July 7, 2012

"Stack ranking" — the management technique that ruined Microsoft.

Stack ranking "forces every unit to declare a certain percentage of employees as top performers, good performers, average, and poor."
"Every current and former Microsoft employee I interviewed—every one—cited stack ranking as the most destructive process inside of Microsoft, something that drove out untold numbers of employees,” Eichenwald writes. “If you were on a team of 10 people, you walked in the first day knowing that, no matter how good everyone was, 2 people were going to get a great review, 7 were going to get mediocre reviews, and 1 was going to get a terrible review,” says a former software developer. “It leads to employees focusing on competing with each other rather than competing with other companies.”
ADDED: Stack ranking seems designed to overcome the standard problem in group projects, that people take advantage of each other. If we're all going to get the same credit, what do you do? Work really hard or let others do the work? What can you do to prevent that dysfunction? Apparently, the answer is to create a different dysfunction.


chickelit said...

This link and story makes me wonder how such companies run things nowadays in outsourced job colonies.

Do they adhere to Maoist egalitarian principles?

edutcher said...

Software development, to a great degree, requires teamwork.

This would be a real killer.

Andrew Koenig said...

Suppose you're a manager with 10 people working for you, and you're told that next year, you're going to have 5% more money to spend on salaries as you had this year. What method of allocating that money do you think would be most beneficial to your organization?

Phil 3:14 said...

My former company did this. Its big downside is not rewarding the ENTIRE high performing team. Its big upside; its easy.

rhhardin said...

I'e been bottom performer out of hundreds of people, causing the previous bottom performer to express envy.

We did not seem to be the dumb ones.

Joe Schmoe said...

Seems to me the stack ranking has less to do with Microsoft's performance than them just being overly reliant on Windows and Office for 20 years. They've hit a problem every company has; making or doing cool stuff is fun, but if you're not profitable, you're not going to survive.

I was a manager at a large company that used stack ranking. I didn't mind it. As horrible as the writer makes it sound, the one good thing it does do is force managers to take a hard look at the really poor performers. Every group has them, and group members are always whispering about how so-and-so is not pulling their weight, when is the manager going to do something about it, etc. Stack ranking makes reticent managers confront poor performance.

Just look at the federal govt. employees. Just about zero turnover, and I read somewhere that you have a better chance of dying than getting fired from a govt. job. Don't you think they could benefit from stack ranking, and clear out some of the accumulated rot? Cuz I can assure you there's a bit of rot among the federal employees.

leslyn said...
This comment has been removed by the author.
leslyn said...

Microsoft is "destroyed?" When did this happen?

NotquiteunBuckley said...

On a track, a runner running alone, will never drown in his or her own sweat.

Rialby said...

Dell was the exact same way.

rick said...

Microsoft using the antiquated bell shaped curve? Who knew?

Steven Jobs used a more modern technique - he would fire people who had the misfortune of riding in the same elevator with him.

Shanna said...

In practice, this tends to mean the couple favorite/suck up/teachers pet employees generally get the good ratings and everybody else just gets 'fine'.

If you do a good job hiring and developing your employees, they might all deserve a great ranking. Somebody else's group might all deserve a terrible ranking.

I hate the 'ribbon' and the new explorer. I hope those people got a bad ranking.

traditionalguy said...

Survivor Island sounds like a fun game. Dominance uber alles.

Obamacare will become our National Survivor's Island game in which the losers will disappear.

But an Islands where the team was under an enlightened leadership that got the best from each tribal member beat the Survivor Island every time.

Silly Progressives never can seem to accept that the human talent pool comes in too many sizes, and pesronality types to start eliminating those not meeting an ideal, which stacking judgements are designed to do. Stacking is actually as dumb as rocks.

chickelit said...

Stack ranking seems designed to overcome the standard problem in group projects, that people take advantage of each other.

cf. "The Little Red Hen"

David said...

Stack ranking is also attractive to companies which are overstaffed or which are lax in their hiring standards. As such, it masks corporate problems as well as causing them.

Larry J said...

The Air Force has used this evaluation technique from time to time. Where it really fails is when you have select units of highly qualified personnel. Back in the 1980s, they applied this to aircrews. When it came to SR-71 Blackbird crews, you ended up giving some exceptional people poor rankings because of an arbitrary quota. Those crews basically had to meet astronaut selection criteria. There were no slackers or poor performers there. The same is true of any truely elite group of people. Arbitrary quotas can be applicable for a large group of ordinary people but if you have high performance teams, it shouldn't be used. A "poor performer" there might be a star in an ordinary group of employees.

