3 ms·

The problem with the 2000s-era Microsoft method, as I understand it, was that the population sizes they used were too small. Stack ranking was done at the indiv

by mjr00 4y ago

The problem with the 2000s-era Microsoft method, as I understand it, was that the population sizes they used were too small. Stack ranking was done at the individual team level, and that did indeed lead to people sabotaging each other, because in a team of 10 only 2 people could get the highest pay raises and 1 person would be placed on a PIP (or whatever the ratios were). But yes, all major companies still do something like this AFAIK, just with a larger population size now. Speaking from my experience at Amazon, the target was around 4%, but this was across entire departments. My department was 300 people. Is it possible that you can have a department of 300 people where you can't find 12 people who are below-average performers? Yes. Is it likely? No. I don't know how many people the lead in the article managed. If it was a team of 15, he has a legitimate gripe. If it was a team of 200, not so much.

whatshisface 4y ago
Who's the next 4% after you lay off the last 4%?
- georgemcbay 4y ago
  Hopefully your hiring processes suck so you can keep refreshing a new batch of people to get fired!
  - mjr00 4y ago
    If you have a hiring process that results in only a 4% bad hire rate, create a business right now because every business in the world will want to pay you massive sums of money for the secret to such an amazing process. from my own experience, a bad hire rate of 15-20% is what I usually see accepted across the industry.
    - whatshisface 4y ago
      Let me check the math here... 20% of the 4% rehired is 1/5th%, leaving 3 and 4/5ths of a percent of to come from the better performers.
      - mjr00 4y ago
        You're incorrectly assuming 4% are let go, which is untrue as many are provided coaching/guidance to fix performance issues, and not accounting for additional headcount.
  - yborg 4y ago
    Managers who have a solid team will deliberately hire a drone to sacrifice annually, and you can blame the hiring process for a long time.
acdha 4y ago
That’s one of the two but the bigger problem is how you measure performance. It’s notoriously hard to measure outside of the simplest jobs and once you tell people they’ll be punished unless they juice a particular number, well, that encourages all kinds of undesirable behaviour: rigging tickets or other metrics, encouraging people not to do maintenance programming or hard tasks, setting competitors up to fail, not doing things to help or train less experienced colleagues because that isn’t captured, etc. Yes, managers are supposed to keep that from happening but this policy tells them that a) you don’t trust them to do their job, b) their job now focuses on protecting their people from the system, and c) being busy playing the same game on their own personal behalf.
- mjr00 4y ago
  The way it was done at amzn was to get all managers for the department together in a room and provide information about each of their team members and why they were ranked the way they were. If an employee was doing a small number of hard tasks and providing a lot of mentorship it was quite the opposite, they were much more likely to get a "highly valued" ranking than be marked as a low performer. Obviously if you use stupid metrics like lines of code or story points closed you will get stupid results.
  - acdha 4y ago
    How did they reconcile that with a hard quota? I can see that working with a degree of trust but not a fixed percentage.
    - mjr00 4y ago
      Largely you're still ranking your performers from top to bottom, so it's a matter of placing the line at the bottom in the right place to get the desired percentage. Even then there was still a lot of debate among department managers around people on the borderline least effective category. There were usually very clear outliers who were definitely LE (people clearly coasting, people who didn't show up to work regularly), and sometimes that was enough. Other times managers did have to push each other to be brutally honest about those who were borderline, essentially asking, "If you fired this person and got a replacement, would the replacement be more effective than the current person within 6 months?" Keep in mind it was never a hard quota, though in my case at amzn I did have a few times where the GM (so manager of ~1500 org) pushed back and said not enough people were in LE for our department, in which case a few borderline people did unfortunately have to get put into LE.
fatneckbeard 4y ago
most Amazon departments lose money. so whatever psychopath games they are playing, its not working.
- ilyt 4y ago
  Actually lose money or provide value to other departments that earn money ?
ilyt 4y ago
>Is it possible that you can have a department of 300 people where you can't find 12 people who are below-average performers? Yes. Is it likely? No. I mean, below average performers will be around 40-60% just because how averages work. The problem is that the question is "is their performance low compared to team", while the question you should be answering is "is their performance low compared to the market. If team's management is on point with hiring the "worst" won't be that much worse than average anyway, so if you want to replace the bottom few % you essentially waste time as new hires might be at same productivity level. On other side if you recruiting sucks and 20% of the dept does the 80% of the job, it isn't going to help either.
grayfaced 4y ago
No manager has direct insight into performance of 300 people (at least in tech). They have to rely on subordinate managers/teamleads, each of which can be unreliable for different reasons. Some may be playing politics, some may just be bad managers. Some will overhype their employees, some will be honest to a fault. Get them all in a ranking board, and all sorts of games get played. If your team leads could be trusted to measure appropriately, you wouldn't need stack rankings because they'd correctly identify those under-performing. Stack ranking is conceding that your managers can't properly measure performance. Shouldn't that reflect more on management?
RandallBrown 4y ago
The stack ranking on my team at MS was done across a team closer to ~60 developers. Each manager managed 5 or 6 and then they would all get together and fight for their team members. It was insanely toxic. Not sure if every team was the same.
dancingBugz24 4y ago
[dead]

by mjr00 4y ago

whatshisface 4y ago

Who's the next 4% after you lay off the last 4%?

