The biggest strength of Video Puppetry is not only that it is real-time, but it allows multiple puppeteers to work on the same animation remotely. This distributes the control of complex puppets among several puppeteers and at the same time promote easy collaboration. Such remote collaboration feature is not fully supported even in many of the popular animation tools.
Small segments of articulated puppets would be difficult to control. If such segments are too small, they might not even be identified because of hand occlusions. Splitting a character into separate segments and joining them again in software is not common in traditional puppetry.
I would extend Video Puppetry as an audio-first project where users can not only say special words to trigger certain events(like mouth animation), but also narrate the story as they move characters in the animation.