AD605 Operations Management Institute for Marketing Deployment Help Desk HW Reading the case, complete all questions in the file.Turn in both an Excel file (as described above) and your case study report (one Word file).The report should include a cover page. The Word file should be clearly organized so that Randy Albright will be able to follow your analyses and easily read all tables and figures.Do not require the reader to open the Excel file when reading the report. BOSTON
METROPOLITAN COLLEGE
UNIVERSITY
DEPARTMENT OF ADMINISTRATIVE SCIENCES
IMD HELP DESK1
The Institute for Marketing Deployment (IMD) is a non-profit organization that assists small startups in
developing countries with marketing their goods and services. The organization is often associated with
microlending, where relatively small amounts are loaned to individuals to start small enterprises in
communities where traditional financing is unavailable. IMD’s headquarters is located in Toronto
Canada. It has subsidiaries in many countries, most notably India, Kenya, Chile, and South Africa. As
such, IMD requires strong information technology (IT) to facilitate the movement of information across
great distances.
The IT department at IMD (IMD-IT) supports the organization by overseeing the purchase, training, and
service of all IT resources. Their responsibility encompasses hardware, software, and associated devices,
such as printers, scanners, projectors, and other peripherals. IMD-IT supports all internal business units,
with its main internal customers residing in three departments: operations, finance, and human
resources. The operations department has personnel in locations around the world. Each week, IMD-IT
is available 6 days (Monday through Saturday) from 0900-1900 (9 am to 7 pm, Toronto time zone), for a
total of 60 hours per week.
IMD-IT is led by 58-year-old Randy Albright, who has worked at IMD since its founding in 1992.
Originally a political scientist, Randy took over IT responsibilities because no one was available with
significant IT expertise. Randy was somewhat of a computer hobbyist and he welcomed the challenge.
Over the years, his IT expertise has grown and he enjoys his work. The department includes three
functions: one devoted to purchasing hardware and peripherals, one devoted to purchasing and
updating software, and one devoted to troubleshooting (which handles both hardware and software
issues). The troubleshooting function operates a help desk through which all customer requests are
processed. Customers contact the help desk by e-mail, telephone, or walk-in.
Every customer contact with the help desk generates a help ticket. Each ticket is immediately routed to a
technician. Two groups of technicians currently exist – one group devoted to hardware problems and one
group devoted to software problems. No preference is given to customers based on how they contact the
help desk. Once the help ticket is written, jobs are assigned to technicians on a first-come first-served
basis. The help ticket provides formal documentation of the request, how the issue is solved, and the
responsible technician. Therefore, it provides a history of each job handled by troubleshooting.
Recently, troubleshooting customers have been complaining about long waits for getting their problem
solved. An analyst from the business analytics department (Mimi Li) is assigned to work with IMD-IT
and find a solution. Mimi is a 27 year-old graduate of Boston University’s Metropolitan College, among
the first to earn an applied business analytics master’s degree. Her undergraduate degree is in
mathematics. This is her first professional job.
Mimi’s first task was to collect performance information from previous troubleshooting help tickets to
determine the source of the complaints. According to data collected last year, about 40% of tickets had a
turnaround time of less than 45 minutes, but about 10% of tickets were closed after 4 or more hours.
Because a turnaround time of 4 hours constitutes about half of an internal customer’s work-day, many
tickets were closed on the day after they were opened. The turnaround time is working hours not real
1
This case was developed by John Maleyeff based on his work in service process capacity planning. All references to
people and organizations are fictional. © 2019 All rights reserved.
IMD Help Desk Case Study
Page 1
time. For example, a ticket opened at 3 PM on Tuesday and closed at 9 AM on Wednesday would be
recorded as a 6 hours turnaround time. The average turnaround time was about 55 minutes.
When interviewing customers, it became evident to Mimi that the main problem was not the length of the
wait, especially for those customers whose problems were complex, but the variations that cause
uncertainty among its customers. As one customer, Natalie Sutera, noted:
I am a frequent user of printing and scanning devices. I make a lot of requests for minor problems,
especially network connection issues. Most times, the problem is addressed and solved in less than
an hour. But, other times, I have to wait until the next day for the problem to be resolved. These
delays can cause my clients to wait and become dissatisfied. In some cases, the days caused them
to miss important deadlines. I wish the IMD help desk were more consistent.
Other customers were happy because the variations resulted in unexpected short turnaround times, even
for seemingly complex requests. However, it was clear to Mimi that IMD-IT needed to have better
control of customer turnaround times for their troubleshooting service.
Randy was not entirely convinced that Mimi’s help was needed. In his opinion, some technicians were
not working as hard as others. He suggested that a system be initiated where long lead times are flagged
so that he could talk with the responsible technician. In response, each month Mimi generated lists of
technicians and dates associated with the longest 10% of turnaround times. The lists, however, showed
that long turnaround times occurred about equally among the technicians. It also showed that almost all
technicians appeared on the list in some months but not in other months. Hence, Mimi believed that the
problem was systemic (i.e., a natural result of the way the system was managed) rather than being caused
by “bad” technicians. She decided to pursue this capacity planning issue in more detail.
After creating a better system for creating shift schedules for technicians at the IMD Help Desk, Mimi Li
set out to determine the specific staffing levels for various work shifts. For each work shift, two options
will be considered. The first option (Figure 1) is similar to the current system, where jobs are segmented.
Hardware problems are solved by hardware technician and software problems are solved by software
technicians.
The second option would consist of a common set of servers that handle both hardware and software
problems. Because much of the training and knowledge base of software and hardware technicians are
common, the amount of retraining would be manageable if this system were implemented. However, in
the long run it is anticipated that all-purpose technicians will be paid 10% more than technicians that
specialize in either hardware or software problems. In both options, problems will continue to be solved
on a first-come, first-served basis. Due to the number of reasons, it would be impossible for servers of the
segmented option to switch to solving other types of problems (e.g., hardware technicians cannot be
temporarily be reassigned to solve software problems).
Figure 1: Segmented Option
Hardware Problems
Hardware
Technicians
Figure 2: Common Option
All Problems
General
Technicians
Software Problems
Software
Technicians
IMD Help Desk Case Study
Page 2
Mimi started by overseeing the collection of service time data corresponding to hardware and software
problems. The analysis of hardware problem service times (not including queue times) is shown in the
Appendix. Based on this analysis, Mimi concluded that hardware problem service times are
exponentially distributed, averaging 45 minutes per customer. A separate time study estimated that
software problem service times are exponentially distributed, averaging 15 minutes per customer. Given
the current mix of hardware and software problems, Mimi estimates that, with the combined option,
problems will be solved in an average of 30 minutes, and that the service times will follow an exponential
distribution. Arrival rates would vary according to time of day and day of the week. But, the service
time distributions would be consistent across all time periods.
The average arrival rates for all customers are shown in Table 1. Customers for both types of problems
will arrive independently of one another the demand rates shown are valid during the duration of each
time period specified.
Table 1: Help Desk Demand
Weekday Early
9:00 am – 2:00 pm
Hardware
Customers
5.1/hour
Weekday Late
2:00 pm – 7:00 pm
3.0/hour
3.7/hour
Saturday
9:00 pm – 7:00 pm
1.2/hour
2.7/hour
Timeframe
Hours
Software
Customers
8.6/hour
Early on, Mimi sensed that although Randy hired her, it is clear that he does not appreciate the operations
management challenges. For example, he explained this about the common line option:
On early weekdays, we expect 5.1 customers per hour. Since service time per customer averages 45
minutes, this equates to 229.5 minutes of service, or 3.8 hours. Therefore, wouldn’t I just need to
assign 4 workers? What’s so hard about that?
As she begun work on analyzing the two options and creating a recommendation on which
option to choose, Mimi was keenly aware that her ability to explain and justify her work would
be just as important (or perhaps more important) than the technical aspects of this otherwise
routine project.
IMD Help Desk Case Study
Page 3
APPENDIX: Analysis of Hardware Service Times
The table below the hardware service time data (in minutes) for 75 customers. The time order of the data
is listed down each column (e.g., the first two data points are 32 and 27).
32
27
54
81
20
2
36
36
125
11
102
34
5
3
12
67
45
187
9
24
21
6
61
15
66
31
29
43
137
64
199
1
63
24
84
159
12
98
4
16
49
4
61
151
22
1
34
43
38
64
106
7
33
17
30
52
126
9
64
3
5
33
39
11
40
31
14
9
18
19
24
17
52
60
14
The time series plot confirms stability of the process generating the data. That is, the process is not
changing over time and therefore can be analyzed as one homogeneous data set. The histogram confirms
that the data are consistent with an exponential distribution. The average service time was 44.6 minutes
and the standard deviation was 44.2 minutes.
Time Series Plot of Service Time
Histogram of Service Time
200
25
20
Frequency
Service Time
150
100
15
10
50
5
0
1
7
14
21
28
35
42
Job Number
IMD Help Desk Case Study
49
56
63
70
0
0
40
80
120
160
200
Service Time
Page 4
BOSTON
METROPOLITAN COLLEGE
UNIVERSITY
DEPARTMENT OF ADMINISTRATIVE SCIENCES
AD 605 Operations Management
IMD Help Desk Case Study (Due November 5, 11:59 PM)
Study the case study document. Work with your team on the following questions. Important: For case study
assignments, it is NOT a good idea to start by splitting the questions among the team members. Each team member
should read the case and draft a response to every question. Then, the team should get together to perform the final
analyses and then (perhaps) split the writing of the final report (which should be edited to ensure continuity).
Important – make your report understandable by Randy Albright. His background and attitude about the
work are included in the case study document. Your grade will include evaluation of your ability to explain
concepts to management personnel who are not familiar with analytical methods like queuing theory.
Case Questions:
1.
Explain to Randy why his estimate of 4 workers needed during the early weekday period is flawed. Remembers
that his background is in political science with acquired IT expertise. Create an explanation that he will
understand.
2.
List up to six important assumption required to apply the M/M/s queueing model at the IMD Help Desk for
evaluating the two alternative presented in this analysis. For each assumption, write 1-2 sentences explaining
why it is a valid assumption.
3.
Perform an analysis using the M/M/s queuing model by evaluating each of the two options, over the three time
periods. Experiment with different numbers of servers and observe the balance of service costs (i.e., number of
servers, required expertise, etc.) and customer service (i.e., waiting time, queue size, etc.). Complete the following
table showing your recommended number of servers for each time period. State the criteria you used to make
these choices and justify your recommendations. Turn in an Excel file with the analysis details (clearly labelled).
Servers Recommended by Waiting Line Analysis
Timeframe
Option 1:
Software Problems
Option 1:
Hardware Problems
Option 2:
Common Problems
Weekday Early
Weekday Late
Saturday
4.
Based on the analysis of each option above, make an overall recommendation. Should IMD implement the first
option (segmented service) or the second option (common service). Be aware that it is not possible to
reconfigure the layout between times, and it is not possible for the segmented lines to share service workers. Be
prepared to have a team member present this result orally at the next classroom session.
Turn in both an Excel file (as described above) and your case study report (one Word file). The report should include
a cover page (with course name, case study title, date, and team members). The Word file should be clearly organized
so that Randy Albright will be able to follow your analyses and easily read all tables and figures. Do not require the
reader to open the Excel file when reading the report.
Purchase answer to see full
attachment