Shaping The Bell Right : Time to Performance Appraisal

 | November 08,2013 01:37 pm IST

A pulp-making unit hired 40 engineers from prestigious institutions, as management trainees who were toppers in their respective branches and institutions.


The management of the plant adopted a freakish policy with regard to performance appraisal - 10 percent of all employees were to be rated below average.

The management did not want all the employees to be ranked high, notwithstanding their excellent performance.
 

The axe fell on the trainees. The raters rated all the 40 trainees below average. Humiliated, these 40 put in their papers even before their training period expired.
 

The above bizarre case is the result of rank and yank method, which is the coolest thing in performance appraisal system these days. Under this method, employees are rated against one another on a subjective scale of 1 to 10. This concept of forced ranking, a tough minded approach, also entails that bottom dwellers get pushed out of the organization, if their relative positioning on the curve does not improve in the three consecutive years.
 

Strong Back
The Bell Curve system was pioneered by the leader of the century, Jack Welch early in his tenure at General Electric under the name "Vitality Curve". At present, 20 percent of the large corporates follow the 'Bell Curve Rating Method', the most prominent being Conoco, Hewlett-Packard, Microsoft, Ford Motors and notoriously Enron!
 

Explaining the term and the system, T. N. Hari, Director, Performance Management and Benefits, Daksh e-Services, says, "Call it by any name, 'Rank and Yank' strategy, 'up or out' policy, 'bell-curve' rating, or Jack Welch's 'vitality curve', what it actually implies is that at least ten percent of the company's strength has to be replaced every year. The system though controversial is fairly popular in HR circles abroad."
 

There is no single way of implementing this rating process - the concept encompasses any system in which individuals are rated against one another. The most common is where 20 percent are rated as 'excellent', 70 percent as the unspectacular but necessary back bone of the company, and the rest 10 percent as bottom-feeders, who are too poor in the performance to ever be trained, so the solution is to cut 10 percent of this superfluous flab every year.
 

Some companies rank their employees on a totem pole, one above the other, while others divide their staff into quartiles as done by Polaris. According to R. Shekhar, Senior VP and Head of HR, Corporate Strategy and Business Excellence, Polaris, "Performance at Polaris is categorized into four levels - Premium, Outstanding, Competent and Learning. These are indexed to 90th, 50th and 25th percentile. The Premium performers are at the top end of the industry, and that is consistent with our policy of institutionalizing meritocracy at all levels of the company."
 

According to Gautam Sinha, CEO, TVA Infotech, a recruitment agency, "The system is based on the normal distribution of employees on a bell-shaped curve. Companies use it to temper their appraisal processes order to correct managers who tend to overrate their people, as statistics prove that in any company only 20 percent of the total population can really be considered exceptional."
 

The normal distribution of employees in a bell shaped curve is shown in Figure 1:
 

Figure 1: Normal Distribution of Employees

Pitfalls
Although Forced Distribution Method is quite popular in western countries, it does not ensure its success in our country. It is severely criticized for being unethical, subjective, and unsuitable for small teams and creating a dysfunctional work environment. These pitfalls far out the way its contribution in terms of minimizing the errors of central tendency and facilitating comparative ranking.
 

Unethical
Bell Curve Method is in fact the most unethical form of performance appraisal system currently used. According to Praneet Mehrish, Country Human Resource Director, ST Microelectronics Ltd., "You cannot forcibly retire a certain section of your staff, every year. This would be unethical."
 

Further, according to Pradeep Nevatia, Vice President (Operations) & Country, Vetri Software India Ltd., "Such a concept may be working fine in the west but it is not certainly suitable for Indian companies. Here, if a person is asked to leave, other issues crop, such as who will feed his family, what happenS to his self-esteem, who will arrange for his re-engagement, etc."
 

Subjective
Bell curve method is highly subjective. As put by Madhukar Shukla, Professor, OB & Strategic Management, XLRI, Jamshedpur, "Since the bell curve is applied, not across all the employees, but to individual department / team / function, there is a good chance that the worst in the high performing group may be better than the best in an average performing group. Finally, the company may be left with low performers, while losing some good ones."
 

Further, applying this method year after year may result in erroneous results. The first time you may be cutting the obvious fat. But the second time you are cutting the interstitial fat. And the third time, you are only getting down to the muscle and the bone.
 

According to Hari Mohan Jha, VP (HR), ITC Welcome Group Hotels, "The highly subjective nature of this evaluation system often results in making people angry or ambivalent. The message that goes out to employees is that an overabundance of people cannot be allowed to perform at an optimum level, because this would skew the 'curve'."
 

Not Compatible for Small Teams
The biggest hurdle comes when bell curve has to be applied to smaller teams. According to Madhukar Shukla, "Since the model is based on the statistical characteristics of large groups of at least 30 people, it cannot be a valid differentiator if applied to, say, a group of 7-10 employees."
 

In smaller teams, there is so much proximity that it only leads to hoarding of knowledge and customers, because the last thing an employee wants to do is to share information with the people he or she is competing with.
 

Logically too, such a model cannot work for a very small group of extremely high or low performers for the simple reason that it force-fits them into predefined compartments. If it works, it can work only for a large, randomly selected sample.
 

.