I am sure there are a lot of factors, I also know a lot of people who were not using the hardware acceleration (display optimization), making even posing difficult. That said, I am in the same (learning) boat, I am barely moving passed the 'making characters' stage, trying to build sets (and props) I am happy with, so what do I know? I have never even set up a single proper scene.
It took some doing but I finally found a way that I can post and set up multiple figures without lag. In my pipeline, I load up the figure and then replace all the surfaces with a "rubber" material. Leave them naked, and then do all the posing. With the active pose tool things move about pretty easily and there is no lag or jerk.
Once I have everything posed the way I want it, then I go back and add the lighting. Since there is only a single material and no background, the test renders happen out pretty quick and I can get the general lighting set in quick order.
Next, I group all the figures together and add the scenery. I move the group to where I want them in the scene, then do the lighting test again to see if I need to add more to light the background.
Once that's done, I throw on the actual skin materials on the people, add the hair, clothes, etc, and then do a final lighting check to see if I need to tweak the intensity up or down.
I know it sounds like a lot of steps, but since you're doing the bulk of the moving and test rendering in an environment with fewer surfaces, it makes all the "guess work" portion go a lot faster.
Of course I'm always open to new work flows. This is just the one I've found be the fastest for me.
