How is a bad configuration change that took out their entire network NOT a single point of failure? It's not like there were twenty different network engineers

How is a bad configuration change that took out their entire network NOT a single point of failure? It's not like there were twenty different network engineers that typed the exact same command across their network. Ultimately it was one person who typed "enter" and submitted the bad change. And the problem wasn't even necessarily the change -- mistakes should be expected. The problem was the lack of a back-out strategy in case this happened.