This is because of fraud. Everyone either does this, or enables the fraudsters. The only third option is to get rid of email, except the fraudsters would just move on to the next thing.
Assume that you’ll have availability problems with email and engineer with that in mind.
Claiming that you either need to ban users without recurse or suffer from enabling fraudsters is why people start feeling like corporations don't have any human employees anymore. If you care about your customers, you could absolutely reach out and verify if the user you are about to/have blocked is a fraudster or not, and then act accordingly.
The reason most just say "You're banned, bye" is because they don't want to do that work, and subsequently don't care about their users one bit.
They ban thousands of accounts per day. Every day. If you can solve this problem you will have discovered a license to print money, so by all means try. But the problem is much harder than you are making it sound.
On the other hand, how many do you think would reach out when they get blocked? Most actual fraudsters wouldn't. Of those that do, you can take a look at what they are sending and quickly see if they are actually doing fraud or not.
I could say the same thing to you, you're making it sound harder than what it is. But if you're set to the mindset of saving money, I understand minimising work makes sense.
Of course they do. It’s just email/twitter/telegram/WhatsApp, and fraudsters are literally professional messaging automators. They open support cases, email execs, I have seen cases of them paying call center workers to call support lines and complain. Fraud is a serious business run by serious people, many of whom have a decade plus of experience.
You cannot attempt to stop fraud with manual effort if you operate at any kind of scale. Full stop.
> Assume that you’ll have availability problems with email and engineer with that in mind.
That goes for all dependencies, not just email. And not just third-party dependencies either.
If PyStorm died today, I could carry on in VIM. I could rebuild my entire architecture on DO if AWS sours. If git takes a dump, I have four machines from which I could copy the code and migrate to e.g. mercurial. And if I get hit by a bus, my code is well documented and testable.
Assume that you’ll have availability problems with email and engineer with that in mind.