They quantified phone screen effectiveness by yield of candidates who passed onsites. I don’t know if they said, but I think those were conducted by different e

They quantified phone screen effectiveness by yield of candidates who passed onsites. I don’t know if they said, but I think those were conducted by different engineers. That’s not infallible, but they’re committed to it being an ok metric—-it’s already what determines who they’re hiring. If you think that their decisions in on-sites are sufficiently bad, you probably should give up on the idea of optimizing how phone screens are conducted--it would be wasted effort, and you'd have no idea how to make sure it worked. You can easily imagine that these results are just noise, but it’s more difficult (though hardly impossible) to spin a story in which it’s making their process worse. I think it would require that these interviewers have a knack for selecting weaker candidates who can still pass the on-site.