People won't abandon Discord until all your private chat/call logs are hacked and dumped. When that happens, only a fraction say 10-15% of the userbase will look for a permanent solution (i.e. something that has always-on E2EE). My guess is about 40% will change to another service with better track record, but that isn't architecturally any more secure (i.e. that's still TLS-only). The remaining 50% will stay on the service because 1) they don't care, 2) their peers don't care or, 3) because they consider the incident response and related PR good enough.
To get people to switch from discord you need something that's faster, that just works better, that fixes some issues faster than Discord staff can fix them, and that adds something the users didn't even know they needed, say, in-game HUD for discord that fits the theme of the game or something, more intuitive management like push-to-talk management commands for the client via secondary button.
As long as it's about decentralization (value/belief system stuff) or about privacy (It's just in-game talk I don't care), and not about helping the user focus on head-shots and kill-streaks, it's not going to be adopted.
https://www.theverge.com/2019/7/12/20691957/mastodon-decentr...
EDIT : Also, interesting that the American right has moved from Twitter to not Mastodon, but to Parler instead :
https://news.ycombinator.com/item?id=25097145
I wonder why ? Mastodon presenting itself as being too left-wing ?
(Though make no mistake, a lot of the American right is 'conservative' only in name, and is instead wildly anti-conservative both in their speech as well as in their deeds.)
Can't we just wait till they run out of money for all the free services they are currently giving away ?
Next they will try to agressively monetize their user base, which will make people migrate to something else - ideally self hosted or federated & based on open protocols this time.