TIME SERIES MODELING OF PRODUCTIVITY DATA FROM CALL CENTERS IN TROPICS

Abstract

Recent studies in Tropics have identified call duration as a measure of productivity in call centers. In this study extensive survey data in conjunction with continuous measure of call duration amidst different agents in a call center has been used. The question examined is: Is there an impact caused by the day of the week or a particular hour of a day on work performance and does this effect varies between agent to agent in the building surveyed? This is looked at by carrying out an elaborate time series analysis of the obtained data. However, field measurements are not obtained under controlled conditions. Tie current study aims to use time series analysis techniques to analyze the data of call duration. The duration of a call is considered as an indicator of the performance of the individual agent who is attending to the call. The call center was ventilated with an outdoor airflow rate of 9.8 1/s/p and 22.7 1/s/p at either constant low and high designated temperature of 22.5?C and 24.5?C respectively. The important characteristic of the experimental design is the 2X2 blind intervention study with a sample size of26 agents. The subjects who were real call center operators were exposed to a set of different conditions for very consequent week. Tie analysis carried out shows that the results obtained have a significant difference in work performance level in the weekly intervention, day of the week and hour of the day among the agents across the nine weeks. These analysis are carried out in four different criteria such as actual average values of the most probable call duration for most number of the calls criteria, for 100% of the calls criteria, for atleast 75% of the calls criteria and for atleast 50% of the calls criteria. The results show that there is significant level of work performance for the weekly intervention; whereas the day of the week effect did not show significance expect Monday of week 5 when compared with week 6. The hour of the day effect shows only significance in the morning between 8.00 a.m. to 9.00 a.m. the implications concerning these four criteria shows that the performance level corresponding to 75% of the calls criteria is better when compared to other calls criteria.