There was certainly mention of "up or out", and a bad review or two would bring about the PIP stick (which seems to be only legal CYA before the firing, not for any hopes of real improvement). And, frankly, it was a long time ago and I haven't gone to any effort to remember details. I will say that in general it was as screwy and demotivating then as it is described to be now.
It was based on a model proposed by Ballmer's hero Jack Welch. What Ballmer forgot is that Welch himself said the "trim the bottom 10%" model didn't work once a company got larger than a certain size.
What Ballmer forgot is that Welch himself said the "trim the bottom 10%" model didn't work once a company got larger than a certain size.
I'm not sure that size is the issue. I think that the percentage of deadwood a company has being static is the problem.
A typical 10,000-person company probably has enough deadwood to make a 10% cut safe: a mature company that's accrued enough slackers and incompetents can endure a cut of the obvious nonperformers and few will be upset about it. Even many of the nonperformers will agree, to some extent, that it's time. Besides, the company will have a cut anyway if it underperforms. But if you cut 10% every year, after two or three iterations you are cutting good people, and the cultural effects of that make the process not worth it.
It was based on a model proposed by Ballmer's hero Jack Welch. What Ballmer forgot is that Welch himself said the "trim the bottom 10%" model didn't work once a company got larger than a certain size.