It is good to see this discussed, but the principle is not news. In http://www.amazon.com/First-Break-All-Rules-Differently/dp/0... the point is made that average managers try to bring up their bottom performers, while the best managers long ago concluded that you get more mileage out of investing energy in the top ones.
Of course the world is full of average people. They have to wind up somewhere. And organizations full of average people are never going to be able to take the advice to work the superstars.
I'll believe that the slogan invest in your best has been internalized by our society when we devote serious resources to making sure that people with IQs in the top 1% stop dropping out of school faster than people with median IQs do. Anyone care to give me odds on this happening in the next 20 years? I'll take the "No" side.
Do you really think people with higher IQ dropping out of school is an issue? Firstly, there are some many other factors at play making a person a "supertstar" (emotional stability, finding out what they love, creativity) but let's agree IQ is one of them and an important one. I'd argue that people on the top 1% are statistically people more interested in learning varied subjects. So unless the person had serious family issues, which is something hard to devote serious resources to pinpointedly correct, whether they drop out of school is not as relevant as them finding what they love and having the tools to learn it, on their own if they're so inclined.
I agree with you that IQ is far from a complete measure of what matters, but I do think that it is an issue.
The reason why is because jobs that are likely to use the abilities of a high IQ person generally require a high school diploma, and frequently require a college degree. Therefore denying these people an equal opportunity to get those credentials limits how effectively society benefits from their abilities.
Also I should note that the matter is personal. I have a good IQ and yet I came within an inch of failing to complete high school. Were it not for a teacher named Bernie Bowker, I would not have graduated, gone to college, or had any prospects of getting jobs where my abilities would be useful. I think that that would have been a tragedy, particularly for me.
This is already happening, in a sense. Elite colleges face extreme pressure to keep their 4 and 6y graduation rate up, in part because successful parents are turned off by sending their kids to schools for outrageous tuition only for them to leave without a degree.
Caltech is perhaps most known for this problem, with graduation rates that were <85%, but changes were made to stop people from "flaming out". Harvard has a 98% 6y rate -- University of Oklahoma is 64%
This kind of helps explain Japan's lackluster performance over the last 20+ years for me. The typical corporate culture there seems to focus on getting the mediocre and lagging employees to perform up to par (mind you, their systems for accomplishing this are rather impressive), rather than sharpening the skills and maximizing the contributions of their Ace players (they definitely exist, but are definitely not given their due compared to their output).
Japanese mega corps are almost designed to be pathological for their best performing employees. I don't know if that is significant enough to have macroeconomic effects (all explanations of that nature strike me as suspicious just-so-stories), but if you hypothetically think you're good at something, a traditionally managed Japanese megacorp will attempt to a) dilute that aptitude, b) put you on a career track which requires decades to reach the autonomy you'd have at a US company in months or a startup in minutes, and c) compensate quality like a teacher's union does (I.e. forbid it and treat anyone asking for it as a threat to social harmony).
That is not projection, btw. I played the game pretty well for 3 years. That is just me describing a small portion of the well-understood rules of the game.
This is an interesting juxtaposition to the assertion that Japan tech companies have lost ground in the market due to a perfectionism that leads to a 5% better product at 50% the cost (can't remember where I read that, but I think it was here).
It's a well-circulated trope that Japanese goods tend to have higher quality but less innovation. It's believed that very few disruptive innovations come out of Japan compared to the USA/Europe/India. This would fit in with the general model of Japan focusing on stability and getting the same level of performance out of everyone. I haven't done the research to test whether this hypothesis is true or not, though.
One thing that backs up your argument is the fact that Japan has extremely low income inequality[1]. This has both good parts and bad parts-it's probably useful for society as a whole in a lot of ways, but it means that if you're truly better than average you'll earn a lot less.
Anecdotally the top-performing 20-somethings I know in Japan make at most 2x the average salary, whereas the top-performing 20-somethings in the USA easily make 5-10x average (just look at how much a Ruby Developer, top-tier lawyer, or investment banker makes compared to the average income).
The problem is most places don't even have a way to find out a average manager. The reason is that management layers in almost every big company works like a cartel, they know darn well how to preserve their self interests. Generally you don't get to know until its often too late to do anything about it. The people who suffer the most are not bad performers, but those who fall out of managements favor or people who just can't get political and benefit out of it. The actual performers are likely to be named arrogant mavericks who are bad at team work. They won't be fired, but they will be treated so badly, until they themselves see the wisdom in leaving.
