I can’t feel nothing but this chain that binds me/Lost track of how far I’ve gone/ How far I’ve gone, how high I’ve climbed/On my back’s a sixty pound stone/ On my shoulder a half mile of line (The Rising, Bruce Springsteen)
Curveball: a slow or moderately fast baseball pitch thrown with spin to make it swerve downward and usually to the left when thrown from the right hand or to the right when thrown from the left hand
The “Bell-Curve” has been the mainstay of performance ratings for a very long time now. The distribution of employee performance is forced to align with the assumption that the organization has a few high performers, a few low performers and a vast majority clustering around the ‘average’ performance level. The whole idea of trying to fit everyone on to a ‘bell-curve’ is thus based on the assumption that performance in an organization tends towards ‘average’ – and I guess you can immediately spot the inherent problem with this assumption.
No organization would like to be configured to be average; yet, everything from compensation distribution to employee engagement strategies continues to be guided by this basic rule of thumb.
Popularity or Performance?
The rather uncomfortable question that needs to be asked is what is the core driver behind the performance evaluation numbers/ranking? Increasing the rewards substantially between the top performers and the rest raises the possibility of a vast majority being disgruntled. Rewarding a few much more than the rest poses challenges for traditional notions of fairness and equality in the workplace. Not making enough distinction removes the incentive to strive for better performance – especially for the outliers.
The crux of the issue lies in the question: what portion of the employees is actually driving the business results.
In an a series of interesting studies, involving 633,263 researchers, entertainers, politicians, and athletes – researchers Ernest O’Boyle Jr. and Herman Aguinis concluded that performance follows a power law distribution more closely than a Gaussian distribution. Put in other words – a few outliers are responsible for a majority of the output.
Pareto’s rule, a popular example of power law distribution – is often referred to as the “80-20 rule” i.e. 80% of the output can be attributed to 20% of the causes. Often subjected to abuse, this ‘rule’ still remains an easy way of summing up the conclusions of the power law distribution.
This insight has interesting challenges for how HR will deal with evaluating employee performance and contribution in the future.
In a traditional setup, the typical manager grapples with one or more of following while doing an evaluation:
- Lack of clear quantifiable goals for the employee: Goal setting, an exercise often considered a mere formality and carried out with minimal participation from the employee and the manager comes back to haunt both when its time to evaluate performance.
- Lack of a proper understanding of ground-realities: Managers are humans and are victims of their own perceptions. In the absence of intelligent means to capture customer and peer feedback combined with the previous point of incorrect or inadequate goal, managers often have an incorrect or skewed opinion of performance – especially in large teams.
- Directive from HR to fit team into a bell curve: Small wavelets build up into a large wave, but small bells don’t make a big bell! The basic requirement that every team must have a Gaussian distribution borders on statistical absurdity – but continues to be a popular practice.
- Fear of (increased) attrition: The team member has been working on the customer account for most of the year. The training is time consuming and team members take almost a month to start contributing. Too low a rating might increase attrition and the manager will need to find replacements – which will again need to be trained! Putting everyone near average is much safer from the manager’s perspective.
The ‘Bell-Curve’ to some extent lets the manager play-it-safe with his evaluations. A Power-Law assumption takes away that safety-net increasing emphasis on getting the evaluations spot-on. Put another way – the Gaussian assumption lets the organization be popularity focused and dilutes engagement, while the Power-Law assumption goes to the other extreme and focuses heavily on performance – posing a major challenge to established notions of engagement.
It might seem tempting to abandon the existing (flawed) system of force fitting employee performance ratings into a Bell-Curve in favour of a Power-Law distribution, but such a move would be fraught with potential pitfalls.
Not only must Justice be done; it must also be seen to be done.
Remember that much of employee disengagement revolves around a sense of justice and fairness. The organization should not only do but also be seen as providing the required tools and facilities for the employee to succeed at the task, the organization should able to and also be seen as being able to correctly evaluate how the employee is performing at the task and then finally the organization should and also be seen as adequately recognizing the contribution made by the employee. Stumbling on one or all of these will lead to employee disengagement sooner or later.
In the power-law distribution a vast majority of the organization will be rated as “below average” The new assumption in no way implies that an organization should be only made up of only top performers – in the long run this is impractical for organizations of any size. The useful insight that can be gleaned from the new distribution is distinguishing between the employees who are vital and those who are not. Alarming as it may sound, when taking decisions of whom to retain or promote, making this distinction becomes a critical success factor.
Performance systems that can highlight top performers serve as powerful tools in the hands of HR and leadership of companies looking to engage with their employees and establish a culture of high performance. The impact of interventions based on inputs from a power-law trend of individual performances will expectedly be far higher and more meaningful compared to traditional systems. The challenge however will be eliminate bias in the evaluation – the impact of any such bias will be disproportionately higher in a Paretian distribution, with potentially disastrous consequences – a true curveball!
Acknowledgements and References:
Image courtesy of FreeDigitalPhotos.net
The bell curve is a myth – most people are actually underperformers, Michael Kelly, May 2012, BusinessInsider
The Best and the Rest: Revisiting the Norm of Normality of Individual Performance, HRMA Research Briefing.