You don’t even need to figure out how long the song is just record song length + 1 min of click. Then, record a track that talks you through what is happening where with say an acoustic guitar and mi mic / spoken instructions. This track wont be heard in the final mix however it is a good idea for the narrator not to say things that will make people laugh too hard.

This will save a lot of problems with people missing cues etc. from that point it is just a case of filling in the blanks and then archiving the click and the place keeper track.