Also this whole GE analogy in these kind of articles is MBA speak. That is talking at high abstract levels, without absolutely knowing what they are talking about. Stack ranking worked at GE not because Stack Ranking is awesome, but because there was some one like Jack Welch to make it work.
Adhering to stack ranking to the letter but not in spirit will basically amplify the problem. Guess who gets the least rank in a political system?
The core insight-- that mediocrity and failure provide about the same payoff (zero) in a modern technological economy-- is accurate.
Here's where the divergence happens. Being a top performer requires a lot, and not all of it's intrinsic. Fit with tools and project are important, and so are relationships. There aren't intrinsic A and B players. For example, most software projects have one or maybe two "blazers" who write most of the new code and make the important decisions, and many more who spend the bulk of their time keeping up. It's a lot better for your career to be in the former category; get out if you're in the latter. This is what language and tooling wars are often about in companies: people wanting to change the landscape so they can be the trailblazers and the others end up as the followers. This is why it's economically worth it (in a corporate context) to invest so much energy into tooling and language fights instead of on the work itself: if you win the battle, the leverage it applies to your work is immense. The truth is that most projects have a rank ordering that becomes self-perpetuating, because the high-ranking people take the best work and progress faster. There isn't one global equilibrium where the "best" people are magically matched to the best work, but N! stable equilibria, where N is the number of participants.
Why is this important? Well, the correct insight is that the payoff curve (between input factors and output) is now convex. In a concave world, you wanted to manage toward the middle: reduce variance, bring up the slackers. This is what management is good at: bringing everyone in to the middle of the performance curve. That's also why traditional management is failing in the technological world. Maximizing a concave function is equivalent to minimizing a globally convex one, which is the easiest class of optimization problems (there's one global optimum). For these, incremental evolution (gradient descent) works: you converge to an optimal solution (in economic terms, an equilibrium). It's not equipped to handle the convex world where there are zillions of possible local optima and non-local insight is required in order to find better solutions.
What's inaccurate is the idea that there are "intrinsically" mediocre vs. excellent people and that the former should just be discarded. Why do we have to stay in school until we're in our mid-20s? Because our economic contributions are worthless until we've had an amount of schooling that would be considered immense by historical standards. The reason for all this schooling is to increase the probability of a person ending up as a "10x" contributor in an economy that has no place for 1x contributors.
As for stack ranking, that's something separate, and it's actually harmful to the top performers. Stack ranking is about intimidation-- nothing else. The idea is ancient: Romans called this method of intimidation and punishment decimation. Split the regiment into groups of 10, of which one will die. Sometimes it was random (drawing lots) and sometimes the groups voted. The reason for the 10% cut was because (a) it's not enough of a force reduction to "disrupt continuity" but (b) it was large enough to scare the shit out of people. People use "decimation" as a synonym for "devastation" but that's not correct, because "decimation" is a controlled, small reduction of force. Stack ranking is the same concept, but with political infighting used instead of randomness to make the decisions, and the "divide-and-conquer" campaign it wages against the proles is intentional.
Intimidation's effect is performance-middling. It kills creativity and innovation, and the top performers suffer, but it brings the slackers closer to the median. Traditional, intimidation-based management worked well in a concave world where people were mostly doing low-margin commodity work and one slacker could cancel out 5 good workers, but it fails catastrophically in the convex world.
Are you suggesting that different languages and tooling would make different players on the team 'A' players? Sorry, but I find this to be an extremely dubious claim.
At the risk of coming off as a completely arrogant ass, I've been an 'A' player in two completely different language, stack and tooling environments so far. Now I find myself shifting the lion's share of my coding time from back-end to front-end work, but aside from an initial adjustment period, I don't find this to be a huge hit to my productivity or effectiveness.
I'm pretty confident most other 'A' players can do the same easily, probably more easily than I can. Am I unhappy in an environment that forces to use tooling that I find to be unproductive, sure. I've been here before, but nonetheless I was still one of the 'A' players.
A players are not different because of having the right fit with tools or projects, or because of more (or less) education. They are different because they have a good fit with their purpose combined with passion for what they do.
Are you suggesting that different languages and tooling would make different players on the team 'A' players? Sorry, but I find this to be an extremely dubious claim.
It depends how far apart they are.
Having written the code is a huge advantage. Getting to write code is a major advantage (if most of your time is spent reading code, your odds of transferring to a new-dev role on the same team are low). Picking tools that one knows well is also an advantage, but probably less of one, because there are far crappier corporate codebases than in-use open source tools (in general).