Firehand said...

One guys' experience:

DADvocate said...

Corporate America is really screwed up. My company does a lot of work for large corporations, usually Fortune 500. We only deal with marketing, but some of the stuff they come up with astounds me. A few companies seem to have it all together, but most stumble from success to success, or success to failure and back again, in spite of themselves.

Mr. D said...

Lotsa companies use this technique and I've been at the top and the bottom. Being at the bottom was more of a blessing than being at the top. Of course, the motto at the Fortune 25 company at which I worked was "the day you're hired, the bullet is fired."

Quayle said...

Stack ranking seems designed to overcome the standard problem in group projects, that people take advantage of each other. If we're all going to get the same credit, what do you do? Work really hard or let others do the work? What can you do to prevent that dysfunction?

Not if the group project is properly run, as follows:

1. There must be only one project manager (or overall project manager); one butt to kick; one person responsible for getting it done regardless of anyone else.

I have a saying, or adopted a saying, "If two people are in charge, nobody is in charge."

2. There must be a clearly defined end result. It can't be mere language (i.e. "Build a house"). It must be clearly specified (i.e. "Build THIS house")

3. The sub-components - the break out into sub-tasks - must be clear and must be realistic.

4. The assumptions on which the project resides and the outside dependencies (dependencies on things outside the project) must be clearly understood and constantly reviewed.

5. The internal dependencies must be understood and constantly monitored.

6. Each team member must be given clear ownership of a sub-task or deliverable. This is how you make them responsible for delivery. They can't blame others. They must meet their deliverable and their date or show the team whose dependent deliverable was late or wrong.

(BTW, it is a certainty that Obama doesn't know the first thing about any of the above. He is an absolute novice.)

Bill Harshaw said...

This was tried in the Federal government for a while--part of the Carter reforms of civil service. It's one way to attack the Lake Woebegon effect, where everyone is above average, which often results when supervisors take the easy way out.

It was abandoned in my agency under Reagan for some of the reasons in the article and comments, plus, I suspect, the fact it was a Carter initiative.

Mark said...

An anonymous Microsoft employee runs a blog called Mini Microsoft, and the yearly posts about performance reviews have hundreds of comments from other employees about the stack ranking process and outcomes. A major problem, in addition to the ones mentioned above, is that Microsoft managers often rank based on the "visibility" of the employee to others outside the team, esp. to higher mgt., regardless of competence.

Methadras said...

One way to solve the problem is to eliminate hierarchy altogether and simply have a flat support structure. That way people contribute equally without having to worry about stack ranking since you'll see right away who belongs there and who doesn't. It takes care of itself while people focus on their competition which isn't their own work associates.

Ann Althouse said...

"Microsoft is "destroyed?" When did this happen?"

Who are you quoting?

Hagar said...

I have only been subjected to one "performance review." I think it became understood that I did not care for it, so they did not try it on me again.

My basic attitude is that if my boss/supervisor have a comment about anything I do, he/she should tell me about it.

Shanna said...

My basic attitude is that if my boss/supervisor have a comment about anything I do, he/she should tell me about it.

That's kind of the thesis of the 'one minute manager'...just give feedback when it actually matters, ie, when something good or bad has happened. Annual performance reviews are pretty ridiculous, imo, and say more about your supervisor/management rules than they say about you.

PatCA said...

It sounds very bureaucratic, like Bill Gates ran out of ideas and didn't trust the creativity of his underlings to keep the business growing. Like public employment or academia.

When the pie is not growing, people act quite viciously to keep a piece of it.

Shanna mentioned the hated ribbon. Office ruined its brand in a quest for quick profits. Didn't work.

bbkingfish said...

Are there any major corporations that DON'T use some form of stack ranking (as it is explained here)?

Hagar said...

All these "management techniques," like TQM ("Total Quality Management") and its various spinoffs and imitators, are fundamentally destructive to employee morale and the profitabilty of any firm that gets ensnared in this happy horseshit (or "Bullshit").

Hagar said...

That said, who knew that Microsoft has been "destroyed"?

(I am with Leslyn on that one; perhaps the one and only thing ever :))

traditionalguy said...

An old saying about people was that they seemed worthless because they were the second best basketball players...on Michael Jordan's team.

bagoh20 said...

