I mean, is it inconsistent, or is it just a judgment call? I imagine dude was trying to thread the needle between a consistent methodology and interesting, digestible data. He could treat every subsidiary individually, but then you could have a highly fragmented list of players. He could have kept it top-level, but then you end up missing out on individual data for key entities. He probably split the difference to make it interesting to his audience while still providing relevant data.

This isn't a scientific study. It's a post on a sub about attractive presentations of data. I think OP succeeded.