If the people are very far apart in ability, then those kinds of changes won't permute the ranking... on the other hand, if there's that much spread among programmers of equal rank, then you have an unusual team and a recipe for chaos.
It's not just about languages. It's about environments and coding practices as well. Sure, an A programmer can write quality code in any language. It doesn't take very long to learn a new programming language itself. Learning a corporate codebase and an environment (often broken) takes a lot longer. Three to six months, easily. Reputations are formed, and people are slotted rather inflexibly into the pecking order, in less time than that.
Being an A programmer won't stop you from writing quality Java code if you get to develop from scratch. The crappiness of the language may hold you back, but not at a career-killing rate, and it won't bring you below expectations because no one expects high productivity when Java is the language of choice. However, if you're dropped into an existing codebase and the code has a bunch of VisitorFactory garbage, being an A programmer hurts you. Good programmers just don't have much patience (or experience, if they can help it) with shit code.
But I think that really invalidates this whole premise that there is no such thing as 'A' players and that they are primarily a product of the tooling and environment.
Teams with shit code and poor unproductive environments, pretty much by definition, don't have 'A' players and the ones they do have don't stay long. Someone once told me "it's easy to retain mediocrity in a company, it's the talented people that are hard to keep." This is a chicken-and-egg problem, but in my mind it is the lack of quality team members and managers that leads to a poor environment all around.
That being said I've worked in environments like this, where tooling was poor, the build chain almost non-existant and there was a legacy of extremely poor architecture. I was still able to make improvements, bring in better code, architectures and dev practices (albeit slowly). For that matter, so was every other 'A' player I know that worked there. The main thing is none of us stayed there more than a year.
Even for Welch, as I understand it, he was mostly firing areas that were under performing by the metrics they had. So they did have something: we reach the sales numbers we expect from that division or it gets the ax. I don't think this kind of strategy will work for all kinds of companies.
>> There's arguably never been a worse time to be a mediocre, average or typical employee. For most firms today, mediocrity is a cost to be managed and a burden to be borne, not a potential to cultivate or a resource worthy of serious investment. Indeed, if an app, web service or an outsourcer can deliver 80% of typical or average performance at 35% to 60% of the cost, then why pay for average employees? These days, no organization can afford to pay a premium for mediocrity.<<
The Indian outsourcing and bodyshopping cartels which thrive on mediocrity and cost-arbitrage would actually consider that to be awesome news, for it hints that they will get more work to keep costs low. But what they fail to realize is that another cheaper destination is emerging elsewhere. :) One man's poison is another's pie.
Somewhat related idea: given an employee who's stellar in some areas and not so stellar in others, do you try to even their attributes out, or do you develop the areas where they're stellar even further?
So many people are focused on bringing everyone up to some base level of mediocrity in every imaginable area, but what if we instead focused on finding brilliance, then doing everything we can to amplify that brilliance?
Which might, of course, include working on some of those areas for improvement, but intelligently, with the goal of keeping those weaknesses from impeding excellence elsewhere.
Gallup has studied this exact question. They have found that working on areas of brilliance, and structuring jobs to work around areas of mediocrity, is hands down massively better than trying to address weaknesses.
Hand waving and made up numbers. It's kind of annoying when someone uses precise numbers that they pulled out of the air with no empirical basis as the foundation of an argument.
At the end he asks about the organizations where 60% of value is produced by 40% of employees. Which organizations? How is value measured? No telling because the numbers are just made up.
Thanks for saying that. I must be getting old and grumpy because that's an idiom that gets under my skin for no good reason as opposed to my prior hatred for the "fail" idiom which I am legitimately annoyed by due to its very nature as pessimistic armchair denigration.
If "agreed" were used instead, it'd perform the same function without being called out by people who don't like internet linguistic trends. Clearly indicating one's position in an argument is valuable, and "this" accomplishes that.
Even though GE is generally considered well run, let's not forget that it needed a bailout by Warren Buffett during the financial crisis.
At risk of sounding like I am defending my own mediocrity, the average worker and manager are going to be the ones who can last the distance.
Top performers have a nasty habit of getting promoted, or headhunted. Or burnout. Institutional memory gets lost. Ultimately, it's the average long distance runners who carry the torch to the finishing line.
Not that I dont think the best performers shouldnt be duly recognised and rewarded, but overemphasis on top performance...that's the stuff Enron is made of.
