There are many ways to approach it like (just first things that come to my head):
- simple exit surveys (e.g. if I unsubscribe from ISU channel on YT, or Olympic Figure Skating channel, or such, the moment I press the "Unsubscribe" button I am sent a survey asking me to share my reasons and explaining it is not to make a marketing effort on me to come back but to help to improve the viewership of the sport), a good list of exit groups/media/social media/channels would need to be compiled and off you go; in a slightly more sophisticated survey you could approach "sleeping accounts" this way, I mean those who did not unsubscribe but just have not been actively using their accounts for some pre-defined prolonged time (even exits/sleeping accounts from GS could bring some interesting data, probably, although drowned in the site related opinions) - this obviously depends on legislation in various countries who can be approached this way and by whom
- focus groups - this is a qualitative research method which you can just check on the net as it is a bit complicated to explain here
- including a series of questions into more general TV viewership survey like asking "Have you ever watched..., Do you watch regularly.... Why not... Why not anymore .... what would have to happen... etc. (which gives you some kind of quantitative data difficult to obtain otherwise).
- addressing fora like this one here for brainstorming sessions

and surveys, especially asking them if they can approach their former members by email with a survey
There are just ad hoc ideas off my head and it is not even my area of expertise - I am sure if you approached some specialized agency they would give you more and more elaborate options.