are presented to them as the goals and parameters, possibly not "occupy these X squares for Y rounds."
Want them to run? Show a force 50x their number coming after them. Want them to hold a point? Put something important (and not man- or available-magic-portable) there. This approach does offer a lot more room for ambiguity, where you have to learn by failure. In a campaign I run, the cleric is quite portly and slow and not all that good in combat. The other PCs figured out after the first couple times they all ran off and left him behind that someone needs to be a blocker for him if they're going to be on the move. This approach engages PCs not just at the simple tactical level but in an ongoing and organic manner.