Karan Chaudhry
ISEN 424-901 Final Report
Spring 2012
Poor Yorick’s Service Time Improvement
INTRODUCTION
The purpose of th...
Time in Line: Time Start Ordering: Time Left Queue:
10:36:55 10:38:00 10:38:50
10:37:15 10:38:25 10:39:35
10:39:30 10:40:1...
2:57:45 3:01:45 3:02:20
3:03:00 3:04:10 3:04:45
3:03:20 3:04:55 3:05:05
3:05 3:06:30 3:06:51
3:08:00 3:10:25 3:11:10
3:09:...
Time Start Time End
machine
1
machine
2 Blender
Total time
(in sec)
1 3:05:00 3:07:09 X 129.00
2 3:05:30 3:06:58 X 88
3 3:...
Table 2: Data for drink processing times
When collecting the data above in Table 2, the “Time Start” was taken to be the t...
31 2:39:12 2:39:30 X 18s
32 2:41:24 2:41:40 X 16s
33 2:43:13 2:43:25 X 12s
34 2:43:30 2:43:47 X 17s
35 2:46:55 2:47:05 X 1...
Figure2: Distribution for drink delivery times
Following that, we used the @Risk application to do an input analysis on th...
1. An additional pick up station should be made over by Machine 2 and the Blender with an
appropriate sign above it. This ...
CONCLUSION
This project was an attempt to reduce customer wait time at Poor Yorick’s. Taking action on my
recommendations ...
of 9

Final Report-Poor Yorick's Coffee House

Published on: Mar 4, 2016
Published in: Education      Business      Technology      
Source: www.slideshare.net


Transcripts - Final Report-Poor Yorick's Coffee House

  • 1. Karan Chaudhry ISEN 424-901 Final Report Spring 2012 Poor Yorick’s Service Time Improvement INTRODUCTION The purpose of this project is to reduce customer wait times at Poor Yorick’s Coffee Shop. The following paragraphs entail the type of analysis that we did, the shop’s simulation model, some recommendations to improve efficiency at your coffee shop and their benefits. PROBLEM STATEMENT Poor Yorick’s Coffee House, located on the first floor of Evans Library, is famous for its wide selection of fruits, muffins, croissants, sandwiches, gourmet coffee or even specialty drinks that it offers to thousands of students, faculty and staff every day. However, apart from its popularity, its big name has notoriously been linked to big waiting lines, thus leaving customers dissatisfied in the speed of service and making them leave the shop without a purchase. Consequently, the shop is not only loosing money but most importantly loosing goodwill of the customers. In an attempt to fix this, we collected data at the shop over the month of March. We then built a computer model for the shop to analyze the data gathered during the study. And then we tested alternative shop layouts and possible solutions to cut back wait times. DATA COLLECTION AND SUMMARY Over the month of March, we made several visits at the shop and recorded data during its busiest time i.e. between 10:30 AM and 3:30 PM. We split up into two teams of two members each. The first team was responsible for recording processing times of the customers while the second team was responsible for recording the operational times of the baristas. The first team positioned itself across from the two registers and recorded the time that the customers entered the line, the time that they spent to order and pay at the cashier and the time that they spent waiting for their order to be processed. Once they received their order, the customers were considered to be out of the system. Moreover, on at least three different days, we began collecting data when the shop had no customers in line and finished when the shop again had no customers in line. The data that we collected in regards to the aforementioned is summarized as follows. 1
  • 2. Time in Line: Time Start Ordering: Time Left Queue: 10:36:55 10:38:00 10:38:50 10:37:15 10:38:25 10:39:35 10:39:30 10:40:15 10:40:35 10:39:40 10:40:35 10:41:00 10:41:10 10:41:25 10:42:10 10:41:50 10:43:05 10:43:40 10:42:40 10:43:20 10:44:00 10:46:10 10:47:35 10:48:00 10:46:50 10:48:10 10:49 10:48:55 10:50:00 10:50:55 10:41:00 10:52:10 10:52:40 10:51:50 10:53:40 10:54:10 10:52:05 10:54:35 10:55:30 10:54:50 10:56:35 10:57:15 10:56:15 10:58:55 10:59:15 10:58:15 10:59:00 11:00:20 10:58:35 11:00:05 11:00:35 11:01:40 11:02:10 11:02:55 11:03:00 11:04:15 11:04:25 11:03:45 11:05:50 11:06:05 11:05:05 11:07:20 11:08:05 11:05:45 11:10:00 11:10:10 11:06:45 11:11:50 11:12:25 11:08:30 11:14:05 11:14:25 11:13:30 11:18:45 11:19:40 11:14:40 11:20:35 11:20:55 2:29:30 2:29:50 2:29:55 2:36:50 2:37:20 2:37:35 2:38:15 2:39:30 2:40:10 2:40:50 2:43:35 2:44:05 2:41:50 2:46:15 2:46:40 2:42:25 2:46:40 2:47:05 2:47:40 2:53:20 2:54:15 2:47:55 2:53:55 2:54:20 2:48:50 2:54:45 2:55 2:56:25 3:01:30 3:01:40 2:57:30 3:01:31 3:02:00 2