Motivating people is a fascinating problem. At my small company which has ranged from 50 - 200 people and currently has around 100, I have tried a number different approaches. In order to prevent creating extra work I usually try to keep it simple. I attempt to share nearly all the discretionary profit with the employees. For a while we just took the profit and spread it around as raises based on merit. It only took a few highly profitable years to leave everyone with high fixed salaries that nearly killed us when the economy went south. It forced me to cut pay and just terminate a lot of people during a bad year. Nobody wants that.

My current strategy is to only raise salaries for changes in level of responsibility or demonstrable skill. We distribute profits through bonuses.

At first everyone got the same bonus, which made people happy and improved general performance, but led to some employees trying less, and others getting to resent them.

It's too hard to find and train new people, so just replacing the less motivated is not adequate. Besides, I want to teach people the advantages of being more productive, whether they stay with us or not.

So, now I have created a very simple 3 point spreadsheet evaluation process that my managers can use to evaluate everyone's performance in a few hours. Then bonus is distributed according to their individual scores.

The amount of bonus available is determined by overall company performance which encourages team effort, and one of the 3 points of evaluation is cooperation and teamwork. What portion each person gets of the bonus is dependent on individual performance. Therefore the very best money comes from performing well as individuals and as part of the team. Everybody has the same potential bonus amount whether top executive or brand new employee.

This seems to be working the best. people are appreciative and actively pursue help in improving themselves before next quarter.

Overall, the best people are not affected by motivating strategies. They simply can't be any other way. They do it because it's who they are, but it does lead to loyalty with them. It's everyone else that needs something more.

bagoh20 said...

"All these "management techniques," like TQM ("Total Quality Management") and its various spinoffs and imitators, are fundamentally destructive to employee morale and the profitabilty of any firm..."

I've read that is statistically true, and in my experience, the appearance of such a system means the company is headed for some bad years. Every customer I ever had who implemented one of these programs either dropped it, scaled it back to near irrelevance or went out of business.

I'm not sure if it's a company killer or just a signal of preexisting management confusion and lack of better ideas.

There is a lot of money in selling this stuff. I see it to be just like alternative medicine for business.

Tari said...

Thank you, Jack Welch - people read books like his when these lousy ideas are touted and then, bang!, they decide this technique will make them a better manager. I worked with a guy who used to read books like that: when he finally left his direct reports threw a party.

Sklar Hast said...

When the days of easy growth were over with for my employer they turned their attention to producing growth through efficiency. Some consequences were regrettable. Chiefly, they began to evaluate performance competitively.

In my job we worked alone 95% of the time and under conditions that were broadly similar but varied widely in circumstantial detail. I saw everyone's circumstances since my job was to cover anyone on leave. My performance varied almost exclusively according to the circumstances I encountered from day to day.

Nothing in the calculations used to evaluate performance allowed for the ten thousand minor details that changed at random daily. It became commonplace wisdom that you would be evaluated by a do-nothing yes man who had no incentive to distinguish between reality and the image on his computer monitor. A workplace characterized by personal regard between management and employees changed very much for the worse.

Management knew well the advantage to them in the personal loyalty their oldest employees had given them in the start up years. Much company propaganda was aimed at reassuring us that we were the same, good old company we had always been when it came to personal loyalty. You could enjoy a hamburger and a hearty congratulatory speech when some goal was met (in fact, you had better) but all the loyalty they boasted about was gone.

This situation would not have caused the cynical hardening of attitudes that is the inevitable result when one set of people claims to have real, personal regard for another set who can see that it's really all bullshit. People don't regard an all business attitude as dishonest as long as you don't pretend that it's anything else. Management wanted the benefit of the old loyalties but none of the cost.

All of them were evaluated the same way. Simple calculation. Strictly numbers. They smiled and back slapped and asked for input and didn't listen and wondered, some of them, why they were casually despised as ass kissers. Gradually all of them, and us as well, who could remember a more human work environment were replaced.

Our replacements were kids who grew up taking for granted that the kind of analysis that the start up generation was reluctant to undertake because it was so labor intensive and costly could nowadays be completed almost as soon as you could conceive some half-assed new angle to chop numbers with and have the results displayed on your monitor in no more time than you needed to take a leak. Their faith in the fidelity of their numbers to reality was as remarkable as it was callow. Making one of their indices jump a tenth of a point and another one go down by as much makes them a star.