This assumes mediocre workers will always be mediocre. I've always assumed the job of managers is to find the right place for for bright people they've hired. I've certainly seen mediocre talent turn onto stars in the right environment.
This is really the elephant in the room when it comes to any kind of ranking. As humans we always want to make rankings like A is better than B who is better than C... and so on. The reality is much more complex.
If you take K1 kickboxing history, you'll see that pretty much all the top guys have all beaten each other at some point. This makes it pretty clear that "best" is something that varies wildly and may even vary from day to day.
It's no different in software. I can tell you I've had times in my career where I was on fire, out producing any teams I knew about. And I've had times where the burn out was so bad I was producing nothing at all for weeks. How good someone is just isn't static.
This reminds me of an article I read here ages ago on IBM's Black Team. It was (and is) fascinating to me. In case someone hasn't see it, www.t3.org/tangledwebs/07/tw0706.html. Discussed on HN here http://news.ycombinator.com/item?id=985965
As someone who is on the 'high performer' list, I can tell you I have seen this first hand and this is fine, we will work harder, provided we're given an even more outrageous salary. Otherwise, I'm going to get burned out doing the job of my coworkers / thinking for my coworkers every day, monitoring them so they're more efficient and don't go off chasing down bugs incorrectly. I don't need to be managed, leave me alone and I'll make customers and your boss'es boss happy. But this all can get tiresome and it sucks. But give me that bonus/salary and I'll continue on like a trooper.
At my last job, I was given more and more work, and I made everything work without intervention, but I left due to low pay, now the old boss wants to hire me back as the old system seems to go down a lot, while his cheap labor source turns out to be more expensive than he first thought. There seems to be a belief that we can just whip up the high performers when we need them, but this just leads to burn out and turnover. just pay the high performers an ungodly amount and all will be well.
High performers are still a great deal, because you can pay them 2x the salary to get 10x the results. Few places realize this, but the ones that do are wonderful to work for.
High performers are still a great deal, because you can pay them 2x the salary to get 10x the results.
The issue, for many companies, is that the perks the high performers want in exchange for this "arbitrage" that exists in hiring them-- very high levels of autonomy, rapid career advancement, implicit trust from the first day of work-- are considered unaffordable in the actual political context of the company.
Companies won't give high performers 10x salaries because the high performer might not deliver if the conditions are bad, but if they give them high degrees of autonomy and interesting projects, that's visible to the less adept workers and it becomes an issue for management. Managers know that if they let their best people follow different rules, everyone will expect the same freedom.
High and low performance don't seem to be intrinsic to people, in the sense that mediocre people can turn excellent given the right conditions, and vice versa. What does seem to be intrinsic is that there are high- and low-variance people. Low variance people are reliable ladder-climbers, and high variance people are the creative ones who might have a mediocre year or two, but then have a huge breakthrough. The top performers are, for the most part, high variance individuals. High-variance people can also fail quite badly, in certain circumstances, so companies are very nervous about hiring them at all. The "corporate ladder" rewards reliable mediocrity, not intermittent excellence.
Intimidation-based management (which has been dominant since 1800) regresses performance to the mean, and its purpose is to reduce variance. When the input-output curves relating skill and effort to productivity were concave, variance-reduction was the way to go, because lower variance was equivalent to higher performance. Now that we're in a convex world (at least in software) variance-reduction fails us.
That's a bit of a generalization. Some anecdotal evidence... years ago when I was but a youngin', I didn't get along at my company, got burned out and eventually curled up in a corner and slept/played games. Needless to say I got on the PIP (performance improvement plan). Then I got laid off eventually. I was immature. My manager sucked, but so did I for not realizing this and taking action.
These days, I am simply more mature, and I regret my past somewhat, but at least I am good at learning in general. If you don't regret your past, you aren't growing, everyone does. Now, after life has kicked me around, I know people well, I know how not to be a condescending jackass, and be pleasant, and get things done, and communicate to my manager honestly vs being passive aggressive etc. And I have that same inner drive to figure things out. I'm still the same person, I just don't bullshit around, I keep my mouth shut, impress people with my actions, then lo and behold, when they come to find out more info about me, I tell them, but before that, I remain an enigma. When I came to my group, some underestimated me, I didn't lash out like when I was younger, I just shut up and fixed their bugs.
