Not so much in my PhD, but during one of my postdocs, my advisor thought me the importance of benchmarking and replicating some previous results or models before using them as foundation for my own work.