The mini-microsoft blog is worth being thoroughly read, mainly because of the
comments! Yes, that's what I said. Blog comments better than its content. Unfortunately, with massive volume, the blog makes for some time-consuming read.
For today, I picked up the two entries on employee review:
Microsoft's 3.0 (or, How I Learned to Stop Worrying and Love The Curve); and
Microsoft Stack Ranking is not Good Management
Employee review systems fascinate me. As I've written
before, I believe organizational units (teams, companies, individuals, partners) maintain (just?) the efficiency required by its surrounding environment. It's the
Invisible Hand all over again.
What is so interesting about employee reviews, is that these are artificial inputs on a dynamic system. Employee
reviewsbonuses[1] exist only to leverage the intrinsic egoism, of individuals, on behalf of the company. They spark the Invisible Hand. As any artificial input on a dynamic system, review systems are fraught with misconceptions, wrong designs, misadapted assumptions or generally
leaky abstractions. Artificial inputs on dynamic systems are difficult.
All of this makes reading on Microsoft's review system quite interesting indeed. I was quite surprised to find out it is pretty primitive. It's a ten point scale, vertical feedback, bell curve-fitted yearly evaluation:
- There is no 360º evaluation, so peers and subordinates do not affect review.
- The bell curve fitting is an uniformization corporate-brain-death toll. (Tom de Marco defines it much better than I do)
360º evaluation can be questioned. It should be possible, in fact, to get good evaluation underway top-down (managers evaluate subordinates). However, it strikes me as odd tossing out the opinion of direct co-workers. This is valuable information, and should be taken into account during evaluation.
The bell curve fitting -- stating that 30% of the team must have 3.0, and no more than 10% may have 5.0 -- on the other hand, is a clear Magnum .357 shot on the foot. You are effectively stating there are no 100% excellent teams on your company (and no dumbass teams either).
Selected comments, with good points:
All in all, there are great ideas here on what a review system should
not be. Nice stuff to build on. I promise to balance this rather
destructive post with one defining my requisites for a good review system.
Notes:
[1] Employee reviews are as effective as the end result in employees' pockets. That may be as a result of a bonus or a promotion. Promotions can be delayed but not too much. Bonuses can be smallish, but never ridiculous. I once saw a company's review system reduced to shreds in five minutes when the audience noticed the yearly bonus ammounted to 25% of the monthly wage. Don't make that mistake.
Recent Comments
Deixei de fora muita coisa nas minhas
Não me expliquei bem. O desconto de ad
Estive a consultar a minha última factura,
Eu estou abrangido por uma promoção de