  • 3. 2:57:45 3:01:45 3:02:20 3:03:00 3:04:10 3:04:45 3:03:20 3:04:55 3:05:05 3:05 3:06:30 3:06:51 3:08:00 3:10:25 3:11:10 3:09:55 3:12:50 3:13:10 3:14:25 3:15:40 3:16:00 3:16:40 3:17:45 3:18:45 3:19:50 3:20:10 3:21:10 3:21:15 3:21:35 3:21:45 3:22:00 3:22:30 3:22:45 3:23:25 3:24:25 3:25:15 3:25:55 3:26:50 3:27:05 3:29:05 3:29:25 3:29:55 3:30:05 3:30:55 3:31:05 3:31:10 3:32:15 3:33:25 Table 1: Data for processing times of the customers at Poor Yorick’s Figure 1: Distribution for processing times of the customers at Poor Yorick’s As a major part of our analysis of the data in Table 1, we used the @Risk application, an extension in Microsoft Excel, to obtain the statistical distribution for processing times of the customers at the shop as shown in Figure 1. We found that triangular distribution best fit our data from Table 1. On the other hand, the second team recorded the time that it took the barista to make a drink, and the time it took the barista to deliver a drink to the customer after the drink was made. The data that we collected in regards to the aforementioned is summarized as follows. 3
  • 4. Time Start Time End machine 1 machine 2 Blender Total time (in sec) 1 3:05:00 3:07:09 X 129.00 2 3:05:30 3:06:58 X 88 3 3:07:00 3:08:55 X 115 4 3:08:00 3:10:27 X 147 5 3:09:40 3:10:50 X 70 6 3:10:00 3:12:45 X 165 7 3:11:00 3:15:30 X 270 8 3:13:20 3:16:20 X 63 9 3:15:45 3:16:55 X 70 10 3:16:02 3:18:20 X 138 11 3:17:25 3:19:20 X 115 12 3:17:38 3:21:53 X 255 13 3:20:50 3:24:25 X 215 14 3:21:37 3:23:32 X 115 15 3:22:15 3:26:58 X 283 16 3:22:20 3:28:21 X 361 17 3:27:23 3:30:48 X 205 18 3:28:01 3:34:28 X 387 19 3:32:04 3:33:45 X 101 20 3:32:23 3:34:28 X 125 21 3:34:42 3:35:45 X 63 22 3:35:02 3:36:15 X 73 23 3:38:30 3:40:20 X 110 24 3:39:30 3:40:38 X 68 25 3:41:58 3:43:37 X 99 26 3:44:21 3:45:36 X 75 27 3:44:30 3:45:03 X 33 28 3:46:40 3:48:20 X 100 Avg. Time for Hot Drinks=> Avg. for Machine 1&2(sec) 132.19 Avg. Time for Blended Drinks=> Avg. for Machine 1&3(sec) 180.29 4
  • 5. Table 2: Data for drink processing times When collecting the data above in Table 2, the “Time Start” was taken to be the time that the barista began making the drink while the “Time End” was taken to be the time when the barista ended pouring the drink in the drink cup. Time Start taking drink to customer Time End taking to drink to customer Machine 1 Machine 2 Blender Delay 1 11:52:50 11:53:00 10s 2 11:53:30 11:53:35 X 5s 3 11:54:50 12:00:30 X 5m 40s 4 12:00:24 12:01:20 X 56s 5 12:01:08 12:01:10 X 2s 6 12:02:03 12:05:30 X 3m 27s 7 12:05:20 12:05:44 X 24s 8 12:05:35 12:08:33 X 2m 58s 9 12:08:27 12:12:25 X 3m 58s 10 12:12:10 12:12:40 X 30s 11 12:12:28 12:13:13 X 45s 12 12:12:52 12:14:16 X 1m 24s 13 12:14:07 12:14:55 X 48s 14 12:14:48 12:14:55 X 7s 15 12:18:59 12:19:04 X 16 12:19:25 12:19:33 X 8s 17 12:21:04 12:21:16 X 12s 18 12:23:25 12:23:33 X 8s 19 12:23:38 12:23:58 X 20s 20 12:27:24 12:27:32 X 8s 21 12:30:25 12:31:32 X 1m 7s 22 12:32:51 12:32:56 X 5s 23 12:33:54 12:33:57 X 3s 24 12:36:25 12:36:33 X 8s 25 12:37:25 12:37:37 X 12s 26 12:38:49 12:39:58 X 1m 9s 27 12:39:45 12:39:50 X 5s 28 12:39:35 12:39:58 X X 23s 29 12:41:30 12:41:35 X 5s 30 2:36:22 2:36:35 X 13s 5
  • 6. 31 2:39:12 2:39:30 X 18s 32 2:41:24 2:41:40 X 16s 33 2:43:13 2:43:25 X 12s 34 2:43:30 2:43:47 X 17s 35 2:46:55 2:47:05 X 10s 36 2:49:30 2:49:41 X 11s 37 2:49:55 2:50:12 X 17s 38 2:51:55 2:52:10 X 15s 39 2:55:45 2:55:57 X 12s 40 2:57:52 2:58:01 X 9s 41 2:59:53 3:00:00 X 7s 42 3:05:30 3:05:45 X 15s 43 3:07:12 3:07:25 X 13s 44 3:09:10 3:09:19 X 9s 45 3:11:57 3:12:14 X 17s 46 3:15 3:15:50 X 8s 47 3:17:34 3:17:44 X 10s 48 3:20:15 3:20:26 X 11s 49 3:23:56 3:24:10 X 14s 50 3:26:10 3:26:16 X 6s 51 3:28:56 3:29:19 X 23s 52 3:31:55 3:32:40 X 45s 53 3:32:49 3:33:00 X 11s 54 3:34:00 3:34:20 X 20s 55 3:34:30 3:34:43 X 13s 56 3:35:35 3:36:00 X 25s 57 3:37:26 3:37:43 X 17s % for mach. 1 0.105263 delivery for machine 1 1.33333 % for mach. 2 0.315789 delivery for machine 2 4.17 % for blender 0.578947 delivery for blender 4.075 Table 3: Data for drink delivery times We then used the data from Table 2 and Table 3 to find its respective distribution(as shown in Figure 2) using the same procedure as explained earlier. The data in both cases was found to be normally distributed. 6