Can't make progress with one analytic approach? Devise a new one - a phenomenon we called flavor of the week. A person in an office remote from any actual work attempts to make innumerable small decisions for a large crew who are responding to thousands of details about which he can never have knowledge. This, I think, is the curse of the modern workplace and particularly a large corporation.

I blame information technology. I do not fear computers. Like Isaac Asimov, I fear the lack of computers. Without them things would be much changed for the worse. But people are too much in thrall to data and act as though their measurements of reality are reality. A manager in the company I worked for is nothing more these days than the human face of a computer program.

I never thought, in the early days, that I would be relieved to reach retirement. I'm so relieved to escape the scrutiny of the dumb formulas that are the mind of corporate culture that it's as if I have been cured of some loathsome disease. I'm looking now for a second career and you can bet I'll never become the soulless bastard whose job it is to read from a report and smile and assure everyone that he's human. Just as human as you.

Joe Schmoe said...

A few thoughts:

Big companies do stuff like this because the status quo isn't working. It's okay to tout alternate management strategies, but 'let everyone do their own thing' isn't an option.

Big companies are geared to maximize efficiency when it comes to their products or service. In order to maximize efficiency, they have to come up with a lot of processes or procedures. Big companies are terrible at innovation. See all the tech companies like Google and Microsoft that hit the jackpot once but haven't innovated anything of significance since then. Big companies can't innovate well; that's why they buy small companies that can.

Stack ranking can fail in a perfect world when you've weeded out all the poor performers, and you're left with reasonably good to great performers. But, in the five years I worked under it, that never happened. I had seven different teams in that time. Turnover and role changes ensures this almost never happens. If it does, great. A desired end result of stack ranking is to weed out poor performers. Managers do spend most of their energy dealing directly or indirectly with poor performance. If you don't have to deal with that, maybe you can focus your team more on productive activities rather than deal with destructive fallout.

Just saying 'make a good hire' isn't realistic. I don't care who you are; you're not going to bat 1.000 when it comes to hiring. You're just not. At least stack ranking gives you a way to deal with it.

Just saying 'properly manage the project or team' isn't realistic either. Please point me to the handbook that has all the answers for how to run the perfect project. There isn't one. Colleges aren't great at prepping people to be managers because they adopt a big-company mindset governed by economics models that are so inaccurate as to be useless. People need to be better at thinking on their feet. That's why someone like Bagoh is successful; he hasn't been brainwashed by outmoded MBA techniques for how to run a company into the ground. That, and it sounds like he doesn't have public shareholders to answer to ;)

Curious George said...

"Tari said...
Thank you, Jack Welch - people read books like his when these lousy ideas are touted and then, bang!, they decide this technique will make them a better manager. I worked with a guy who used to read books like that: when he finally left his direct reports threw a party."

I did too. I remember a marketing consultant that we hired commenting "Someone should take away his library card"

Chris said...

Assessing employees isn't hard. Microsoft is a very bureaucratic company that wanted the process overregulated. If you want to find 10% of people to fire, it's easy - you say "who do you want to fire?" and look at who gets the votes.

Joe Schmoe said...

In practice, this tends to mean the couple favorite/suck up/teachers pet employees generally get the good ratings and everybody else just gets 'fine'.

In my experience, not so much. At the beginning of the year, we worked with our reports and teams to set up a list of things they would accomplish throughout the year. I explained to them what was good, and then we added stuff that could get them to a higher level. If things changed during the year, such as switching to a different project, we would revisit the goals accordingly.

So results against goals, all of which was very measurable by the amount and quality of work output, was the biggest factor in the ranking. At least where I worked.

Let me add one thing. I'm not a dyed-in-the-wool advocate of stack ranking, but I am familiar with the challenges of managing a group in a large corporation. It's hard and challenging. A lot of people who have never managed people think managers are clueless. Some are, but most know more of what's going than you think. By far. But to reiterate, it's hard to make changes at big companies. They have a ripple effect throughout your whole business unit, so it takes a lot of work to justify changes. Making changes at big companies has been compared to turning an aircraft carrier around in a small harbor. In my experience, that's apt. It's one of the reasons why I now work for a teeny tiny company.

jeff said...

My company does that. You are ranked against the people in your group. The group is allotted so many high average and lows. So the same guy could be ranked as high, average or low doing the exact same job depending entirely on what group he is in. It's entirely possible for that guy to get a strong review, nice raise and a nice bonus OR a poor review and urged to look for employment elsewhere for doing the exact same job. And everyone knows it.

bagoh20 said...

