In a stunning reversal of policy, Microsoft announced that it was abandoning the practice of “stack ranking,” in which every employee is ranked and rated, and those with the lowest ratings are fired.

Lisa Brummel, head of human resources, said in a statement:

No more curve. We will continue to invest in a generous rewards budget, but there will no longer be a pre-determined targeted distribution. Managers and leaders will have flexibility to allocate rewards in the manner that best reflects the performance of their teams and individuals, as long as they stay within their compensation budget.

No more ratings. This will let us focus on what matters – having a deeper understanding of the impact we’ve made and our opportunities to grow and improve.

The article says,

In the stack-ranking system, Microsoft managers had to place a set percentage of their direct reports into each of five silos, ranging from a “1″ silo for top performers to “5″ for the bottom performers. The bottom-ranked employees typically ended up seeking opportunities in other parts of the company or elsewhere.

Stack ranking has drawn continual fire from employees, many of whom felt the system rewarded internal politicking, withholding of information, and backstabbing, rather than innovation or cooperation.

A Vanity Fair article last year blamed Microsoft’s “lost decade” in large part on stack ranking.

In the stack-ranking system, Microsoft managers had to place a set percentage of their direct reports into each of five silos, ranging from a “1″ silo for top performers to “5″ for the bottom performers. The bottom-ranked employees typically ended up seeking opportunities in other parts of the company or elsewhere.

Now, if only we could get the stack ranking system out of the public schools.

Audrey Amrein-Beardsley has started a valuable new blog where she reports the latest news on VAM and interprets the latest research. She is one of our best researchers on the topic and, time and again, she has put a pin in the inflated hope that teachers can be measured like potatoes or corn.

n this post, she dissects Mathematica’s recent research on the value of moving highly experienced NBCT teachers to low-performing schools. She agrees that it makes a difference, but disagrees with the comparison group (which included 20% brand new teachers) and doubts that policymakers would be prepared to carry out the lessons on a grand scale.

What if we found that a class size of 10 was optimum for low-performing students? Would we be willing to implement the policy implications?