  • 7. Figure2: Distribution for drink delivery times Following that, we used the @Risk application to do an input analysis on the average times of the two machines and the blender to fit to a normal distribution. Our results are tabulated as follows. Input analysis on drink machines (in seconds)Input analysis on drink machines (in seconds)Input analysis on drink machines (in seconds)Input analysis on drink machines (in seconds)Input analysis on drink machines (in seconds)Input analysis on drink machines (in seconds) Machine 1Machine 1 Machine 2Machine 2 BlenderBlender Standard Deviation 1.33 Standard Deviation 1.29 Standard Deviation 1.26 Mean 0.516 Mean 4.17 Mean 4.08 Table 4: Input Analysis on drink machines Lastly, we observed that customers arrived individually at an inter-arrival time(time gap between each arrival) of 150.74 seconds. SIMULATION MODEL and RESULTS Using the results from the data collected above, we built the model of the shop using the simulation software called ARENA. Since our data collection was from 10:00AM until 03:00 PM, we set our model to run for that 5-hour period. Also, in order to get a much more appropriate cycle time(i.e. the time that a customer spends in the shop from the point to when he enters the line to the point after he gets his drink), we set the model to run for a 10-day period. As a result, we found the cycle time for the current shop layout to be approximately 9.32 minutes. RECOMMENDATION(S) Upon careful analyzing and then re-analyzing Poor Yorick’s current layout and taking into account the data collected, I have three recommendations to make, and they are as follows: 7
  • 8. 1. An additional pick up station should be made over by Machine 2 and the Blender with an appropriate sign above it. This will significantly cut the time that it takes for the barista to take the drink to the customer since the barista would literally be just able to turn back and hand over the drink to the customer as opposed to walking over to the other side of machine 1 which originally took an average of 4 seconds. As a result, this would reduce drink processing times. 2. An extra Blender should be added to keep up with the growing demand of iced drinks, especially with summer being around the corner. Adding this extra blender will enable the barista to make several drinks at the same time as opposed to waiting for the first drink to be made before starting on the second one which is how it was in the first case. As a result, this would reduce processing times of iced drinks even more and thus, in turn would cut back wait time of customers. 3. An extra cashier should be added to keep the lines smaller. Also, cashiers should be divided by the type of customer. In other words, customers who needed hot drinks would go over to the cashier by Machine 1, and then would pick up their drink by Machine 1. Similarly, customers who needed cold drinks or specialties would go to either cashier 2 by Machine 2 or go to cashier 3 by the Blender and then pick up their drinks at the common pick-up station as discussed in 1. This would cut back wait time of customers. I then built my simulation model for Poor Yorick’s which I called, “Efficient Poor Yorick’s Layout” which takes into account the above recommendations, i.e. having 3 cashiers, an extra blender and an additional pick-up station. I have tabulated the results as follows: System Cycle Time (in min) Current Poor Yorickʼs Layout 9.32 Efficient Poor Yorickʼs Layout 8.14 Table 5: Comparison of the results of the current model and its alternative From the above table, it can be seen that the alternative shop layout has reduced the cycle time by over 1 minute, which means that the customer wait time has been reduced by over 1 minute. 8
  • 9. CONCLUSION This project was an attempt to reduce customer wait time at Poor Yorick’s. Taking action on my recommendations would give you more business by improving shop efficiency; in fact, all customers stopping by would actually be making a purchase as opposed to leaving the line due to long wait times in the initial case. Moreover, this will also improve customer satisfaction and leave the customers with a piece of mind, thus giving them room to actually enjoy the products that your coffee shop has to offer. If you have any additional questions, comments or concerns about this project, please do not hesitate to contact me. I can be reached by phone at (979) 997-8565 or by email at kchaudhry@tamu.edu. Sincerely, Karan Chaudhry Junior Industrial & Systems Engineering major, Texas A&M University 9