"That's why someone like Bagoh is successful; he hasn't been brainwashed by outmoded MBA techniques for how to run a company into the ground. That, and it sounds like he doesn't have public shareholders to answer to.;)"

Oh, I went through my phases too. When you don't have experience, you assume other people must know better, so you read and trust ideas put out there. Some are good, but a lot are crap, and you can't tell the difference. I have managed to grow companies for almost 3 decades with only 2 unprofitable quarters in all that time, but I believe a lot of that was luck. I couldn't write a book on how to do it, because although it's hard, it's simple:

Don't over-think it, be honest, and treat people well. You need to give your job away, and then do it again and again. Management for me is really just the golden rule - it works everywhere. What would you want your boss to do with you? Do that for them. I don't know if you can build a muti-national corporation that way, but it sure works on the small scale.

And no, I have no shareholders. I can be as good to people as they deserve, which is a dream come true. Nobody enjoys payday as much as me.

Robin said...

Microsoft is like an immense dinosaur, its very powerful. But when it dies, it takes time for the signal to reach the tiny brain. But when that signal finally reaches the tiny brain, the whole dinosaur falls over more suddenly than expected.

New growth in personal electronics is in fields where Microsoft has no monopoly and the result is that they can't shove mediocre products down the consumers throat.

It also means that Microsoft's inefficiencies can't be covered up. Microsoft can spend huge amounts of money but it does not guarantee profitability. An example is the Xbox division, which has still not earned enough profit to pay back the immense amounts of money dumped in its development.

Alex said...

I love all these "geniuses" at Forbes and Vanity Fair. Let them try to run a company with 95,000 permanent employees and 2x as many contractors at any given moment. Microsoft is a massive enterprise and keeping it profitable is not easy.

Alex said...

Robin - yet the Kinect is the fastest selling gadget in history. Windows 7 has sold 600 million copies in less then 3 years. Now Microsoft is coming up with a brand new tablet - The Surface with a completely new manufacturing method and new keyboards. To say they are a dinosaur is the height of ignorance. I bet you don't even know what Azure is. I bet you haven't seen how many design awards Windows Phone has won.

DADvocate said...

I bet you haven't seen how many design awards Windows Phone has won.

Awards don't pay the bills, or employees.

rcocean said...

The "stack method" didn't work at MS because you're dealing with 10 very smart, very motivated people. As Althouse points out the problem with business group projects is the slackers take advantage of the workhorses, who then get discouraged.

However, if you have 10 workhorses on a project, "stacking" causes the problems indicated.

tiger said...

Management in certain divisions of GE (under Welch, IIRC) set up reviews so that 10% of the employees were fired every year. I worked for a company where the VP/Sales said that he thought 30% of the sales force should be fired every year. EVERY YEAR!!!
Nice to see how GE and my old company - sold a few years later - have fared.
'Not very well' is the answer.

bagoh20 said...

This idea seems doomed to failure. Who gets in the line that says "we cull 10% of those in this line regardless of how good they are"?

I always thought that people perform not only best, but exceptionally well when they are having fun with it, and when the environment is safe enough that they only have to think about the challenge at hand, free of politics and posturing. I eliminate anything that encourages that stuff.

I defend any employee attacked, and my people know it. I do what has to be done, but a person knows they deserve it when they lose me in their corner. I would say 90% of the time I fire someone, they admit they knew it was coming and they accept that it's their fault. The last person I fired, just a couple weeks ago, after many years together; hugged me, apologized, and said thanks for the opportunity.

That may seem like unrealistic softness, and I do get taken advantage of occasionally, but I bet a lot less than most companies.

Kirk Parker said...

Joe Schmoe,

Did they stack rank the managers with the same vigor as they did the line employees? If not, then there's your answer.

Oh--and also what Jeff said@ 1:05pm.

Kirk Parker said...

As far as the book-reading managers go, things actually weren't that bad in the business/management book arena until the appearance of In Search of the One-Minute Megatrend.

Gabriel Hanna said...

Stack ranking might be fair if you were evaluated over multiple projects.

But I think it has the same problem as curve grading in academia. At high-ranked schools students don't work together. At low-ranked schools, students conspire to keep the average as low as possible.

Shanna said...

Don't over-think it, be honest, and treat people well.

I like this. I think people will put up with a lot at work if they are treated well and are able to respect the manager.

Just saying 'make a good hire' isn't realistic.

