Ti copio cosa ho spiegato a uno su reddit, e’ in inglese, fattelo bastare che non ho sbatta adesso 
There is a very good chance that it’s not a single component that is to be fixed. Modern software engineering rely on concept like the separation of concerns which means that an action like a “login” might actually pass through several different functions or “states” and it takes time to properly identify where and when is the chokepoint.
On top of that, when identified, you might encounter good reasons for which that choke point hasn’t been addressed prior to this moment (limitation on third party services, tech debt which is workarounds meant to be temporary but with no time to go back and fix properly, limitation on the infrastructure - server/network/budget, etc)
After all this, you might end up with a problem that requires multiple moving parts to be changed and some of which that are meant to “never stop” or that if changed would cascade in a series of other changes necessary (imagine as in a black box, and you need to change the way the input is sent, you’ll need to rewrite the inner working and also everything that was using that input to now provide that different input)
There are multiple other aspects but to keep it simple, this is the gist of it.