6 ms·

Notice that, when they say they only hire the top 1%, there's a vicious sampling bias at work: even in a worst case where they're being literal they're actually

by saulrh 8y ago

Notice that, when they say they only hire the top 1%, there's a vicious sampling bias at work: even in a worst case where they're being literal they're actually hiring the top 1% of their applicants, not the top 1% of engineers. Think about how many jobs you usually apply for before getting one. Now imagine someone that's less successful than you at getting hired and think about how many times they're interviewing. Now think about the population that the interviewer sees. Now add college students that are just practicing and people that are required to interview for every job on the board to maintain unemployment. Now think about how much time interviewing you spend for every day working and how that might change as a function of quality - staying longer at better places, maybe? I obviously don't have any statistics, but I think I could imagine a situation where the top 95% of software people constitute the top 1% of the interviewing population. Also one where interviewers are terrible at math, but that's no surprise. Anyway, you're fine, just go for it. That fact that you are self-aware enough to worry about it puts you in good standing all by itself.

nestorD 8y ago
There is a great article on the fact that everybody hires only the top 1% : https://www.joelonsoftware.com/2005/01/27/news-58/ https://www.joelonsoftware.com/2005/01/27/news-58/
- spectre256 8y ago
  The fact that people still parrot the top 1% idea after it was pretty thoroughly debunked by Joel almost 15 years ago is...interesting.
  - mmirate 8y ago
    When you discover that an intuitive result is wrong, the production of a thorough debunking is but a tiny sliver of the total work needed to change the world's minds away from that wrong result.
  - jobigoud 8y ago
    We are still making many of the mistakes laid out in The Mythical Man-Month so…
cirgue 8y ago
We hire the top 5% of applicants. This says a lot about our applicant pool. It says very little about our team. No one here is from a top program. Many don't even have CS degrees.
- pmiller2 8y ago
  How do you measure this? I would wager that, at best, you're hiring the top 5% of applicants based on interview performance. It's likely you know little or nothing about what that says vis a vis job performance, because your interview probably doesn't exercise skills used on the job, except in a very tangential way.
  - eberkund 8y ago
    I think he just means that they hire 5% of applicants, by definition they are the top 5% because they are the ones they chose to hire. Unless of course they have a significant number of people going through the entire interview process only to decline the offer.
    - pmiller2 8y ago
      That's pretty close to the point I was making. They call it the top 5% because 5% of people pass their interview. The company doesn't have any idea whether those people (out of the entire pool) are the best candidates to do the job they were hired to do, because the company probably doesn't measure that. And, they sure don't know that their "top 5%" is representative of the top 5% of all engineers. For starters, how do you measure that?
- rorykoehler 8y ago
  No you don't. I know you don't because you have no idea who the top 5% of applicants are. Not after interviews, not after 6 months on the job, not even 10 years in the job. You can't distill something as complex as humans working in teams that form a part of a larger organisation into a single number like that. That's not how the world works. For example Bob was doing great work but then his wife cheated on him and he is doing less great work. Is Bob in the top 5% or not?
Spooky23 8y ago
I've observed that in pursuing the 1% unicorn candidate, many people step in unicorn shit and hire smart people who are bad employees. My goal is to not hire the bottom 50% of the workforce in a given discipline. IMO, you're better off building teams that make the average+ person more productive.
- cheez 8y ago
  Yes! I love to build teams with average people that create above average results due to their configuration. It's my secret sauce.
  - Double_a_92 8y ago
    Reminds me of that baseball movie where a manager hired low-value players, that very extremely good at only something.
    - chupasaurus 8y ago
      "Moneyball" is the name of both the movie and the book it was based on.
- fizx 8y ago
  >I've observed that in pursuing the 1% unicorn candidate, many people step in unicorn shit and hire smart people who are bad employees. What's fun is that this isn't just a possibility, its likely! Despite the following being a clearly artificial argument, the general shape of it is true. Let's say that you can rate employees 1-10 on each of technical skill and employability/soft-skills, and that both are equally sought by the market. Given how attractive your company is to work for and what you pay, its quite often true that you can allocate around 12 total points between the two categories. 13 point candidates have better offers from Google; 11 point candidates fail your interview process. If you only hire people that strongly pass a strong technical interview, ensuring that their technical skill is at least an 8/10, then your employees will have at most 4/10 soft skills.
  - 8y ago
ethagnawl 8y ago
> people that are required to interview for every job on the board to maintain unemployment. I'm pretty sure you're being hyperbolic, but that sounds hellish. I have GAD and dread interviewing so much that I'd forego unemployment and just work as a Task Rabbit or something. To preempt the inevitable responses: Yes, I do have a family to support and, yes, I have worked as a Task Rabbit or done manual labor when faced with similar situations in the past.