But if you have managed to make good hires and train them properly so you have a very good, high performing team, this way of assessing performance doesn't work. Some of this depends on how big or small your group is, of course.

I know they talked a little bit about switching over to this kind of ranking last year but backed off. In practice, it would have meant someone who was not your supervisor would be making decisions on how good you are.

And as mentioned, if you employees and they are all different levels (ie, some managers, some clerical, some mid level) and someone has to have a lower ranking, it tends to be the lower paid employees who get that. And that kind of sucks, because it has nothing to do with performance and everything to do with position.

Gabriel Hanna said...

I can see the logic: fire the lowest performing 10%, and hire new people; some of them are bound to be better than the 90% you kept. Eventually all your workforce is high performers.

Problem is people figure that out. New employees are competition for your job. Shut them out, don;t teach them anything. Make no use of their talents, make them unhappy, let them get fired.

High turnover and a hostile environment follow. Sounds productive.

ed said...

As someone who has participated in more software development teams than I care to remember the basic rule is this:

If the team leader or project manager doesn't have a very good grasp on who are the stars, the worker bees and utter dregs then that team leader or project manager needs to be fired ASAP and escorted out of the office right then and there.

When I've lead teams there was never any question in my mind as to who performed and who did not. To have such an idiotic system in place frankly makes me think upper management in Microsoft needs to be cleaned out wholesale.

PS: And Microsoft; STOP changing the damn menus in MS Office! Every new version I've got to waste time finding stuff. It annoys the shit out of me. I've got better things to do than to waste my time goofing around with your shit.

Michael Haz said...

The Microsoft employees who don't like Microsoft's management techniques are free to seek employment that better suits them, aren't they? They aren't being held hostage, locked into their cubicles every day. They can leave.

But most don't. That's because there aren't other employers who pay as well as Microsoft for many software and CE jobs.

Also, employees love to bitch and moan about the job. I did, you did, everyone does. War stories, often overblown.

ed said...

@ Gabriel Hanna

"I can see the logic: fire the lowest performing 10%, and hire new people; some of them are bound to be better than the 90% you kept. Eventually all your workforce is high performers."

A couple things:

1. There is an 80/20 rule. 80% of the productive work is done by 20% of your developers. You carry the other 80% to do the stuff that you don't want the 20% to waste their time on and in the hopes of developing one of them into a 20%'er.

2. You are very very rarely ever going to have an entire team of 20%'ers. Some problems are simply personality wise. Other problems involve insufficient challenge to keep their interest. Some are extended immaturity. It happens.

3. Just having one superstar on a team can make the team. Lacking at least one will probably sink your project.

4. And no specific metric really works in identifying the superstar. They just show up and you can tell right away. You can have a dozen developers beavering away at a project and this one developer can do 100 lines of code that makes everything work better as if by magic.

ed said...

@ DaDvocate

"Corporate America is really screwed up. My company does a lot of work for large corporations, usually Fortune 500. We only deal with marketing, but some of the stuff they come up with astounds me. A few companies seem to have it all together, but most stumble from success to success, or success to failure and back again, in spite of themselves."

It really is amazing isn't it?

I was part of a $70 million refactoring of a major software suite for a large corporation. After the money was spent, the software done ... the company got sold and the new owners junked the whole thing because they wanted to use their own in-house system. *shrug* I got paid so ymmv.

But it is stuff like this that keeps me going as a contractor. If someone thinks, for whatever reason, that I'm not pulling my weight then they have the option to just tell me the contract is done. No muss, no fuss. And I off and working for someone else.

My first day I always tell clients that if they don't think that they're getting a terrific value for their money every single day then they should not keep me.

Gabriel Hanna said...

@ed:There is an 80/20 rule. 80% of the productive work is done by 20% of your developers. You carry the other 80% to do the stuff that you don't want the 20% to waste their time on and in the hopes of developing one of them into a 20%'er.

That's a good point: the lowest 10% might not be significantly lower than the next lowest 50%. In which case the strategy is basically firing 10% at random, hoping to get better people.

Ameryx said...

This approach to evaluating employees has lots of perverse effects:
1. As someone else noted, it decreases cooperation within a team. Why help someone else look good?
2. It decreases incentives to move into more challenging parts of the business. Why not stay a big fish in a small pond?
3. Employees are discouraged from taking risks: failure could mean dismissal.
4. The horizon for employees maxes out at one year. Some projects may produce nothing for an extended period before yielding results.
5. It makes middle and upper management lazy. They can be less involved in evaluations, hiding behind the statistics.
6. It is discouraging for first-line managers, as it reduces their discretion and their ability to treat employees fairly.
Organizations seem to adopt this approach when management lacks the integrity and courage to manage .

ed said...

