I'm currently experimenting with low poly modeling. I  hit a snag with my workflow with animation and face animation because of my high polycount. My facial animation experiments came up short because i did not have full control of the topology. I think if I model the head from scratch this will improve my workflow and give me more control over facial animation.