It is not looking for the loudest parts
Loud is easy to find and it is not the same as watchable. A stretch where you shout at a jump scare for four seconds is not a video.
Four signals decide, and the first carries the most weight. Topic cohesion: does this stretch keep returning to one subject across its whole length? Sustained reactions: energy spread across the block rather than one spike. Audience interaction: how much of the talk is aimed at chat. Speech continuity: a block that is mostly dead air is not an activity, however exciting its few moments are.