I do think, as you allude to, certain people just have that knowing, they have it in them, but for whatever reason, maturity, bad manager, etc., they don't rise to the top. I think though, if you are mature, you will realize your manager is crappy right away and address that. Part of being a star performer, to me, is not just tech. ability, but non-technical ability.
All these senior management types that praise Welch's philosophies seem disconnected from the real world. They seem to think that by constantly managing out under performers (even mediocre staff) and striving to only employ high performers, everything will be peachy keen.
Sorry but this just doesn't fit the real world because in every company, there will be menial tasks. Menial tasks that high performers will refuse to perform.
Look at sports teams. The Chicago Bulls needed utility players to surround Jordan. Does anyone think a team of 5 MJ's or Lebron's would succeed over the long run?
>>The Chicago Bulls needed utility players to surround Jordan. Does anyone think a team of 5 MJ's or Lebron's would succeed over the long run?
To add further, if Chicago Bulls itself hire all Jordan grade players. Will they want to the menial job of surround the actual Micheal Jordan or try to become Micheal Jordan's themselves?
Go all the way. Most of human value has been created since 1700- medicine, transport, mechanisation etc. Most of that is down to a few thousand individuals, from James Watt to Feymann. If we focus on increasing their productivity (every tenured professor is given state funded housekeepers) we may well increase the invention rate
What we don't increase is the spread of that through society to the benefit of all.
That requires the useless middle managers and useless peons implicitly referred to in this elitist piece of short sighted rubbish
Mexico has an interesting take on this: if you get into the upper echelons of academia you can get a salaried do-what-you-want-publish-if-you-want job. Some of the country's best thinkers are paid to think.
It's different than tenure, they don't have to teach and I don't believe it's associated with a specific institution. I'm not sure exactly what it's called or how it works, I'll have to call a friend or perhaps someone here from Mexico can comment.
Can you please post links about this? It's incredibly interesting, and it would be great to see what's the reason behind it coming into existence and what the output has been because of it.
This is tenure. Most universities have it - and very few have onerous requirements.(#) Getting it is hard (MIT famously has <50% rate - that is after working outrageously hard till your 30's to become world class professor material, you still get 50/50 chance of packing your desk up - eat that "up or out" Management consultancies.
That being said I suspect the tenure rate elsewhere of faile MIT tenures is pretty good.
(#) Newton famously had to teach three classes, they were so impenetrable no-one attended the third. Literally the best minds of the universities could not understand what the hell the best mind was saying. He did not have to teach any more classes.
Kinda fails when you don't have A grade directors running the company. Also, how do you identify an A manager, unless you're really picking the productivity of their A workers? Surely weaker managers attribute their success to the weaker employees (to avoid being 'found out')?
Rather than getting rid of "average people", it seems more common to sequester top talent. I think this is what's happening with small consulting shops and startups. You often see small teams of incredibly talented people at a startup or consulting firm that create a lot of the major innovations, then sell them to the large companies that handle distribution, etc.
This makes sense because most people are average or below average. There is no way that a GE or IBM can function without large swathes of ordinary people. I doubt we'll see average performers disappear from big companies any time soon-instead we'll see the continuing rise of small startups and consulting shops with a few incredibly talented people.
The thing is, you can fill your whole company with go-getting geniuses, but not every task in the company requires that, and if you give the boring-but-necessary work to the whizkids, they're going to get p-ed off and leave.
Stick the smartest on the cutting edge stuff that is the true source of competitive advantage for your firm. Keep the rest for the important-but-dull work - HR, Payroll, etc.
Which one are you? The problem with the magical numbers and hand waving is that there is a 80% chance that the author is one of the mediocre. Would you take this advice from someone who is mediocre?
52 comments
[ 3.6 ms ] story [ 89.7 ms ] threadOf course the world is full of average people. They have to wind up somewhere. And organizations full of average people are never going to be able to take the advice to work the superstars.
I'll believe that the slogan invest in your best has been internalized by our society when we devote serious resources to making sure that people with IQs in the top 1% stop dropping out of school faster than people with median IQs do. Anyone care to give me odds on this happening in the next 20 years? I'll take the "No" side.
The reason why is because jobs that are likely to use the abilities of a high IQ person generally require a high school diploma, and frequently require a college degree. Therefore denying these people an equal opportunity to get those credentials limits how effectively society benefits from their abilities.
Also I should note that the matter is personal. I have a good IQ and yet I came within an inch of failing to complete high school. Were it not for a teacher named Bernie Bowker, I would not have graduated, gone to college, or had any prospects of getting jobs where my abilities would be useful. I think that that would have been a tragedy, particularly for me.
