Make sure your explanation is understandable and in colloquial English. Avoid enumerations, be to the point and do not dwell on things that do not matter. Support the videos with words on screen if that's possible. Try to limit your videos to 1 minute. A teleprompter can help the speaker to speak smoothly. There are apps that can help you with this. Make sure the speaker looks directly in the camera, so that they look at the participant directly.
By showing a bad example first, and following up with a good example, the viewer will quickly understand the essence of the theory. Once again, make sure the texts are short and to the point. Aside from that, a good example is recognizable for the viewer. Make sure you record in a familiar environment and with realistic actors. When the actors have to show a certain emotion, put it in the script clearly. In examples, actors are talking to each other. This means that the actors don't need to look directly into the camera.
In order to encourage trainees to redo exercises, it is important to watch the duration of the Role Play. As a rule of thumb, the duration of a good Role Play varies between 5 and 15 seconds. Within the Role Play itself you don't need to worry about describing context because you specify the context in the situational description beforehand (this helps you to make great Role Plays). Make sure the actor speaks directly to the participant, and thus looks directly into the camera. With Role Plays, always specify the emotion the actor needs to show. Make sure the text of your script is written in everyday language also, this may include slips of tongue and coughs.