Try hard Not To
Try hard not to do this, is my first instinct. In a straight-up D&D type game, this situation gives me the hives, because what was a difficult juggling act has suddenly become two linked and synchronized juggling acts, one in each hand. My job gets exponentially harder to do with split parties, much less split parties over extended time periods.
All of a sudden, you have to keep each branch of the adventure roughly the same in table-time (i.e., two hours for each group, for a four hour session), roughly the same in game-time (so they can rendezvous later), balance them in terms of spotlight time, make sure your combats are scaled right so the players don't get killed horribly, make sure the smaller groups have the right skills for traps and social encounters... it's a mess, and it's very hard to do.
I feel no shame in steering my players away from that if I can.
That said....
Play To The Point Of Tension
With each group, play approximately to the point where a TV show would cut to commercial, which is to say, when it becomes clear that a combat is looming, or when they have to make some major decision, or they have just learned something critically important or interesting or threatening.
First, that's just good narrative design-- always leave 'em wanting more.
Second, though, it gives the players something to do when you switch back to the other group. Specifically, they can plan their approach to the combat, or hash over the decision to be made, or discuss what they just learned. It cuts the boredom factor down a little bit.
What this means, though, is that your branches have to be structured this way ahead of time. If your combat-heavy branch is really just one big two hour slug-fest, that's going to make this approach more difficult.