One thing which has begun to bother me about a lot of recent science fiction and action films is that they always go straight to putting the whole world in peril. In Guardians of the Galaxy the climax is an attempt by the villain to destroy the Good Guy Planet. In Interstellar the hero is looking for a new home for humanity because Earth is doomed. In The Fifth Element a giant cloud of Evil is coming to consume the Earth unless the heroes can find the MacGuffin to stop it.
I think this is a mistake.
In screenwriting guides like Robert McKee's Story, the advice for writers is "GSU" — give the characters a Goal, keep raising the Stakes, and generate Urgency. Now certainly this is good advice if you want to make an exciting film (or novel, or whatever). You don't want characters wandering around with no purpose, you don't want the audience to stop caring, and you don't want anyone to get bored.
But! There's a paradox. If you raise the stakes too high, you destroy all the suspense. Unless the film is a black comedy like Doctor Strangelove, the audience is pretty sure that the world won't be destroyed. Which means all the hero's heroics are not that suspenseful because they are destined to succeed.
There are ways around that: if the "hero" is a group of people, the writer can kill off some of them along the way, creating uncertainty and suspense about who will make it to the ending. Or in a movie, the director can simply dazzle the audience with footwork: in Raiders of the Lost Ark we're pretty sure Indiana Jones will accomplish his goal, but we never know how he's going to manage it and are always entertained to see what's coming next.
And, of course, there's the heavy-lifting method: keep the stakes small but make them important to the characters, then make the audience care about the characters. In Casablanca even the noble Resistance leader Victor Laszlo won't affect the outcome of World War II very much, but the audience nevertheless cares very much about how the love triangle between Victor, Rick, and Ilsa plays out.
Of course, this method takes a bit more work, but I think the payoff is ultimately worth it.