> One year they did extremely well and decided to simplify and give everyone a straight $10,000 bonus, from the janitor to the CEO
So the guy that watches YouTube videos for 3 hours, and the guy that takes 2 hours worth of smoke breaks a day got the same as the gal that works 50 hours a week and fixes the majority of the product issues?
Contributions are not equal so please don't treat them that way or you will only disenfranchise your BEST employees.
And remember, your best employees are the ones that have the most options elsewhere....
I expected this response, thanks for taking the time to write it.
The problem is one of statistics. Your employees will always, no matter how hard you try, fit on a bell curve.
Here's the hilarious way most people think about this bell curve (just imagine this defines the set of employees actually working at a company, not just new hires; because it does): http://www.intellectcorp.com/images/bell-curve-hiring.jpg
Say you have 100 employees. There will always be 4-5 who will be exceptional—the upper quintile. There will always be 4-5 who are at the bottom—the lower quintile. The problem is, this will always be true. If you fire the bottom 5, then there will be a new "bottom 5". People will always be worried about their job, the pressure of not being fired will be a negative-feedback motivator (or that's the idea), and the top people at the company will continue to be rewarded.
Two questions:
1. Do you truly think that the top employees are working harder because you reward them? How much harder?
2. What about the other 90 employees?
Your top employees aren't what you should be worried about. Firstly, if you depend on them and (as you say they have the most options elsewhere) heaven forbid they leave, where does that leave you? Fucked, that's where. Second, even if you think the top employees are doing the work of 2 people each, or even 3 people (which is a lot to expect of anyone), 80-90% of your work output is still being done by the thickest part of that bell curve.
What's the answer?
Improve the system. Making systematic, broad improvements will help everyone do better work. This is the true reason the straight 10k bonus worked so well: it was a systematic reward. But other things work even better: provide free lunch and your employees won't leave the office as much. Provide free dinner and they'll stay late. Bam, systematic across-the-board improvement in working hours and efficiency that no bonus would ever come close to. Allow frequent employee skill-improvement days, make them focus on honing their skill sets for 1 day a month without the stress of real work. Bam, everyone gets better, not just the top 5%. There are countless others.
Figure out how to make the 90% in the middle work the hardest—and trust me, it's not by promising them bonuses if they can make it into the top 5%. That might be motivating for the upper 10% or maybe 15%, but not the rest. Improvements to the system they're in are the only good answer. The antiquated structure of bonuses and rewards and the punishment of firing is a religious myth left over from protestant and catholic beliefs in punishment and shame. They're not based on behavior, they're not based on science, and they're simply not true.
Are monetary rewards entirely bad? No, of course not, people should be rewarded for doing good work. I personally think that this reward should be in the form of simple, understandable fair compensation, with as few barriers to that as possible. Other reward systems may work too. But they should not be the focus—the focus should be on improving the system your employees work in, as it will result in gains an order above what's possible with carrot and stick.
Interesting points and I'm glad you shared this information...seriously.
Something that I hadn't considered is that the largest gains in productivity could/would come from your "average" (so to speak) employees .
I incorrectly assumed that money is the most important carrot...like you say, possibly for the top 10%, but everyone else may be motivated in other ways.
After reading the Wikipedia article I'm honestly compelled to learn more about this man and his methods.
Appreciate that, thanks, and I'm happy to share it.
Deming was a fascinating guy. His ideas (mostly applicable to factories but easily applied to all companies) helped bring Japan out of its post-war recession and give it the respected reputation for high-quality goods they have had for some time. His concepts have been largely ignored since (partly due to his horribly dry writing style) but I think they're very valuable to our current economic situation in the west as well.
The focus is on quality, and the path is through process-based management. You have to realize your employees are a system just like your software is: mostly understandable, and statistically predictable. You can shape your system of people just like you can a manufacturing line, and traditional beliefs about punishments and rewards simply do not apply if you look at what's truly effective in improving quality and improving production.
It was remarkably successful in Japan partly because their philosophical backgrounds are so completely different: it is easier for someone brought up with ideas of Zen buddhism or other eastern religions to think of the whole rather than the individual; the process rather than the part.
In the US, where we're very much an individual and self-focused society, we track immediately to individual motivation without thought to how it affects the whole company. The unsubstantiated belief that "what's best for the individual is best for the whole" is embedded in us and very difficult to shake. Indeed, it's fairly core to our democracy and has ties to political beliefs as well. Deming himself had the breaking of this habit as one of his prerequisites (his System of Profound Knowledge): that everyone in the entire company must understand this holistic way of thinking, not just the management. Surely there is a balance to be struck, but Deming covers that as well, focusing on the psychology of personal motivation and work quality rather than belief systems. And it's all backed up by great process engineering and experience.
