5 ms·

Randomness in interviews is underrated because it offsets these over-selection effects. I'd love to see the results for a company where every applicant that pa

Randomness in interviews is underrated because it offsets these over-selection effects. I'd love to see the results for a company where every applicant that passed basic screening tests was given a 1% chance of an offer; with additional probability weight awarded for performance in interviews.

rightbyte 3y ago
I believe most recruiting processes are worse than chance so skip the weights. Roll a dice, pick someone, fact check the resume for scammers, done.
- ghaff 3y ago
  I honestly don't think it's that bad. But if one or two (for a sanity check) qualified people look at resumes and have a 30 minute conversation/technical weeding out question with a candidate, you're probably getting into diminishing returns past that. (Though you probably also lean heavily on referrals and "pedigree" given that you're not really trying to get some non-obvious signal in the interview process.)
  - rightbyte 3y ago
    Sure I agree. I was abit hyberbolic, but not much, if you account for all the bad processes.
- mistrial9 3y ago
  As an American in California, watching the extreme lengths people went to, to find specialists fitting some very exacting requirements and literally bringing people from around the world and then rejecting them in large numbers, I was told that in China in the dot-com boom time they formed companies with very skilled people but basically formed the teams with whoever showed up. It was a completely different method, and the difference is the virtue signalling and status cues for the founders and their inner circle. In Silicon Valley, it was anti-status to NOT search the world for particular resume contents, and that got weird sometimes. Secondly the "google style" interrogation of new candidates with "you just got out of computer science graduate school so you learn to present like this" whiteboard sessions, evolved in the hands of callous or worse managers, into a sort of repeated hazing of programmers, by non-programmers or the like.
- red-iron-pine 3y ago
  A lot of the back of the envelope math points to new hire failure rates of around 50%, where failure is if the candidate is still there in 18 months. Only about 19% would be considered "successful" hires. Also only about 11% were not successful due to purely lacking technical skills. like I'm sure we all have stories about a noob that doesn't know how to do something basic like ssh with a key or something, but in many cases that's not the issue. https://sowelo.eu/unsuccessful-recruitment/ https://sowelo.eu/unsuccessful-recruitment/
  - FirmwareBurner 3y ago
    >I'm sure we all have stories about a noob that doesn't know how to do something basic like ssh with a key or something Not knowing how to do that, or not knowing how to google how to do that? I also joined some jobs as a noob when switching domains, where I was lacking skills that the more seasoned people would consider as being "basic", but I could also google what I was lacking most of the time and learn on the fly.

rightbyte 3y ago

I believe most recruiting processes are worse than chance so skip the weights. Roll a dice, pick someone, fact check the resume for scammers, done.

ghaff 3y ago
I honestly don't think it's that bad. But if one or two (for a sanity check) qualified people look at resumes and have a 30 minute conversation/technical weeding out question with a candidate, you're probably getting into diminishing returns past that. (Though you probably also lean heavily on referrals and "pedigree" given that you're not really trying to get some non-obvious signal in the interview process.)
- rightbyte 3y ago
  Sure I agree. I was abit hyberbolic, but not much, if you account for all the bad processes.
mistrial9 3y ago
As an American in California, watching the extreme lengths people went to, to find specialists fitting some very exacting requirements and literally bringing people from around the world and then rejecting them in large numbers, I was told that in China in the dot-com boom time they formed companies with very skilled people but basically formed the teams with whoever showed up. It was a completely different method, and the difference is the virtue signalling and status cues for the founders and their inner circle. In Silicon Valley, it was anti-status to NOT search the world for particular resume contents, and that got weird sometimes. Secondly the "google style" interrogation of new candidates with "you just got out of computer science graduate school so you learn to present like this" whiteboard sessions, evolved in the hands of callous or worse managers, into a sort of repeated hazing of programmers, by non-programmers or the like.
red-iron-pine 3y ago
A lot of the back of the envelope math points to new hire failure rates of around 50%, where failure is if the candidate is still there in 18 months. Only about 19% would be considered "successful" hires. Also only about 11% were not successful due to purely lacking technical skills. like I'm sure we all have stories about a noob that doesn't know how to do something basic like ssh with a key or something, but in many cases that's not the issue. https://sowelo.eu/unsuccessful-recruitment/ https://sowelo.eu/unsuccessful-recruitment/
- FirmwareBurner 3y ago
  >I'm sure we all have stories about a noob that doesn't know how to do something basic like ssh with a key or something Not knowing how to do that, or not knowing how to google how to do that? I also joined some jobs as a noob when switching domains, where I was lacking skills that the more seasoned people would consider as being "basic", but I could also google what I was lacking most of the time and learn on the fly.