Little update - I'm actually having some success just creating a 3D solid, putting it at the right part of the head I want to track, and then multiplying it's position by the X, Y, and Z offset percentages that the face tracker creates.
Something like this for position, where I add some expression sliders ("Offset X", etc) to tweak the amount of translation:
tempX = thisComp.layer("FaceFootage.mov").effect("Face Measurements")("Offset X") * .01 * effect("Scale Offset X")("Slider");
tempY = thisComp.layer("FaceFootage.mov").effect("Face Measurements")("Offset Y") * .01 * effect("Scale Offset Y")("Slider");
tempZ = thisComp.layer("FaceFootage.mov").effect("Face Measurements")("Offset Z") * .01 * effect("Scale Offset Z")("Slider");
[transform.position[0] + transform.position[0] * tempX, transform.position[1] + transform.position[1] * tempY, transform.position[2]Â + transform.position[2] * tempZ]
It's pretty noisy on Z, but I bet I could smooth the data and then send this to C4D via Cineware.
Still open to hearing any suggestions from anyone that has played with this!
Thanks again,
 - Nathan