It's been said so many times, but the competitive "stack-ranking" review process is toxic and broken. While engineering management pays lots of lip-service to teamwork and collaboration, when it comes down to actual execution, employees are (rightly) more concerned with individual achievement and the ever-elusive "visibility" in order to get high review scores which translates into promotions and potentially lots of monetary reward.
For PMs in particular, the stack-rank reviews cause people to either consciously or unconsciously over-engineer product features to have more cross-team dependencies and greater complexity, since this leads to increased visibility and scope. People are rewarded for this, regardless of success or failure of the product/feature.
The fact is that individual work requirements and deliverables vary so widely that it hardly makes sense to compare them to each other -- it's an apples and oranges exercise. Attempting to apply the mythical man-month to creative engineering roles is inherently folly.
Another side effect is that there ends up being a huge amount of collaborative and dependency overhead, which limits teams' agility and ability to deliver on short timelines. In turn, this prevents Microsoft from being able to innovate or even "fast-follow" in an ever more dynamic technology market.
Microsoft has huge resources and influence, but on the whole there are relatively few products/teams that are able to deliver real value to customers in a timely way. The cash-cows of Windows and Office continue to bolster the company financially (i.e. "nobody ever got fired for buying Microsoft"), which is a good thing for keeping the company profitable and stable, but it masks the fact that it's a lumbering beast of a company that's finding it increasingly difficult to compete and stay relevant in emerging technology markets.