@ Ameryx

"This approach to evaluating employees has lots of perverse effects:"

For me the most perverse effect of something like this is to drive top workers out of some teams or projects and into others. If you have 10 people and this management technique is in effect. Then if that team has 3+ top workers then at least one of them will transfer to another team or project.

Or even more perverse all of the top workers will apply to transfer to another team or project. Or simply get a new job with a new organization.

ed said...

btw one interesting way to encourage software developers is to give them 1/3'rd of the hours they save on development back to do whatever they want.

As an example if someone is allocated a total of 30 days to complete his portion of a project and he is done in say 23 then take 1/3'rd of that (about 16 hours or 2 days) and let the guy do whatever he wants. Play with some software. Learn something. Write a prototype with some new tech.

Nerds gotta be nerds.

Joe Schmoe said...

Kirk Parker, yes, the managers were stack ranked with more vigor than the individual contributing employees. Two years I wound up paying higher bonuses to my top people than I received from my boss. The debates over team rankings and manager rankings were very, very intense. You had to bring your A-game to advocate for you and yor team, and to get the most bonus pool for your team. That said, at the managerial level, we were all realistic about the pecking order we were in. The developer teams were making the intellectual property, thus in a good year they were going to get the lions' share of the bonus. If you were on a support team, you weren't going to get as much money. If you didn't like that, tough. And my teams weren't the developer teams.

And hey, if you know who is working well and who isn't, then this is an easy system to administer.

Some of you have proferred hypotheticals where deserving people get overlooked or fired, but you're going to extremes. It's not like that. It's something companies with multiple levels of management can use to implement a consistent process among their management teams. Just expecting someone to know how to manage is unrealistic. Managers need tools and training, too.

Shanna, everybody tries to make good hires. My point is that if you hire ten people, one or two of them aren't going to work out long term. Good hiring practices and proper training don't guarantee that it's always going to work out. If you've ever hired a decent number of people, you'd know that. Saying that there's some magic formula to ensure good performance is something that only exists in the minds of academicians.

Dante said...

Microsoft is having trouble because of the paradigm shifts. Yes, IBM is around, but how many people are writing in COBOL these days?

The essential problem is the technology moves fast, people invest a lot of time aligning their neurons to perform with a paradigm, and when it changes it's hard to adapt. For instance:

IBM, the LAN, and PCs.
Novell and the internet.
Sun and Linux.
Kodak and the digital camera.
Newspapers and blogs (having lost their monopoly distribution channels, oops)

Apple nearly died at the hand of the 80386 PC, but was reinvented by a magical man.

Microsoft continued to try to push their crappy Windows software onto phones and other embedded devices (I think they called it WINCE, for windows CE, like get a clue). But, other more nimbler folks have come in and snapped up that and tablet computing. You can't lug the old ball and chain forward sometimes, you have to start all over.

The change and destruction of companies is why productivity continues to increase. Now, compare that to government.

Uh. I don't even want to think about how government monopoly control over the health care system is going to slow it down forever. Long live Solyndra.

Dante said...


You are obviously working for (or owner of) a company focused on success and survival. That's not the way very large companies work, in my experience. I would guess it would be easy to get rid of thirty percent of the folks, and no one would notice the difference. Sure, processes would have to be streamlined.

I left such a big company I worked for for seven years. I don't know how I put up with it. Now I'm at a much smaller company. Sure, there are dicks walking around, but not as many of them, because the company has to be successful. It's not like you can screw up a product line and it won't hurt everyone. Not so at the big companies. They will use their marketing and sales team to cover up the screw ups, so it ends up being political.

And then there are all the process hounds, who think the right answer to screw ups is to add more process, so nothing can get done.

This is a true story. An onerous time consuming process was added in by some senior director guy that affected all the worker bees. This guy was heading up the India team, and was pushing it here in the US too. I spent my weekend, probably about sixteen hours of time, trying to get through this. Finally, I complained up to the VP. I went through and found that this guys team weren't doing it at all, of course, so I recommended a way of making the whole thing automatic. I met with the India Director, and he told me "It's not possible my team isn't doing this process because I told them too. That's the kind of person you get in a big company. Adding worthless time consuming process, then demanding it on high so he can brag to his superiors about how he is improving things, and adding value. The right thing happened this time, and the guy was let go, thank goodness.

