Listening again. What I'm hearing most is an inconsistency in the level of your vocal. It's in the first verse too, actually. Certain lines are popping out well, others sound masked. For example, notice the difference in the first line of the verse and the last ("you don't even see it yourself"). I'm not saying the vocals need to be out front, necessarily. But when the lines are coming in and then fading, it's frustrating to try to follow, like trying to follow a conversation in a loud place. Personally, I would go through the vocal track and do some manual edits, boosting the softer bits, getting the levels more even.
Nice break at 1:05. Again, the vocal in the middle section that begins about 1:39 comes is very low. Not a clue what you're saying. That harmony at 1:42 is dissonant, and again when you repeat it a few lines later and later in the song.
Second verse at 2:18 is probably the most consistent in terms of level, apart from a couple of lines around 2:30 that I struggled to understand. Again, in the part that follows that verse starting around 3:00, the levels are inconsistent. Hear how that line "can we do anything" (I think) is lower than the ones before it?
Nice outro.
Anyway, just one man's opinion. For me, the gold standard is that every word of every line of the lyric should be clearly intelligible.
