Just curious. Whats the point?

Good question. It bugs the shit out of me. Even Criterion does it.

I've never been particularly clear on the whole process, perhaps it's cheaper? Maybe they use a different software for encoding? Maybe it takes up less data space on the disc?

When source material is mixed, like when they mix film at 24fps+video at 30fps, or 25 for PAL, or even 60 fps footage etc together, they can't switch from progressive to interlace and back, so they encode it all at 1 rate, like 30fps NTSC for the featurette. That's 1 reason.