Everything We Knew about the Bell Curve Is WRONG!

Share This
December 20, 2016 | Ankita Poddar | HCI
bell curve gaussian distribution

Well, almost everything. The Gaussian distribution aka the bell curve has unwittingly become the most important mathematical function for HR professionals around the world. It stars in one of the most important HR processes – performance management; and consequently goes on to determine compensation. However, in recent time, more and more organizations have begun to question Gaussian’s presence at the workplace. Meanwhile, there has been an elder brother hiding behind the curtains who is finally ready to steal Gaussian’s limelight. Not unfamiliar to those who play in the field, it is none other than the ‘Power’ law!

The Gaussian distribution gained popularity because of its ability to explain the distribution of most things in nature – height, weight, speed, snowflake sizes and error rates. This distribution is easy to use, explains most occurrences, and has an average and standard deviation that helps predict the future to some extent. It is easy to see why people can come to love this distribution. However, a Gaussian distribution can’t realistically represent some things – exceptional performance being one of them.

This is where the Power law comes to play. Gaussian distribution fails to map accurately the output delivered by those in your organization whereas a Power law comes closer to the truth. One would expect that majority of the output is delivered by the bulk in the middle of the bell curve; however this is not quite true. The small group of exceptional performers overtake the delivered output. You may say that the tail of the bell curve represents them appropriately but how do you commensurate in pay? Is compensation then a true representation of their contribution? Also just how often have you tried to squeeze people out of the one end of the bell curve and wanted more people to enter the other end?

As Laszlo Bock mentions in his book ‘Work Rules’, exceptional contributors perform at a level so far above that of most, that they are able to pull the average way past the median. However, most compensation ranges fail to account for such exceptional performers. By a complicated combination of compa-ratio, performance and pay ranges, an exceptional performer gets a series of good raises until they hit the max of the range and begin to get minimal increases and sometimes none. One then begins to bend rules to ensure that your best person does not walk out the door.

So what do we do with our exceptional performers now that you know that the Gaussian distribution fails ever so often? I don’t have an answer to this question yet. One way would be to pay your top performers exceptionally well. There’s good reason why the top NBA players are paid far above any of their peers. Unfortunately, it is not easy to implement this in organizations. It takes a great deal of maturity and risk appetite to be able to pay one person in the same role 5x times the other. Maybe this is precisely why organizations choose to live in the false perception that the Gaussian distribution explains performance. It is the easier option.

Look around in your organization and think about your top performers. Look at their compensation and ponder on how many people would you trade for that single individual? There’s a high chance your bell curve doesn’t account for a lot of things and maybe this is when you need to evaluate whether Gaussian should stay in your organization or if he should walk out.