Caltech is perhaps most known for this problem, with graduation rates that were <85%, but changes were made to stop people from "flaming out". Harvard has a 98% 6y rate -- University of Oklahoma is 64%
That is not projection, btw. I played the game pretty well for 3 years. That is just me describing a small portion of the well-understood rules of the game.
Anecdotally the top-performing 20-somethings I know in Japan make at most 2x the average salary, whereas the top-performing 20-somethings in the USA easily make 5-10x average (just look at how much a Ruby Developer, top-tier lawyer, or investment banker makes compared to the average income).
[1]: http://en.wikipedia.org/wiki/List_of_countries_by_income_equ...
The problem is most places don't even have a way to find out a average manager. The reason is that management layers in almost every big company works like a cartel, they know darn well how to preserve their self interests. Generally you don't get to know until its often too late to do anything about it. The people who suffer the most are not bad performers, but those who fall out of managements favor or people who just can't get political and benefit out of it. The actual performers are likely to be named arrogant mavericks who are bad at team work. They won't be fired, but they will be treated so badly, until they themselves see the wisdom in leaving.
Also this whole GE analogy in these kind of articles is MBA speak. That is talking at high abstract levels, without absolutely knowing what they are talking about. Stack ranking worked at GE not because Stack Ranking is awesome, but because there was some one like Jack Welch to make it work.
Adhering to stack ranking to the letter but not in spirit will basically amplify the problem. Guess who gets the least rank in a political system?
http://news.ycombinator.com/item?id=4331584
Here's where the divergence happens. Being a top performer requires a lot, and not all of it's intrinsic. Fit with tools and project are important, and so are relationships. There aren't intrinsic A and B players. For example, most software projects have one or maybe two "blazers" who write most of the new code and make the important decisions, and many more who spend the bulk of their time keeping up. It's a lot better for your career to be in the former category; get out if you're in the latter. This is what language and tooling wars are often about in companies: people wanting to change the landscape so they can be the trailblazers and the others end up as the followers. This is why it's economically worth it (in a corporate context) to invest so much energy into tooling and language fights instead of on the work itself: if you win the battle, the leverage it applies to your work is immense. The truth is that most projects have a rank ordering that becomes self-perpetuating, because the high-ranking people take the best work and progress faster. There isn't one global equilibrium where the "best" people are magically matched to the best work, but N! stable equilibria, where N is the number of participants.
Why is this important? Well, the correct insight is that the payoff curve (between input factors and output) is now convex. In a concave world, you wanted to manage toward the middle: reduce variance, bring up the slackers. This is what management is good at: bringing everyone in to the middle of the performance curve. That's also why traditional management is failing in the technological world. Maximizing a concave function is equivalent to minimizing a globally convex one, which is the easiest class of optimization problems (there's one global optimum). For these, incremental evolution (gradient descent) works: you converge to an optimal solution (in economic terms, an equilibrium). It's not equipped to handle the convex world where there are zillions of possible local optima and non-local insight is required in order to find better solutions.
What's inaccurate is the idea that there are "intrinsically" mediocre vs. excellent people and that the former should just be discarded. Why do we have to stay in school until we're in our mid-20s? Because our economic contributions are worthless until we've had an amount of schooling that would be considered immense by historical standards. The reason for all this schooling is to increase the probability of a person ending up as a "10x" contributor in an economy that has no place for 1x contributors.
As for stack ranking, that's something separate, and it's actually harmful to the top performers. Stack ranking is about intimidation-- nothing else. The idea is ancient: Romans called this method of intimidation and punishment decimation. Split the regiment into groups of 10, of which one will die. Sometimes it was random (drawing lots) and sometimes the groups voted. The reason for the 10% cut was because (a) it's not enough of a force reduction to "disrupt continuity" but (b) it was large enough to scare the shit out of people. People use "decimation" as a synonym for "devastation" but that's not correct, because "decimation" is a controlled, small reduction of force. Stack ranking is the same concept, but with political infighting used instead of randomness to make the decisions, and the "divide-and-conquer" campaign it wages against the proles is intentional.
Intimidation's effect is performance-middling. It kills creativity and innovation, and the top performers suffer, but it brings the slackers closer to the median. Traditional, intimidation-based management worked well in a concave world where people were mostly doing low-margin commodity work and one slacker could cancel out 5 good workers, but it fails catastrophically in the convex world.
