Jason Van Orden brings up the error of using "hits" to measure the approximate number of times a  Podcast was Downloaded. 

"…Looking at the # of hits to a podcast file (MP3 or other) does not indicate how many times the entire file was downloaded. Some of those hits are the result of simple requests for information (such as the file size) or partial downloads. Further, if someone starts a streaming player, but stops listening ten seconds into the program, it still results in a hit to the file in the logs.

In Podcasting Metrics Series Part 2, I pointed out the problem with looking at the # of hits to a file for estimating the # of downloads. A single download can result in many hits. The hits stat will always be bigger than the actual number of downloads."

Most Analytics don't do much about RSS Feeds, Podcasts or Vidcasts and really ought to - and I've been suggesting this to every Web Analytics Vendor I talk to.

Right now, I gather the bytes downloaded of a file by the size of a file to get an approximate number of downloads for a file. Another possibility I found gets you close the first set of numbers is taking the number of Hits to a file and multiplying by .80 which seems to be consistent.  

Every podcast is usually downloaded a little more than 1x per visitor - same is true of transcripts of a podcast - that .80 multiplier seems to work but I can't tell  you why it's .80 except that I  observed this ratio works several times.