There’s been quite a bit of buzz lately on Microsoft and its future. One topic that’s been covered is the management practices of the company, where Mr. Balmer’s leadership – or lack of it – has been questioned.
In particular, there’s been a lot of criticism about the “Forced Ranking System” that utilizes the principles of the normal distribution (“Bell Curve”) to categorize staff into categories, such as high performers, often denoted by “1” or “A-players”, and low performers, denoted by a higher number or a succeeding letter.
To me, the notion of using the normal distribution for ranking employees is a clear sign of complete ignorance about statistics, as well as of human beings and their motivations.
Let’s start with the statistics part:
to use the normal distribution, you need a couple of things, first and foremost sufficient data points on identical attributes, in our current context, employees with identical jobs and tasks. Secondly, you need a very well defined metric, and a unit for that metric, for each of the attributes that you want to compare. An example of such an attribute is height (measured in cm or ft), an other would be shoe size.
So, even with only these two requirements, it’s easy to see that any attempt to use the Bell Curve for ranking employees is not going to work: how many professions or jobs are there where you have a large enough number of employees doing exactly the same job, i.e. having a totally identical attribute (job/task/responsibility) that you want to compare…? What metric that is not totally arbitrary and trivial will you employ ?
This type of thinking, that you will be able to rank professionals by using the Bell curve, to me smells really badly of the Tayloristic ideas about jobs and productivity from the 1900th century, and to me these approaches are totally out of date in our current knowledge based economy, but our MBA trained managers and HR folks still seem to believe that they can optimize organizations using this ancient paradigm.
The second reason why these forced ranking schemes are counter productive has to do with psychology, motivation and inspiration: again, our MBA’s seem to believe professionals behave like Pavlov’s dogs, that we will immediately drool if offered a reward, and change our behavior if punished, i.e. that we can be managed by the carrot & stick approach.
As the two articles below clearly demonstate, employee ranking schemes based on naive use of statistical methods, and particularly on the naive assumption that the Bell curve can provide some help, have clear counter productive effects for any organization.
It’s not just Microsoft that employs these naive attempts to rank staff, I’ve been exposed to very similar rankings over my entire career, and I’ve personally seen the perils of them (despite myself having most of the time been in the “A” category) .
As an example: if an organization has a unit, say consisting of some 30 staff with very different jobs, roles and responsibilities, but where the unit is responsible for a large part of the larger organizations (thousand’s of staff) total revenue, and always as a unit performs on top, and where each individual objectively is among the overall top performers of the overall organization: will you still place 30% (or whatever the current low water mark happens to be) of them in the low performer category…? And if you decide to do so, where do you think you will find replacements that are better…?
Let’s say you’d apply this naive MBA thinking to the group of 2012 Nobel Prize winners : applying the “logic” of the MBA’s, obviously, 30% of the laureates must by definition (given by the Bell Curve) be low performers… should they then really be awarded the Nobel Price…?
Or lets apply the same naive MBA logic to a sports team, such as a soccer team as another example: out of the 11 players on the field, 4 must then by the MBA definition be low performers. So lets sack the golie, he hasnt scored in a long time anyway, and the same goes for the defensive players….
See how rediculous this type of forced ranking is…?