by saulrh 8y ago

nestorD 8y ago

There is a great article on the fact that everybody hires only the top 1% : https://www.joelonsoftware.com/2005/01/27/news-58/ https://www.joelonsoftware.com/2005/01/27/news-58/

spectre256 8y ago
The fact that people still parrot the top 1% idea after it was pretty thoroughly debunked by Joel almost 15 years ago is...interesting.
- mmirate 8y ago
  When you discover that an intuitive result is wrong, the production of a thorough debunking is but a tiny sliver of the total work needed to change the world's minds away from that wrong result.
- jobigoud 8y ago
  We are still making many of the mistakes laid out in The Mythical Man-Month so…

cirgue 8y ago

We hire the top 5% of applicants. This says a lot about our applicant pool. It says very little about our team. No one here is from a top program. Many don't even have CS degrees.

pmiller2 8y ago
How do you measure this? I would wager that, at best, you're hiring the top 5% of applicants based on interview performance. It's likely you know little or nothing about what that says vis a vis job performance, because your interview probably doesn't exercise skills used on the job, except in a very tangential way.
- eberkund 8y ago
  I think he just means that they hire 5% of applicants, by definition they are the top 5% because they are the ones they chose to hire. Unless of course they have a significant number of people going through the entire interview process only to decline the offer.
  - pmiller2 8y ago
    That's pretty close to the point I was making. They call it the top 5% because 5% of people pass their interview. The company doesn't have any idea whether those people (out of the entire pool) are the best candidates to do the job they were hired to do, because the company probably doesn't measure that. And, they sure don't know that their "top 5%" is representative of the top 5% of all engineers. For starters, how do you measure that?
rorykoehler 8y ago
No you don't. I know you don't because you have no idea who the top 5% of applicants are. Not after interviews, not after 6 months on the job, not even 10 years in the job. You can't distill something as complex as humans working in teams that form a part of a larger organisation into a single number like that. That's not how the world works. For example Bob was doing great work but then his wife cheated on him and he is doing less great work. Is Bob in the top 5% or not?

Spooky23 8y ago

I've observed that in pursuing the 1% unicorn candidate, many people step in unicorn shit and hire smart people who are bad employees. My goal is to not hire the bottom 50% of the workforce in a given discipline. IMO, you're better off building teams that make the average+ person more productive.

cheez 8y ago
Yes! I love to build teams with average people that create above average results due to their configuration. It's my secret sauce.
- Double_a_92 8y ago
  Reminds me of that baseball movie where a manager hired low-value players, that very extremely good at only something.
  - chupasaurus 8y ago
    "Moneyball" is the name of both the movie and the book it was based on.
fizx 8y ago
>I've observed that in pursuing the 1% unicorn candidate, many people step in unicorn shit and hire smart people who are bad employees. What's fun is that this isn't just a possibility, its likely! Despite the following being a clearly artificial argument, the general shape of it is true. Let's say that you can rate employees 1-10 on each of technical skill and employability/soft-skills, and that both are equally sought by the market. Given how attractive your company is to work for and what you pay, its quite often true that you can allocate around 12 total points between the two categories. 13 point candidates have better offers from Google; 11 point candidates fail your interview process. If you only hire people that strongly pass a strong technical interview, ensuring that their technical skill is at least an 8/10, then your employees will have at most 4/10 soft skills.
- 8y ago

ethagnawl 8y ago

> people that are required to interview for every job on the board to maintain unemployment. I'm pretty sure you're being hyperbolic, but that sounds hellish. I have GAD and dread interviewing so much that I'd forego unemployment and just work as a Task Rabbit or something. To preempt the inevitable responses: Yes, I do have a family to support and, yes, I have worked as a Task Rabbit or done manual labor when faced with similar situations in the past.