It was impossible to ban people from Usenet, because it was a distributed system.

It was possible to ban people from moderated groups, or complain to their ISP and get their ISP to ban them (much harder), or it was possible to simply cancel their messages.

I believe at the peak 1/3 of all usenet traffic was spam and 1/3 was spam cancels.

Back in the day the IRC ban would most likely be handed by a human, or if automation was involved, the person behind the automation would be reachable. There's the potential for assholes of course, but at least those assholes weren't trying to monopolize humanity's entire social fabric so you could always join another IRC network if you were banned on one.