To prepare for any narration I write out the script and then start reading it out loud. I check for flow and I make sure I really focus on articulating every syllable. I have a small bit of a Southern accent when I talk, so when I narrate I make sure to use my "teacher voice." I remember the first time I heard myself narrate; I thought, "Wow! I speak so slow!" I try to pick up the pace when I narrate now, but I'm always nervous that someone is going to think I narrate too fast.
Then I read an article by Ray Pastore (2012). Pastore conducted an experiment to find out if people's comprehension of audio would change if the speed was compressed. I was surprised to find out that there wasn't a huge gap between the 0% compressed audio and the 25% compression. There was a dramatic decrease in comprehension with the 50% compression. The decrease in comprehension between the 0% and 25% was less than one full point. I think if I were to create a very long podcast, the amount of listening time saved might be worth the risk lowering comprehension. Of course, that would totally be subjective to the content and purpose for listening.
Pastore. R. The effects of time-compressed instruction and redundancy on learning and learners' perceptions of cognitive load. (2012). Computers & Education, 641-651.