How to replicate this effects?

First things first, here is the video in question = V̴̬͂̍i̵̪͘r̷̬̈t̷̗̿͐u̵̜͊́á̵͕l̴̫͌̈́ ̵͈̈͆R̶̘̟͋ê̸̦̪ǎ̴̛̟l̵̠͝i̵͈̲͠t̸̫̣͋y̷̞͑̚ ̷̠͒ă̸̢͘ġ̴̪ͅa̶̰̓i̵̗͒͠ǹ̵̞̒ - YouTube

I am looking to replicate this effect in a video of my own, but I cant work out how the designer manages to get all of his layers (the shapes, outlines, app icons, videos and images) to follow the phone screen. I have played around with perspective point tracking and managed to replace a phone screen in a test run, but that is all. I cant get the same level of detail and I cant make the objects 'stick out' of the phone and follow the motion of the phone.

Are the details in their own composition? Is it all one composition? How does he do it?

If anyone could help I would greatly appreciate it.

Thanks

Chris