georgemcbay 4y ago
Hopefully your hiring processes suck so you can keep refreshing a new batch of people to get fired!
- mjr00 4y ago
  If you have a hiring process that results in only a 4% bad hire rate, create a business right now because every business in the world will want to pay you massive sums of money for the secret to such an amazing process. from my own experience, a bad hire rate of 15-20% is what I usually see accepted across the industry.
  - whatshisface 4y ago
    Let me check the math here... 20% of the 4% rehired is 1/5th%, leaving 3 and 4/5ths of a percent of to come from the better performers.
    - mjr00 4y ago
      You're incorrectly assuming 4% are let go, which is untrue as many are provided coaching/guidance to fix performance issues, and not accounting for additional headcount.
- yborg 4y ago
  Managers who have a solid team will deliberately hire a drone to sacrifice annually, and you can blame the hiring process for a long time.

acdha 4y ago

That’s one of the two but the bigger problem is how you measure performance. It’s notoriously hard to measure outside of the simplest jobs and once you tell people they’ll be punished unless they juice a particular number, well, that encourages all kinds of undesirable behaviour: rigging tickets or other metrics, encouraging people not to do maintenance programming or hard tasks, setting competitors up to fail, not doing things to help or train less experienced colleagues because that isn’t captured, etc. Yes, managers are supposed to keep that from happening but this policy tells them that a) you don’t trust them to do their job, b) their job now focuses on protecting their people from the system, and c) being busy playing the same game on their own personal behalf.

mjr00 4y ago
The way it was done at amzn was to get all managers for the department together in a room and provide information about each of their team members and why they were ranked the way they were. If an employee was doing a small number of hard tasks and providing a lot of mentorship it was quite the opposite, they were much more likely to get a "highly valued" ranking than be marked as a low performer. Obviously if you use stupid metrics like lines of code or story points closed you will get stupid results.
- acdha 4y ago
  How did they reconcile that with a hard quota? I can see that working with a degree of trust but not a fixed percentage.
  - mjr00 4y ago
    Largely you're still ranking your performers from top to bottom, so it's a matter of placing the line at the bottom in the right place to get the desired percentage. Even then there was still a lot of debate among department managers around people on the borderline least effective category. There were usually very clear outliers who were definitely LE (people clearly coasting, people who didn't show up to work regularly), and sometimes that was enough. Other times managers did have to push each other to be brutally honest about those who were borderline, essentially asking, "If you fired this person and got a replacement, would the replacement be more effective than the current person within 6 months?" Keep in mind it was never a hard quota, though in my case at amzn I did have a few times where the GM (so manager of ~1500 org) pushed back and said not enough people were in LE for our department, in which case a few borderline people did unfortunately have to get put into LE.

fatneckbeard 4y ago

most Amazon departments lose money. so whatever psychopath games they are playing, its not working.

ilyt 4y ago
Actually lose money or provide value to other departments that earn money ?

ilyt 4y ago

>Is it possible that you can have a department of 300 people where you can't find 12 people who are below-average performers? Yes. Is it likely? No. I mean, below average performers will be around 40-60% just because how averages work. The problem is that the question is "is their performance low compared to team", while the question you should be answering is "is their performance low compared to the market. If team's management is on point with hiring the "worst" won't be that much worse than average anyway, so if you want to replace the bottom few % you essentially waste time as new hires might be at same productivity level. On other side if you recruiting sucks and 20% of the dept does the 80% of the job, it isn't going to help either.

grayfaced 4y ago

No manager has direct insight into performance of 300 people (at least in tech). They have to rely on subordinate managers/teamleads, each of which can be unreliable for different reasons. Some may be playing politics, some may just be bad managers. Some will overhype their employees, some will be honest to a fault. Get them all in a ranking board, and all sorts of games get played. If your team leads could be trusted to measure appropriately, you wouldn't need stack rankings because they'd correctly identify those under-performing. Stack ranking is conceding that your managers can't properly measure performance. Shouldn't that reflect more on management?

RandallBrown 4y ago

The stack ranking on my team at MS was done across a team closer to ~60 developers. Each manager managed 5 or 6 and then they would all get together and fight for their team members. It was insanely toxic. Not sure if every team was the same.

dancingBugz24 4y ago

[dead]