PDQ Office Products. U.S. Distribution Strategy. Network Design Analysis Validation Recommendation

Transcription

1 PDQ Office Products U.S. Distribution Strategy Network Design nalysis Validation Recommendation 1

2 ontact Information For any questions about this case study or other Peach State capabilities please feel free to contact us. Peach State Integrated Technologies 3005 usiness Park Drive Norcross, G (678)

3 Table of ontents! Project Overview! Demand nalysis! ustomer Locations! Network Model and aseline! Scenario Evaluation! Summary and Recommendation! ppendix 3

4 Project Overview 4

5 Project ackground PDQ orporation is seeking a Distribution Network plan that will deliver high service levels during a period of growth.! Existing Facilities: PDQ currently has a single distribution center (D) located in central NJ. ll customers in the United States receive their shipments directly from this D.! Expansion: PDQ initially targeted the New York metro area and the northeastern states. large untapped market exists in the rest of the country, and PDQ must consider how to serve these customers.! Service Levels: PDQ s promise of quick delivery is central to the philosophy of the company. In Europe, most customers can expect 24 hour delivery. The goal of this study is to explore several service level scenarios for the US market and show the type of distribution network that would be required to achieve each target. 5

6 ! Examine the historical demand: uild profiles of typical customer orders to understand the current network.! Project future demand: Identify key markets that will be served in the future.! Rate the current network: Detail the service levels that can be achieved with the existing facility.! Design a network for the following scenarios:! aseline entral NJ only! How many D s to reach of the US in 24 hours?! est 1, 2, and 3 D networks (3 scenarios)! est entral NJ + 1 D network! est entral NJ + 2 D network Project Objectives Peach State and PDQ agreed on the following project goals and objectives to ensure success.! Identify service levels: With each scenario, calculate how many customers can be served in 1 day, 2 days, or more.! Review existing warehouse: Make recommendations around space efficiency and process improvements. 6

7 Project pproach & Methodology Peach State used historical shipment data and information about target markets to build a model of PDQ s network.! Input Data: PDQ provided detailed information about products, order activity, and shipment methods. The sample data provided covered June 2001 to June 2002.! Demand nalysis: Peach State examined the data and built profiles of the typical customer order, daily order activity, and shipment method. This data was also used to do a product velocity study which identified the fastest moving products.! ustomer Location: PDQ provided the PI buying power index which describes the population of office workers in each U.S. county. This data was used to identify the largest metropolitan areas in the U.S. which are important strategic targets for PDQ.! Network Modeling: detailed model of the network was built using the data profiles. Using this model, Peach State located facilities to meet customer demand. This approach minimizes distance to the customer base and meets service level targets.! Scenario nalysis: baseline scenario was compared with several other alternatives. Detailed maps and service levels are provided for each scenario. These will provide PDQ with a strong foundation for strategic expansion. 7

8 Demand nalysis What does the typical order look like? 8

9 Demand nalysis Peach State used historical data to build profiles of PDQ s shipping activity and customer demand. Data was used to answer key questions:! How many orders are received each day?! re sales levels increasing over time?! How many shipments are sent via LTL carrier? Parcel carrier?! What does the average order look like? How many products? verage sizes? Total weight?! Which products are the fastest moving? What percentage of activity do they represent? 9

10 Orders per Day The number of orders processed per day fluctuates, but it is most common to ship 55 to 65 orders per day. Number of Days Includes only 2002 activity. Orders Per Day Daily Order ount # Orders Shipped/Day Frequency of occurrence % % % % % % % % % % % % % % % % Total % 10

11 Shipment Volumes January 2002 marked a step increase in shipment volumes Monthly Order ount Data files analyzed only contained partial shipment history for June Daily Orders 0 Jun-01 Jul-01 ug-01 Sep-01 Oct-01 Nov-01 Dec-01 Jan-02 Feb-02 Mar-02 pr-02 May-02 Jun Shipment history after January 2002 most accurately reflects PDQ s current shipment rates Jun-01 ug-01 Oct-01 Dec-01 Feb-02 pr-02 Jun-02 11

