The problem I have with stack ranking, and one that I would think engineering companies like MSFT and AMZN would share, is that it is just bad science.
I can get on board with the idea that it is good to drop the bottom 10% of the company on a regular basis. But they're essentially taking very small sample sizes (individual teams), and assuming their distribution matches that of the entire company.
Or to put it another way, let's say you ranked every employee in the company 1-40,000 or whatever. You are trying to drop employees 36,001-40,000. An absolutely awful way to make that happen would be to create 4000 small groups of 10 people and then drop the bottom person from each group. Anybody who's even somewhat proficient in stats can tell you that this setup will yield results WAY off the mark. You could even quantify it with some simple analysis, and though I don't have time to do so, I'm sure the results would be horrifying.
And that's assuming a totally random distribution of people, which isn't true, and totally ignoring psychological effects of such a system. But at the very least, stack ranking is just bad science, and that should matter to these companies.
The idea is to create specific psychological effects.
Suppose that of the three attributes motivation, initiative, and talent two of them are very much within the control of the employee and the other less so.
The goal of the incentive is to make the employee want to maximize all three rather than doing other things (such as wasting time, etc.).
Imagine the slow, cynical organizations that give everyone a 5% bonus year in and year out no matter what. Just show up, put in your 9 to 5, and collect your bonus.
Bonus/incentive programs are intended to create more of a dynamic. Yes it's more stressful but the idea is that it's worth it.
But the motivation given is to beat the people around you. Which in many cases does not help the company as a whole. It also puts a damper on team coordination.
Nobody here is suggesting a total lack of merit pay. We're just saying that it is awful to assume that in any small group in an organization there is always someone who deserves to be fired, and someone who deserves a big bump, etc.
I wonder if the following would work: company is divided into small to mid-sized teams. Say 5 to 30 people. Each team has to do something that results in profit. More profit from a team implies larger raises and bonuses for that team. There's a catch: if the company as a whole doesn't meet a certain profit goal then no teams receive raises or bonuses. It seems like this structure encourages teammates to cooperate with each other in the "we all succeed or fail together" theme. Also there's some incentive for team A to ensure team B is successful since team B's failure could bring the entire company down. The psychological implications could be interesting. Seems like there's a snowball effect: n successful teams => team n+1 will probably be successful. This effect probably works in reverse: n lousy teams could imply that team n+1 is highly unlikely to succeed.
I'm no management guru but I can see a couple of major problems:
(1) Some teams are necessary but won't make a direct profit, eg. internal IT, customer service.
(2) It could encourage a team to make a profit at the expense of another team; eg. the Office team might release a buggy product quickly (maximizing their profit by not testing it) knowing that the customer service team are the ones who'll suffer a loss supporting it.
Good points. Some creativity is in order when it comes to structuring the teams. For example the Office team could be responsible for customer support themselves! I'm only half joking. Think about the incentive that creates for making an intuitive and easy-to-use product.
>>There's a catch: if the company as a whole doesn't meet a certain profit goal then no teams receive raises or bonuses.
No this will be a disaster, Expect severe attrition in such a company. No one wants to work and then not get paid, just because you missed the mark by a meager x%.
>>It seems like this structure encourages teammates to cooperate with each other in the "we all succeed or fail together" theme
Read this as, we all succeed or go and work elsewhere.
To be clear employees are still paid their salaries when the company doesn't meet its quota or mark. Just the bonuses and raises are held back. Either way I think you're right about high attrition rates.
I have lots of criticisms of a stack type of system. It should be reset more than once per year, should be light hearted, and one big part of it should be being a team player, etc.
My point is just that it's easy to have the whole team become fairly complacent and start focusing on other activities. Do you want your star team member spending all her time collaborating on a non-business-relevant fun project, or pouring that energy into work stuff to earn the $50K bonus.
I can get on board with the idea that it is good to drop the bottom 10% of the company on a regular basis. But they're essentially taking very small sample sizes (individual teams), and assuming their distribution matches that of the entire company.
Or to put it another way, let's say you ranked every employee in the company 1-40,000 or whatever. You are trying to drop employees 36,001-40,000. An absolutely awful way to make that happen would be to create 4000 small groups of 10 people and then drop the bottom person from each group. Anybody who's even somewhat proficient in stats can tell you that this setup will yield results WAY off the mark. You could even quantify it with some simple analysis, and though I don't have time to do so, I'm sure the results would be horrifying.
And that's assuming a totally random distribution of people, which isn't true, and totally ignoring psychological effects of such a system. But at the very least, stack ranking is just bad science, and that should matter to these companies.