For calculating beam transmission per surface, I would suggest a polarization ray-tracing approach (as opposed to polarized POP) based on the Geometric Image Analysis tool.  This assumes that diffraction can be neglected.  (Note: Unfortunately, the Polarization -> Transmission tool only provides chief ray transmission per surface.) 

For a Gaussian beam, it’s important to explicitly set the aperture apodization to be Gaussian and to also use an Apodization Factor of at least 1.5 (or larger), so that the wings of the Gaussian are not clipped.  The Aperture size can then be adjusted to properly set the diameter of the beam (e.g., see: Apodization Factor).  Here is an example using the Double Gauss sample model:

Then select the appropriate Geometric Image Analysis parameter settings:

For a single field point, just set the field size to zero.  The source should be “Uniform.” The image size should be large enough to capture the entire beam footprint at *any* location along the lens train.  “Use Polarization” should obviously be selected.  Then the beam transmission after propagation to any selected surface is reported as “Percent Efficiency” in the header.  This includes Fresnel/coating loss, material absorption and vignetting (if present).

Lastly, to streamline the process, the IMAE operand can be used to find the transmission values.  In this case, first “Save” the Geometric Image Analysis settings.  Then the IMAE will use the saved values unless overridden by any of the various input parameters to this operand (such as the surface number), as shown here: