James O'Neill's Blog

May 19, 2006

The Mis-appliance of science pt2 – working for Microsoft

Filed under: General musings — jamesone111 @ 10:54 pm

Microsoft is a better place to work than it was last week.

I was out of the country on Friday. My phone downloaded the start of a long mail from Lisa Brummel – who I mentioned recently: A mail from a major name saying he was excited by Lisa’s mail, made it “Read NOW!” instead of “Read, eventually”. It takes something big for me to go “YES!” and throw my phone in the air when I’m in a customer’s building. That hoopy Lisa Brummel has only gone and got rid of the worst thing about working here: The Bell Curve

Mini-Microsoft feels like “Drawing hearts around her name” and Robert Scoble loved the changes (when he could say more than “Wow”). To quote the Seattle Times

The changes likely to have the biggest impact involve evaluation and compensation practices.The existing system doles out bonuses and promotions based largely on a controversial numerical rating scale. The number of employees who can receive a top score is fixed, sometimes forcing managers to give a lower score to a worker even though he or she might have performed at the same level as a peer.

Until Thursday Microsoft insisted that staff performances fitted a normal distribution. Other companies believe this myth of mis-applied science – that everything fits a normal curve. It doesn’t – not if you recruit the best people (and Microsoft does – new starters are often overwhelmed with the calibre of people.). If you do get a Bell Curve you’re failing. Why ? Let me draw you a picture.

The diagram on the left shows a Bell Curve for how the whole population would perform in a job: number of people is shown on the y axis, and performance on the x axis, 100% fall under the Green curve. If we recruit from top 10%, we’d get 100% under the red-curve – which is just part of the green one, stretched The actual performance people are supposed to deliver, is added in blue on the diagram on the right. The best possible performance is fixed, but the big spike in the “potential performance” (red) has been turned into lower performers (that’s the two shaded areas). There are 3 explanations for this

  • We didn’t recruit the people we thought

  • Our environment means about a quarter of people aren’t delivering their full potential.

  • The basis of the grading system is wrong,.

Previously, there was a distrusted and secret process where people were “stack ranked” (so helping others wasn’t in your best interests) . My previous manager sent his team a link to “Stack Ranking as a popularity contest”. Once ordered, grading forced the number of each grade in each group to match the Bell Curve – some people had to be doing well, the same number had to doing equally badly. Microsoft made $250,000 profit before tax per employee last year, how could 1/3 of our people be lousy ? THEY WEREN’T But the Bell Curve said they were. Common sense said some groups would have a greater proportion of “stars” the Bell Curve said they couldn’t

As a friend who runs European IT for another large American company who use the Bell Curve put it: “If a third of my people are turning in poor performances, why have I kept them? It’s just as well that every manager has the same number, otherwise I’d deserve to be fired”

The Bell Curve wasn’t enough to walk away from a job at Microsoft – though it would count against any company wanting to tempt me away. Even those who benefitted from it couldn’t make a case for it. How could we let it go on and still hold our values like taking on big challenges, openness, integrity, commitment to constructive criticism and self improvement ? A few railed against it, but most believed it couldn’t be changed, so said nothing, did nothing: proof that negative thinking is more potent than positive thinking, because it stops the possible from happening.

I mentioned I did a spring clean of my documents. Among them I found “True communication is possible only between equals, because inferiors are more consistently rewarded for telling their superiors pleasant lies than for telling the truth.” There was a widespread assumption that no-one would tell Steve Ballmer that this needed to change. I also found some stuff entitled “Colin Powell on Leadership”, which contains some corkers such as “Never let your ego get so close to your position that when your position goes, your ego goes with it.” Some aasumed Steve had too much invested in the Bell Curve to accept change.
I was introduced Thomas Kuhn’s “The structure of scientific revolutions”. last summer via William Leith’s book “The Hungry years” when it was book of the week on BBC Radio 4. It is about his experience of the Atkins diet; and he refers  to Kuhn’s idea that revolutions in thinking come from outsiders – Atkins was a cardiologist, not a dietician: people inside a comminity cling on to old assumptions. Lisa’s Channel 9 video explains that her first job in HR was to be head of it at Microsoft, at Steve Ballmer’s instance…did he put an outsider in because he saw we needed a revolution ?

On the flight home I made a list of 5 great things about this announcement – In descending order:

  1. Values. It proves they’re not just a puff piece, We do prefer the route which aids integrity. We do look for improvements. And we do take on big challenges.
    We shouldn’t assume that the person who prefers the hard truth to the easy lie can’t end up at the top of the company.

  2. Demonstrable Fairness. Some people stay out of management to avoid handing out unfair grades with “there weren’t enough arguments to stop others being pushed above you” .Now, if someone truly Underperformed they can be told what’s wrong (and what the possibilities are). And team members should see why one of them graded as “Exceeded”.

  3. Team working. It gets rewarded, rather than pushing others up the stack rank at your own expense. I think this means better products, and greater efficiency. We talk about “the people ready business” now we’re acting like one.

  4. Steve Ballmer’s ego isn’t an impediment to doing the right thing.

  5. A lack of cynicism. They could have announced this just before the annual MS Poll which measures organizational health. They waited till it had closed.

This change makes Microsoft a better company for all it’s stakeholders.

One other change caught the headlines; in a triumph of legumetrics over sense (or excessive attention to costs) in 2004 the company stopped providing towels in Redmond locker rooms. Adam Barr grasped that a towel has immense psychological value, and the change spawned public ill feeling. Lisa’s reversed that. I never thought I’d call our HR supremo “ a frood who really knows where the towels are” but as I said of her before it’s good to challenge those assumptions.


This post originally appeared on my technet blog.


Blog at WordPress.com.

%d bloggers like this: