One of the fundamental principles that's coming out of a century of hiring research is you have to evaluate candidates on the same basis - any other approach is

One of the fundamental principles that's coming out of a century of hiring research is you have to evaluate candidates on the same basis - any other approach is an invitation to bias and subjective decisions. If you evaluate two similar candidates, candidate A on their github and candidate B on their performance on Hackerrank then it's incredibly hard to fairly compare those two candidates without adding subjective bias. In practice what happens is the company goes "oh well the person with a github clearly has more initiative". But is initiative part of your evaluation process ? - have you ensured you've fairly allowed all candidates to demonstrate it, or have you ended up evaluating candidates unfairly on totally different metrics. If you're hiring you can vastly improve your hiring process by doing two things: (1) Adopting a standardized structured interview where questions map directly to your needs and (2) using a work sample test (getting candidates to do a small piece of work that resembles real-world work as closely as possible).