Dr. Manish Gupte, has an interesting story to tell us with regard to peer appraisals at companies. Being a software engineer himself in the early part of his career, he had noticed unfair practices in evaluating top performers during appraisal cycles using the bell curve (grading employees as A, B, C & D) and decided to do something about it.
At the age of 26, while working as a Team Leader at a prominent IT firm, he developed a mathematical model based on Game Theory to encourage peer co-operation in Organizations. Game Theory uses mathematical modelling to obtain better results and decisions through interpersonal interactions.
He went on to pursue MS in economics, followed by PhD with a dissertation on peer appraisals at Purdue University, USA. Dr. Gupte has worked as a senior data scientist with noted companies like Oracle and has even taught Managerial Economics to MBA graduates. He is also the Founder of Mechanism Analytics, working in the areas of econometrics, human resource analytics, retail analytics, image processing, speech recognition, big data, text analytics and health.
Here is what Dr. Manish Gupte has to say:
My co-authors and I analyzed annual appraisals data, 360 degrees. We found that employees were forming groups on language and inflating ratings of their friends. We also found good employees were being pulled down by peers by giving them lower ratings. So, we used game theory and statistics to solve this problem.
We looked at names of people who were likely suspects. Then we used game theory. My co-author started weekly one on one sessions, where we watched for possible conflict. During these 1 on 1s we did not reveal the information managers had, and that way policed the employee, as any difference from manager’s observation was viewed with suspicion and reprimanded. So, we could make the employee honest and get more information from him.
There are situations where people lied! If might be because some think that if you pull down the top guy, you or your friend can take his place. This kind of collusion effects productivity of good performers. Such kind of behavior was handled in our case study.
We also asked less information whenever there was conflict. That way the employee found less gain by lying and on top risked getting punished. So, he spoke the truth. This way we could create a culture of honesty and positivity. Employee productivity and morale went up significantly.
At another company in Pune, my co-authors and I used continuous assessment to make employees work hard. We did not use divide and rule (like usual managers) even though employees had a strong group. We made sure employees remained friends but collected information on work through continuous assessment. That way productivity increased again.
Dr. Manish Gupte is now actively working with companies, where huge datasets can be analyzed and solutions devised to boost productivity and inculcate ethical appraisal practices at workplaces. Top performers can now be well recognized and get better job roles hence increasing productivity and morale.
Manish is also analyzing how middle managers can be made more honest. Game theory shows there is an incentive for middle managers to say they have done lots of “managing”. They can couple this with showing actual worker’s effort lower. One of Manish‘s friend, who is a founder, is not hiring middle managers for such reasons and he is ok not scaling his business. Statistics and econometrics based detectors for such game playing can make middle managers honest. Recognizing this game is the first step and it solves half the problem.
We hope many more companies come forward to review their HR practices to help create better work environments where performance is not bogged down by office politics.