At the risk of coming off as a completely arrogant ass, I've been an 'A' player in two completely different language, stack and tooling environments so far. Now I find myself shifting the lion's share of my coding time from back-end to front-end work, but aside from an initial adjustment period, I don't find this to be a huge hit to my productivity or effectiveness.
I'm pretty confident most other 'A' players can do the same easily, probably more easily than I can. Am I unhappy in an environment that forces to use tooling that I find to be unproductive, sure. I've been here before, but nonetheless I was still one of the 'A' players.
A players are not different because of having the right fit with tools or projects, or because of more (or less) education. They are different because they have a good fit with their purpose combined with passion for what they do.
It depends how far apart they are.
Having written the code is a huge advantage. Getting to write code is a major advantage (if most of your time is spent reading code, your odds of transferring to a new-dev role on the same team are low). Picking tools that one knows well is also an advantage, but probably less of one, because there are far crappier corporate codebases than in-use open source tools (in general).
If the people are very far apart in ability, then those kinds of changes won't permute the ranking... on the other hand, if there's that much spread among programmers of equal rank, then you have an unusual team and a recipe for chaos.
Totally disagree. A level programmers are almost by definition, A level programmers in every language they've ever encountered.
For B level programmers, you're right, it's worthwhile to play tug-of-war on languages and tooling.
Being an A programmer won't stop you from writing quality Java code if you get to develop from scratch. The crappiness of the language may hold you back, but not at a career-killing rate, and it won't bring you below expectations because no one expects high productivity when Java is the language of choice. However, if you're dropped into an existing codebase and the code has a bunch of VisitorFactory garbage, being an A programmer hurts you. Good programmers just don't have much patience (or experience, if they can help it) with shit code.
Teams with shit code and poor unproductive environments, pretty much by definition, don't have 'A' players and the ones they do have don't stay long. Someone once told me "it's easy to retain mediocrity in a company, it's the talented people that are hard to keep." This is a chicken-and-egg problem, but in my mind it is the lack of quality team members and managers that leads to a poor environment all around.
That being said I've worked in environments like this, where tooling was poor, the build chain almost non-existant and there was a legacy of extremely poor architecture. I was still able to make improvements, bring in better code, architectures and dev practices (albeit slowly). For that matter, so was every other 'A' player I know that worked there. The main thing is none of us stayed there more than a year.
>> There's arguably never been a worse time to be a mediocre, average or typical employee. For most firms today, mediocrity is a cost to be managed and a burden to be borne, not a potential to cultivate or a resource worthy of serious investment. Indeed, if an app, web service or an outsourcer can deliver 80% of typical or average performance at 35% to 60% of the cost, then why pay for average employees? These days, no organization can afford to pay a premium for mediocrity.<<
The Indian outsourcing and bodyshopping cartels which thrive on mediocrity and cost-arbitrage would actually consider that to be awesome news, for it hints that they will get more work to keep costs low. But what they fail to realize is that another cheaper destination is emerging elsewhere. :) One man's poison is another's pie.
I'm a fan of this line of thought: http://headrush.typepad.com/creating_passionate_users/2006/0...
So many people are focused on bringing everyone up to some base level of mediocrity in every imaginable area, but what if we instead focused on finding brilliance, then doing everything we can to amplify that brilliance?
Which might, of course, include working on some of those areas for improvement, but intelligently, with the goal of keeping those weaknesses from impeding excellence elsewhere.
At the end he asks about the organizations where 60% of value is produced by 40% of employees. Which organizations? How is value measured? No telling because the numbers are just made up.
And let's add the obvious remark that if this 40% is that good, than why it is not firing the other 60%? Why do they keep them?
All in all, this is the kind of poor business talk that makes MBAs so irritating.
At risk of sounding like I am defending my own mediocrity, the average worker and manager are going to be the ones who can last the distance.
Top performers have a nasty habit of getting promoted, or headhunted. Or burnout. Institutional memory gets lost. Ultimately, it's the average long distance runners who carry the torch to the finishing line.
Not that I dont think the best performers shouldnt be duly recognised and rewarded, but overemphasis on top performance...that's the stuff Enron is made of.
If you take K1 kickboxing history, you'll see that pretty much all the top guys have all beaten each other at some point. This makes it pretty clear that "best" is something that varies wildly and may even vary from day to day.
It's no different in software. I can tell you I've had times in my career where I was on fire, out producing any teams I knew about. And I've had times where the burn out was so bad I was producing nothing at all for weeks. How good someone is just isn't static.
