Problems Identified
Problem through the year
October November demand is higher than normal
Solution - Part time employees for high demand months
Problem through the month
2nd week has higher demand than normal
Solution - Overtime work for a week by employees
Problem through the week
Variation - Monday and Friday high demand
Supply Side Solution- Overtime on Mon and Fri
Demand Side Solution - Discount on lean days
Problem through the day
Heavy during 10 am to 11:30 am and 1 pm to 3:30 pm
Supply Side Solution - Have overlapping shifts such that there are multiple shifts in heavy workload period of the day
Decentralised
Indianapolis TC Kansas City TC Dallas TC
BTC BTC BTC
08:00-08:30 19 5 22 6 17 5
08:30-09:00 45 10 32 8 30 7
09:00-09:30 42 9 25 6 28 7
09:30-10:00 58 12 20 5 16 5
10:00-10:30 58 12 18 5 27 7
10:30-11:00 50 11 31 8 27 7
11:00-11:30 63 13 18 5 22 6
11:30-12:00 42 9 23 6 30 7
12:00-12:30 31 8 40 9 15 5
12:30-01:00 24 6 25 6 12 4
01:00-01:30 30 7 16 5 14 4
01:30-02:00 60 13 20 5 18 5
02:00-02:30 44 10 31 8 23 6
02:30-03:00 49 11 18 5 15 5
03:00-03:30 29 7 13 4 11 4
03:30-04:00 45 10 23 6 6 3
04:00-04:30 26 7 26 7 7 3
04:30-05:00 17 5 11 4 5 3
05:00-05:30 15 5 19 5 2 2
05:30-06:00 2 2 5 3 2 2
749 172 436 116 327 97
Centralised
Total Expected Calls TC required TC Required Difference \
(Decentralized) (Centralized)
58 16 12 4
107 25 21 4
95 22 19 3
94 22 19 3
103 24 20 4
108 26 21 5
103 24 20 4
95 22 19 3
86 22 17 5
61 16 13 3
60 16 13 3
98 23 19 4
98 24 19 5
82 21 16 5
53 15 11 4
74 19 15 4
59 17 12 5
33 12 8 4
36 12 8 4
9 7 3 4
1512 385 305 80
Ans 1)
The basic performance measure for megacard w.r.t the waiting line can be as follows-
1. Call waiting time- The time for which a caller has to wait before he/she is connected to a travel counselor.
2. Number of calls waiting- The number of call that are still to be connected to TCs determine the conversion speed of calls with respect to the demand(expected calls).
3. Calls handled per TC- Number of attended calls successfully handled by a TC in the half hour slot determines his or her efficiency.
4. Service time per call- Time for which a TC handles a call attended
5. Satisfaction of the traveler from the service provided starting from waiting to the call being handled.
Standard deviation measured in call waiting time, number of calls waiting, calls handled per TC and service time per call can be some other factors.
Ans 2a)
The important parameters for analyzing the Megacard system at different locations:
λ (Arrival Rate) – The rate at which customers are calling BTCs
μ (Service Rate) – The rate at which customers are being served in the system. This is 12 customers per hour
S (Number of serving channels) which is the number of TCs in the given case
Lq (Average Queue Length)
Wq (Average expected queue length) = Lq/ λ
Our aim is to find S where Wq <= 1 minute. In other words, we want to find S such that Lq < λ/30.
Ans 2b, 3, 4) Costs in Decentralised System
Indianapolis Kansas Dallas Centralized
Max Calls in any Half Hour 63 40 30 108
Number of TC req for that Half
Hour 13 9 7 21
Salary expenditure $ 390,000 $ 270,000 $ 210,000 $ 630,000
Total Employee Cost $ 573,529 $ 397,059 $ 308,824 $ 926,471
Total Cost $ 759,642 $ 525,906 $ 409,038 $ 1,341,644
Additional Expenditure
Costs Amount
Start Up Fixed Cost $ 100,000 This is one time
Maintenance Cost $ 60,000 This is per year
Reduction in Costs due to Centralization
Assumption : Costs other than employee Costs remain same, even if TC numbers go down; The maximum number for any half hour is used for
Cost Reduction $ 192,941
Assumption : Supply was being matched exactly to demand for each half hour; hence we save 80 TC Half Hours (385-305 HH)
TC Half Hours Saved 80
TC Hours Saved 40
TC Hours per Day 8
Number of TC decreased 5
Cost Reduction $ 220,588 Divided by 0.68 as salary expenditure is 68% of employee expenditure
Ans 5) Conclusion
These savings would add up over the years leading to a significant amount of Cumulative Surplus for the company. In a
Centralized System, we would be able to utilize the difference in time zones. Despite peak times being same across centres,
the time zones varied. Hence, people who were underutilized at one location, would compensate for the high workload at
other locations.
Hence, the above analysis guides us to recommend a Centralized system
pected calls).