Honestly I think it's better if you completely ignore the monetary rewards and simply focus on what makes people work the best; the process of work. Here's another great TED talk on the subject: http://www.ted.com/talks/dan_pink_on_motivation.html
> even if you think the top employees are doing the work of 2 people each, or even 3 people (which is a lot to expect of anyone)
Claim: in software or engineering, your top employees are doing the work of a thousand people, or an infinite ratio of some of their fellow employees. No combination of administrative assistants at Google is going to program Google Chrome or even a sorting algorithm. No combination of HR people is capable of calculating structural loads.
Claim: (more arguable on Hacker News). Successfully managing a medium or large organization is likewise a skill shared by very few people. Managing humans is like programming on the most testy and finicky AWS nodes you could ever imagine. They don't do what they're told, they do other things, and they may return wrong results. You are responsible for making sure the group performs as a whole. This is engineering of a different sort.
There is a fundamental impedance mismatch between the amount of dollars required to sustain life (outside of serious health issues, this does not greatly vary from person to person) and the dollar productivity of each individual.
Arguably every great tech organization (commercial or non-commercial) has a few people who are 1000, 10000, or 1 million times as productive as others. People like Linus Torvalds, Richard Stallman, Jeff Dean, Urs Holzle, Guido van Rossum, or DHH. That is how they became great tech or engineering organizations.
The Bell Curve may describe the population distribution of intelligence (though ability on any given trait may be different). But as a manager or recruiter your duty is to select highly nonuniformly from the right tail of that curve. Your ability distribution should be as radically right skewed as possible, and should not really resemble a bell curve.
In short, "equal bonuses for all" -- with the same amount going to both the guy who didn't do much work and the guy who picked up after him -- is a great way to demotivate your best people. And this will cost you a lot more than 2-3 of your medium employees, it will cost you a guy who did 1000X their work, and perhaps it will cost your whole business. All that matters are your best, with them you can rebuild anything.
Nit: a quintile is a 5th of the population. The upper quintile would be the 20 people. You probably mean the upper vigintile (top 20th) if you want to limit it to 4-5 people.
> So the guy that watches YouTube videos for 3 hours, and the guy that takes 2 hours worth of smoke breaks a day got the same as the gal that works 50 hours a week and fixes the majority of the product issues?
If those guys have jobs at your company, you have bigger problems than your bonus structure. Unless they are doing more work than they appear to be doing, which is entirely possible.
Comments
> One year they did extremely well and decided to simplify and give everyone a straight $10,000 bonus, from the janitor to the CEO
So the guy that watches YouTube videos for 3 hours, and the guy that takes 2 hours worth of smoke breaks a day got the same as the gal that works 50 hours a week and fixes the majority of the product issues?
Contributions are not equal so please don't treat them that way or you will only disenfranchise your BEST employees.
And remember, your best employees are the ones that have the most options elsewhere....
I expected this response, thanks for taking the time to write it.
The problem is one of statistics. Your employees will always, no matter how hard you try, fit on a bell curve.
Here's the hilarious way most people think about this bell curve (just imagine this defines the set of employees actually working at a company, not just new hires; because it does): http://www.intellectcorp.com/images/bell-curve-hiring.jpg
Say you have 100 employees. There will always be 4-5 who will be exceptional—the upper quintile. There will always be 4-5 who are at the bottom—the lower quintile. The problem is, this will always be true. If you fire the bottom 5, then there will be a new "bottom 5". People will always be worried about their job, the pressure of not being fired will be a negative-feedback motivator (or that's the idea), and the top people at the company will continue to be rewarded.
Two questions:
1. Do you truly think that the top employees are working harder because you reward them? How much harder?
2. What about the other 90 employees?
Your top employees aren't what you should be worried about. Firstly, if you depend on them and (as you say they have the most options elsewhere) heaven forbid they leave, where does that leave you? Fucked, that's where. Second, even if you think the top employees are doing the work of 2 people each, or even 3 people (which is a lot to expect of anyone), 80-90% of your work output is still being done by the thickest part of that bell curve.
What's the answer?
Improve the system. Making systematic, broad improvements will help everyone do better work. This is the true reason the straight 10k bonus worked so well: it was a systematic reward. But other things work even better: provide free lunch and your employees won't leave the office as much. Provide free dinner and they'll stay late. Bam, systematic across-the-board improvement in working hours and efficiency that no bonus would ever come close to. Allow frequent employee skill-improvement days, make them focus on honing their skill sets for 1 day a month without the stress of real work. Bam, everyone gets better, not just the top 5%. There are countless others.
Figure out how to make the 90% in the middle work the hardest—and trust me, it's not by promising them bonuses if they can make it into the top 5%. That might be motivating for the upper 10% or maybe 15%, but not the rest. Improvements to the system they're in are the only good answer. The antiquated structure of bonuses and rewards and the punishment of firing is a religious myth left over from protestant and catholic beliefs in punishment and shame. They're not based on behavior, they're not based on science, and they're simply not true.