There is a sense of eliteness on a good team. Team members feel they’re part of something unique.
[1]- http://www.amazon.com/Peopleware-Productive-Projects-Teams-e...
At my last job, I was given more and more work, and I made everything work without intervention, but I left due to low pay, now the old boss wants to hire me back as the old system seems to go down a lot, while his cheap labor source turns out to be more expensive than he first thought. There seems to be a belief that we can just whip up the high performers when we need them, but this just leads to burn out and turnover. just pay the high performers an ungodly amount and all will be well.
The issue, for many companies, is that the perks the high performers want in exchange for this "arbitrage" that exists in hiring them-- very high levels of autonomy, rapid career advancement, implicit trust from the first day of work-- are considered unaffordable in the actual political context of the company.
Companies won't give high performers 10x salaries because the high performer might not deliver if the conditions are bad, but if they give them high degrees of autonomy and interesting projects, that's visible to the less adept workers and it becomes an issue for management. Managers know that if they let their best people follow different rules, everyone will expect the same freedom.
High and low performance don't seem to be intrinsic to people, in the sense that mediocre people can turn excellent given the right conditions, and vice versa. What does seem to be intrinsic is that there are high- and low-variance people. Low variance people are reliable ladder-climbers, and high variance people are the creative ones who might have a mediocre year or two, but then have a huge breakthrough. The top performers are, for the most part, high variance individuals. High-variance people can also fail quite badly, in certain circumstances, so companies are very nervous about hiring them at all. The "corporate ladder" rewards reliable mediocrity, not intermittent excellence.
Intimidation-based management (which has been dominant since 1800) regresses performance to the mean, and its purpose is to reduce variance. When the input-output curves relating skill and effort to productivity were concave, variance-reduction was the way to go, because lower variance was equivalent to higher performance. Now that we're in a convex world (at least in software) variance-reduction fails us.
These days, I am simply more mature, and I regret my past somewhat, but at least I am good at learning in general. If you don't regret your past, you aren't growing, everyone does. Now, after life has kicked me around, I know people well, I know how not to be a condescending jackass, and be pleasant, and get things done, and communicate to my manager honestly vs being passive aggressive etc. And I have that same inner drive to figure things out. I'm still the same person, I just don't bullshit around, I keep my mouth shut, impress people with my actions, then lo and behold, when they come to find out more info about me, I tell them, but before that, I remain an enigma. When I came to my group, some underestimated me, I didn't lash out like when I was younger, I just shut up and fixed their bugs.
I do think, as you allude to, certain people just have that knowing, they have it in them, but for whatever reason, maturity, bad manager, etc., they don't rise to the top. I think though, if you are mature, you will realize your manager is crappy right away and address that. Part of being a star performer, to me, is not just tech. ability, but non-technical ability.
Sorry but this just doesn't fit the real world because in every company, there will be menial tasks. Menial tasks that high performers will refuse to perform.
Look at sports teams. The Chicago Bulls needed utility players to surround Jordan. Does anyone think a team of 5 MJ's or Lebron's would succeed over the long run?
To add further, if Chicago Bulls itself hire all Jordan grade players. Will they want to the menial job of surround the actual Micheal Jordan or try to become Micheal Jordan's themselves?
Go all the way. Most of human value has been created since 1700- medicine, transport, mechanisation etc. Most of that is down to a few thousand individuals, from James Watt to Feymann. If we focus on increasing their productivity (every tenured professor is given state funded housekeepers) we may well increase the invention rate
What we don't increase is the spread of that through society to the benefit of all.
That requires the useless middle managers and useless peons implicitly referred to in this elitist piece of short sighted rubbish
Ahh... tenure... how we miss ya.
http://en.wikipedia.org/wiki/Tenure_(academic)
That being said I suspect the tenure rate elsewhere of faile MIT tenures is pretty good.
(#) Newton famously had to teach three classes, they were so impenetrable no-one attended the third. Literally the best minds of the universities could not understand what the hell the best mind was saying. He did not have to teach any more classes.
This makes sense because most people are average or below average. There is no way that a GE or IBM can function without large swathes of ordinary people. I doubt we'll see average performers disappear from big companies any time soon-instead we'll see the continuing rise of small startups and consulting shops with a few incredibly talented people.
Stick the smartest on the cutting edge stuff that is the true source of competitive advantage for your firm. Keep the rest for the important-but-dull work - HR, Payroll, etc.