We would only be able to conclude that it's unfair for "great start ups" if the experiment showed that random acceptances performed worse than the interview process.