Segmented masters treat each portion of a song (Intro, Build, Chorus, Drop, Breakdown, etc.) differently, tailoring it to the rest of the song as a whole. Mastering in stems allows for increased control on my end, which allows me to maintain dynamic range, a huge factor in the era of music attenuation. Platforms such as Spotify and iTunes attenuate songs based on overall loudness (LUFS), measured cumulatively. The classic approach to this issue is working to lower the LUFS through the entirety of the song, but with segmented mastering, your song can maintain an overall low LUFS whilst still being mastered to maintain full loudness on the parts that require it (Drops, chorus, etc.). This is possible due to the low in LUFS, high in dynamics portion(s) of your song (intro, breakdown, build) that will leave room for a more pronounced drop/refrain/chorus.