Stack ranking - or curve ranking - where team members are rated against one another - is used at Microsoft and other companies. Basically, it's rare or near impossible to rank all team members at the same high rate, even if all team members have earned it, because the system requires that some are ranked highly, some are ranked as average, and some are ranked as below average.

This blog entry describes why it's a bad idea to rank employees against one another - it creates conflict, internal competition, and game-playing. I'm familiar with this, as I experienced it firsthand at one of my prior jobs. (Yes, the company was and is quite profitable but that's irrelevant to me - an organization can be successful in spite of its internal workings - the company was super-competitive internally and had [has] a reputation for not caring much whether you stay, go, live, or die, unless you are a 'star performer').

Back to the stack ranking system. I was in Borders at Redmond Town Center and I stumbled across Joel Spolsky's new book The Best Software Writing I: Selected and Introduced by Joel Spolsky (which was sheer luck - I had seen it before but mistaken the current cover for the previous book's cover). I flipped through it and by sheer luck compounded hit the article by Mary Poppendieck put in the section called "Team Compensation." (it can be found off of Ms. Poppendieck's site as well: http://www.poppendieck.com/pdfs/Compensation.pdf ). I read the article a little and decide to plunk down some cash for the book (I'm quite glad I did) and sat down in Starbucks and continued reading.

The article deals with the classical story of a team that did a fantastic job all around but now the manager is faced with stack ranking her employees. She says they all are the best (4.0) and has to deal with the consequences. I certainly recognized the listing of dysfunctional consequences that result from the competitive stack rank system:

1. Competition.
2. The perception of unfairness.
3. The perception of impossibility.
4. Suboptimization.
5. Destroying intrinsic motivation.

Something I didn't realize is that good ole Deming himself decried ranking review systems and I thank the article for bringing that up. Back when I was a Deming-nerd I didn't work for a stank-ranking organization so that particular key point was lost of me. The Deming article Gone But Never Forgotten has the following succinct gem:

* Remove barriers that rob people of joy in their work. This will mean abolishing the annual rating or merit system that ranks people and creates competition and conflict.

I'd like to hear from the Microsoft executive who thinks they have greater insights into team leadership than Deming.

Our current review system lacks honesty and integrity when you have to fit people to a curve and then tailor your feedback according to where they ended up. You're lying to them about their accomplishments so that they can be fitted into a compensation model. You can't give them truly honest feedback during the rest of the year because the end-result-curve may not match your kudos.

The only feedback I see mandated to give people during the year? Give them feedback and "message" them if they are at risk of getting a 3.0 so that it's not a surprise. We spend more time preparing the soft-landing for bad-news that folks who are doing great don't get to hear it.

and this gem of a comment in reply to another commenter who thought one should treat the company like a big family:

...It is a serious mistake to liken a company, especially a large one like Microsoft, to a family. The basic nature of the relationship(s) is far different.

Do you stack rank your children and spouse each year? Do you get rid of or "push out" the low performer(s) in your family? Do you allocate the benefits disproportionately to the highest performing members of your family? Do you warn your middle-ranking children that they had better improve or they are in danger?

This one is good, too. Presents an alternative.