Saturday, October 1, 2005

Get Rid of The Curve to Reduce Microsoft's Headcount

Whew, is it time bring my hands together in a "T" and call time-out? There's a bit of nyah-nyah threading itself through the comments. Part of that is probably due to the lull between significant events: the build-up here at Mini-Microsoft over the past year timed with the one-two punch of Business Week and Forbes publicly bringing forth the discontent right before the reorganization of business groups.

And then? grumbling fatigue.

A comment from Saturday morning from someone flourishing at Microsoft:

I compare us to the rest of the companies in the world I think we do darn well. I have been in the workforce for over 24 years and have worked for several companies during those 24 years. In my 6 years at MS this is the best company that I have ever worked for. I have a great manager, I like the (client) group that I support, I love the work that I do and I have been promoted because of the results that I have delivered and customer focus not my political savy. [...] Like I said, if you think that there are high tech companies that are better and will compensate you what you think you are worth, then you should probably go there. You are obviously not happy at MS and you seem to lack any faith that anything is going to change.

If you're not happy at Microsoft and don't have hope and are not rolling up your sleeves to rewire every aspect of the Microsoft daily-business engine (and get lots of scorched oil squirted on your face), you need to freshen up that resume and look elsewhere. We have a lot of shipping happiness ahead to bolster our spirits, but we need to clean our cluttered house and that's not going to be pretty or enjoyable.

The house cleaning, starting from the bottom of the organization and cascading upwards, has to happen for Microsoft to be agile and efficient and productive and to ship more often and on time. We need our own one-two punch of:

  1. Organizational flattening to reduce waste of effort, and
  2. Headcount reduction for efficiency and cost savings (and to show The Street we're committed to a leaner, efficient running of our business).

Headcount reduction. Something to think about is why Microsoft is so bad at firing people (until, cataclysmically, entire groups are paved over or remade). First, firing someone for lack of performance is a terrible thing for a manager to through (and yes, it's a terrible thing to be fired). Unless as a manager you relish unleashing such misery, you're going to be resistant to it. You have to connect the dots that your poor performers are actually putting features at risk and most likely will result in the product either being buggy and slow or having less compelling features. You have to realize that you're not pleasing customers and increasing revenues and shareholder value by keeping around marginal or poor performers.

You have to get rid of your poor performers.

So you do. Let's say you get rid of a few folks over the course of a year. And since good, talented people are hard to find right now, expect those positions to still be open by the time the review model comes around. Then along comes the curve. Say you still have to provide a 25% fit for 3.0s, maybe 45% 3.5s, and 30% 4.0s. If your group is big enough, dropping a few folks didn't make a difference in the number of 3.0s you need to provide. You just screwed some of your 3.5 and, most likely, some of your 4.0 people to fit the curve.

D'oh. Where are those talking points from HR on giving hard review scores again?

Managers learn that lesson really well the first time they go through it. Better to keep low performing reports around to keep the 3.0 bucket full until you're absolutely required to flush them out. Or get some new hires to slap the "welcome!" 3.0 on. Otherwise, your high performers aren't getting the rewards you want and your eyes are twitching as you stare at that blue "Love 'Em or Lose 'Em" book on your shelf.

When Microsoft was roiling with growth, there was enough turn-over going on that the curve got lost in the noise. When we slowed and the dust settled, suddenly all aspects of the curve started to hurt, including 3.5 for people who have done great jobs. And 3.0 for people who had done better than expected. When I was at that manager meeting with HR, one lead related how his boss was ready to go on a hiring spree to help review-time compensation and morale in their group. What kind of hiring spree?

"We need to go hire us a bunch of 3.0s!"

Damn! The Microsoft manager was going out there to specifically hire poor-performers so that the rest of the people in his group could get better review scores (the HR presenter winced upon hearing the story but didn't say much). And you know those new hires didn't just go away quickly but rather settled down like a tick into Microsoft and managed their own success.

Sorry, this is indeed the bad management of The Curve when applied to Stack Ranking. This is the bad management of spoiling Microsoft with persisted poor performers that I doubt anyone has the strength to terminate. And you know that managers know how to avoid the people-review for 3.0s: how many calibration meetings have you been in when folks ruminate over the multi-3.0s and manage to get them to bob up to a 3.5 this time. The radar stops blipping.

My first thought here was to have a safe-harbor for firing people: every person you move out of Microsoft that's good attrition gets to fill one of your required 3.0s for the upcoming review model. The 3.0 bank, so to say. That sort of falls apart when you consider the next year and your bank account is empty. Then you're back to screwing people. I half like this idea. It does give managers the comfort of knowing that they are not going to trend down their really good and great performers just because they thinned their herd for the good of Microsoft and its shareholders.

So consider: if we got rid of The Curve, would it actually make it easier to move on the poor performers at any time during the year? I think so. I believe so. The Curve, for some groups, is actually causing Microsoft to stack up the dead wood so that the bottom of the Stack Rank and the bottom of The Curve are pre-filled.


No comments:

Post a Comment