Harold said...

Any rating system that requires a subjective rating is inherently flawed.

My part time job is in sales. There are metrics that can be measured. Sales per hour, attach rate for service plans and accessories, merchandise return rates, etc. We all have access to total sales numbers for everyone.

In the store I'm working at now, my department head didn't like me. My evals were done by asst managers who knew me from other stores. Got a new ASM a short time before evals were due. He called me in for my eval. I looked at it, and said "So you're fireing everyone else in the department, right?" He looked confused. I went on. "My sales per hour are routinely double that of everyone else in the department. There are 2 people in the store with total sales more then mine over the last year- both full time. You've got me as a poor performer in several areas and there's criticism of how I sell things, without telling me how to improve. And since I'm selling more then the rest of the department, there's obviously something wrong with this evaluation." I went on for a bit,
and got up and left. An hour later he called me back in and I signed my evaluation as a solid reliable employee with no criticism.

I have a severe allergy to subjective evaluations, especially ones that require ladder reanking. If there are metrics that can be measured, they can and should be used as the total basis for the evaluation. In my full time job, there are no metrics. You show up, relieve the previous operator, work for 8 hours, and get relieved by the next. Evaluations are done, because all bureacracies require them, but they are useless. They all read the same, "Shows up, does his job in a competent manner."

Bob Jones said...

I am a long time Microsoft employee and manager (managing teams of about 50 people).

The Microsoft performance evaluation system certainly has its problems, but this article has no clue as to what they actually are.

The basic idea of "stack ranking" is fine. It's based on the notion of "differentiated compensation". You want to reward your most valuable employees -- the 20% -- very well, and you want to reward your least valuable employees -- the 10% -- very little, if at all. The ones in the middle get a nice, industry-competitive, but not amazing, compensation package.

This is meritocracy at work -- in theory, anyway.

The problem is in the execution.

All large organizations impose policies to try to drive uniform behavior. Microsoft is no exception. Compensation and performance evaluation are areas where large organizations are particularly interested in uniformity and consistency.

Over time, the Microsoft performance evaluation process has become significantly more regulated and rigid. There are endless "checks and balances" that erode the ability of line managers to control the evaluation of their own employees. Quotas must be met; employees must be rigidly fit into particular buckets, and deviations from standard practice or historical data are quickly caught by the HR number-scrubbers and must be made right.

On top of that, there is special analysis and scrubbing that happens for female employees and URMs (under-represented minorities -- meaning, not Asians). This further alters the results, especially around the margins.

This often results in line managers feeling completely disempowered in the review process. It's not at all uncommon to hear of managers who disowned the review results they had to deliver to their employees. "I wanted to give you X, but management had to hit the curve, so you got Y instead." This breeds a great deal of cynicism in the process overall.

These problems are not unique to Microsoft. The real problems that Microsoft has faced over the last decade have little to do with the review process. But that's another topic.

bagoh20 said...

I have never wanted to work for a large company. I don't get the attraction considering the negatives. I don't even want my own company to grow much larger. I would rather break off new independent entities, and start over.

Small is beautiful.

Joe said...

In the early 90s Novell did this. The big problem was that the rankings weren't really based on merit, but who the managers, and especially, directors liked and who they didn't, regardless of effectiveness. I repeatedly saw smart, skilled, brilliant engineers get laid off while losers kept their jobs.

damikesc said...

Wasn't that basically what Enron did and it effectively poisoned their culture as well?

Joe said...
This comment has been removed by the author.
Joe said...

And yes, I soon became one of the engineers laid off. I was in good company.

(Novell had great engineers and first level managers and terrible directors and VPs. In my dealings with Microsoft, I got the same impression with the exception of some engineers on the Windows Mobile/CE team, who were completely unhelpful in an astonishingly arrogant way. This isn't a minor trend. I've dealt with many software vendors in my career; the vast majority had good to great engineers and shitty to unbelievably shitty management (I'm looking at you Wind River [VxWorks].))

Joe Schmoe said...

Dante, IBM survives because it completely changed its model. It went from a hardware manufacturer to a global IT consulting firm. IBM doesn't make many machines anymore. They just want you to pay them lots of money to come in, design, and set up your enterprise networks.

Jeny said...

employees performance management system is the best for any kind of company. Thanks.