Explanation
It solves a problem talking to camera cannot: describing a screenshot is always worse than standing in front of one. A claim with its source visible behind it is a structurally different argument from the same claim spoken alone, which is why the format dominates commentary and has outlasted every trend around it.
The failure mode is legibility. A screenshot that cannot be read on a phone is functionally a blank background, and the most common mistake is a presenter standing directly over the detail the video is about.
The common mistake
Fitting the whole source page into frame. At phone size that makes all of it unreadable. Crop hard to the part that matters and zoom on it at the moment the claim is made.