Are monetary rewards entirely bad? No, of course not, people should be rewarded for doing good work. I personally think that this reward should be in the form of simple, understandable fair compensation, with as few barriers to that as possible. Other reward systems may work too. But they should not be the focus—the focus should be on improving the system your employees work in, as it will result in gains an order above what's possible with carrot and stick.
Again, please read http://en.wikipedia.org/wiki/W._Edwards_Deming
Interesting points and I'm glad you shared this information...seriously.
Something that I hadn't considered is that the largest gains in productivity could/would come from your "average" (so to speak) employees .
I incorrectly assumed that money is the most important carrot...like you say, possibly for the top 10%, but everyone else may be motivated in other ways.
After reading the Wikipedia article I'm honestly compelled to learn more about this man and his methods.
I think you may have just made me less dumb ;-)
Appreciate that, thanks, and I'm happy to share it.
Deming was a fascinating guy. His ideas (mostly applicable to factories but easily applied to all companies) helped bring Japan out of its post-war recession and give it the respected reputation for high-quality goods they have had for some time. His concepts have been largely ignored since (partly due to his horribly dry writing style) but I think they're very valuable to our current economic situation in the west as well.
The focus is on quality, and the path is through process-based management. You have to realize your employees are a system just like your software is: mostly understandable, and statistically predictable. You can shape your system of people just like you can a manufacturing line, and traditional beliefs about punishments and rewards simply do not apply if you look at what's truly effective in improving quality and improving production.
It was remarkably successful in Japan partly because their philosophical backgrounds are so completely different: it is easier for someone brought up with ideas of Zen buddhism or other eastern religions to think of the whole rather than the individual; the process rather than the part.
In the US, where we're very much an individual and self-focused society, we track immediately to individual motivation without thought to how it affects the whole company. The unsubstantiated belief that "what's best for the individual is best for the whole" is embedded in us and very difficult to shake. Indeed, it's fairly core to our democracy and has ties to political beliefs as well. Deming himself had the breaking of this habit as one of his prerequisites (his System of Profound Knowledge): that everyone in the entire company must understand this holistic way of thinking, not just the management. Surely there is a balance to be struck, but Deming covers that as well, focusing on the psychology of personal motivation and work quality rather than belief systems. And it's all backed up by great process engineering and experience.
Honestly I think it's better if you completely ignore the monetary rewards and simply focus on what makes people work the best; the process of work. Here's another great TED talk on the subject: http://www.ted.com/talks/dan_pink_on_motivation.html
> even if you think the top employees are doing the work of 2 people each, or even 3 people (which is a lot to expect of anyone)
Claim: in software or engineering, your top employees are doing the work of a thousand people, or an infinite ratio of some of their fellow employees. No combination of administrative assistants at Google is going to program Google Chrome or even a sorting algorithm. No combination of HR people is capable of calculating structural loads.
Claim: (more arguable on Hacker News). Successfully managing a medium or large organization is likewise a skill shared by very few people. Managing humans is like programming on the most testy and finicky AWS nodes you could ever imagine. They don't do what they're told, they do other things, and they may return wrong results. You are responsible for making sure the group performs as a whole. This is engineering of a different sort.
There is a fundamental impedance mismatch between the amount of dollars required to sustain life (outside of serious health issues, this does not greatly vary from person to person) and the dollar productivity of each individual.
Arguably every great tech organization (commercial or non-commercial) has a few people who are 1000, 10000, or 1 million times as productive as others. People like Linus Torvalds, Richard Stallman, Jeff Dean, Urs Holzle, Guido van Rossum, or DHH. That is how they became great tech or engineering organizations.
The Bell Curve may describe the population distribution of intelligence (though ability on any given trait may be different). But as a manager or recruiter your duty is to select highly nonuniformly from the right tail of that curve. Your ability distribution should be as radically right skewed as possible, and should not really resemble a bell curve.
In short, "equal bonuses for all" -- with the same amount going to both the guy who didn't do much work and the guy who picked up after him -- is a great way to demotivate your best people. And this will cost you a lot more than 2-3 of your medium employees, it will cost you a guy who did 1000X their work, and perhaps it will cost your whole business. All that matters are your best, with them you can rebuild anything.
"All that matters are your best"
I agree with some of your ideas, but in my humble opinion, this attitude will cost you your business faster than anything.
Yea, but that can be solved by using separate bonuses for each role.
Nit: a quintile is a 5th of the population. The upper quintile would be the 20 people. You probably mean the upper vigintile (top 20th) if you want to limit it to 4-5 people.
> So the guy that watches YouTube videos for 3 hours, and the guy that takes 2 hours worth of smoke breaks a day got the same as the gal that works 50 hours a week and fixes the majority of the product issues?
If those guys have jobs at your company, you have bigger problems than your bonus structure. Unless they are doing more work than they appear to be doing, which is entirely possible.