The question is whether Microsoft management can figure out who is dead wood. Based on their recent product fiascoes, I am not sure they can. It can also be that management itself is dead wood.

good point.

I agree with your thoughts on this. Its always easier to layoff people and point to the economy

That is a good point, but I'm not sure. I have talked to a lot of Microsoft people in the past, and asked them a lot about Microsoft, and I would be extremely surprised if they had that much dead wood.

Especially because as they get bigger, their demand for 90th percentile types rather than 99.99th percentile types grows.

Nah, you get lots of dead wood at any big company. Nature of the beast. The hiring practices are more like an initiation rite than a real quality indicator.

Plus, 15% would be a pretty small amount of deadwood at a company that size. It may just be specialists hired for departments that don't exist anymore or people that stopped caring 5 years ago, but have been around so long that they can't be fired.

From what I've heard from friends at MS, it seems their dead wood is in management - there's a culture of ass-covering and doing the bare minimum on many teams, which goes a long way to explain why MS rarely does anything revolutionary. Great ideas that are risky don't make it very far in that company.

Exactly 10%-20% isn't that much. An company should be able to fire the worst at the bottom, hire new people to take their place distributed evenly across the spectrum & come out net ahead.

A sure sign of dead wood is when people deny any dead wood exists.

That's just as bad as "never eat on an empty stomach."

I would be extremely surprised if they had that much dead wood.

...

when people deny any dead wood exists.

The implication of the comment is that the poster spoke with MSFT employees who said things like "everyone is really hard working", "The people I work with are so smart", "Everyone pulls their weight". That is to say denials are rarely blunt. None the less they are still denials. In a organization of any significant size these generalizations have near zero chance of being true.