I like this version a lot more than the first - the elements aren't fighting as much between themselves, so it's less frantic and easier to listen to.
Ultimately though, the trouble is the vocal is so dominant, and isn't reacting to the music in a natural way. Listen to the narration on something like The Snowman audiobook (YT video is desynced slightly from the video so best just listen) to hear how the narrator reacts to the dynamics and energy in the music.
I'm not familiar with too much spoken-word stuff, however my dad introduced me to the eccentric works of Vivian Stanshall as a lad, and though it does contain music (drastically out of time/tune at parts), I would not necessarily say it would sit well with OCR's library, enjoyable as it is. Stanshall also knew that the music would distract from the story, so the majority of that record is just him narrating over silence.
A more popular example might be Fire Coming Out of the Monkey's Head by Gorillaz. Spoken-word verses with a sung chorus. I love this song, and although it contains a lot of prose, it's punctuated by choruses, and it works because there isn't much melodic content behind the narration, and execution is flawless. Also Dennis Hopper.
I might be more open to this if the vocal was re-recorded, removed, or the pacing was more like in the Gorillaz track (less vocal, couple of choruses, no melody under the vocal). It still feels like 2 separate entities playing over the top of each other.
NO