12 ctivity Levels Per Day Though total volumes have increased, the spread of daily orders has remained constant. Number of Orders Order ount by Day of Month Order ount by Day of Week Day of Month Shipment volumes tend to drop off toward the end of months Sun Mon Tue Wed Thu Fri Sat Mondays seem to be the busiest day of the week. 12

13 Shipment Mode Parcel shipments constitute the majority of all shipments - more than 60%. Percent Parcel Shipments - Weekly 90.00% 80.00% Percent Parcel Shipments - Daily 70.00% 60.00% 50.00% 90% 40.00% Percent of Shipments via UPS 80% 70% 60% 50% 40% 30.00% 6/4/01 8/4/01 10/4/01 12/4/01 2/4/02 4/4/02 6/4/02 Though there is high daily variability, the trend line shows a stable ratio of parcel shipments over time. 30% 6/4/01 8/4/01 10/4/01 12/4/01 2/4/02 4/4/02 6/4/02 13

14 Lines Per Order Profile More than 50% of all orders were a single line only. Order picking can be designed to leverage this for increased efficiency. Lines Per Order Distribution Number of Orders % 74% 84% 89% 93% 95% 96% 97% 98% 98% 99% 84% of all orders have 3 lines or less. 90% 80% 70% 60% 50% 40% 30% 20% 10% umulative Percent of Orders Lines 0% 14

15 Units per Order Distribution Units and Weight Per Order Line, weight, and unit per order profiles can be examined to define the most suitable order fulfillment methodologies. Order ount % 88% 89% 91% 92% 93% 94% 94% 95% 90% 78% 83% 74% 80% 66% 70% 56% 60% 37% 50% 40% 30% 20% 10% lbs % Units Shipped 82% 86% 89% 92% 93% 94% 95% 96% 96% 97% 97% 97% % % 2500 Number of Orders umulative Percent of Orders Weight Per Order Distribution % kg 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% umulative Percent of Orders 15

16 Velocity Profile Product Movement Profile Order activity was analyzed for each product to distinguish fastest moving items from the slower moving ones. % Orders 80% 60% 40% 20% 0% # Products Product Movement # SKU's Picks % SKU's % Picks Fastest Movers ,712 33% 80% Medium Movers 870 3,324 33% 15% Slow Movers 930 1,109 35% 5% Total 2,669 22,145 16

17 Product Movement Profile Using the product velocity profile, Peach State was able to identify operational improvements PDQ should consider.! n overview map of the warehouse was developed which shows where products are stored.! Fastest moving items are shown in red, and are evenly distributed in the warehouse.! revised approach to product storage could decrease labor costs and shorten order cycle times OFFIE Receiving, Packing, Shipping rea Total Facings = Future ddition 987 = Fastest Moving Items 604 = Medium Moving Items = urrent 422 = Slowest Moving Items 17

18 Space and Process Improvements Peach State recommended the following enhancements for the existing warehouse, based on a one day site assessment:! Slotting -! Zone the warehouse by velocity, fast moving items = front, movers = middle, & Dogs = back.! Keep fastest moving items toward the dock & on the floor; opportunity exists to re-slot the warehouse.! Putaway store and putaway receipts of product by velocity zone; use first part of shift to replenish floor/forward slots from reserve.! llocate the right amount of product to a forward location; i.e., don t tie up 3 floor slots with same product if slots would be better served with other, fast moving SKU s.! ontinue to store items that ship together close to one another.! Operations -! djust workstation layout for packing to minimize travel time between work table, terminal, printers & UPS manifest.! Evaluate increasing the size of the UPS waves to create more efficient picking tours; update system parameters as needed.! Storage -! Rack out remaining floor space with selective pallet rack, minus space required for staging.! reating deeper bays for longer products will only offer marginal space improvements.! Use dense storage for small cube items (e.g., more half pallet locations, bin shelving, &/or case rack).! Put mezzanine over shipping docks; returns or small cube items could be processed on mezzanine.! Rack out over dock doors to store packing materials and empty pallets.! onsolidate dead items on pallet, inventory, ID, and put in back of warehouse.! Possibly use floor storage for high cube, stackable items. Detailed slotting will yield the greatest operational benefits to PDQ. 18

19 Summary of Demand The observations of the demand analysis were used to model PDQ s U.S. distribution network.! For the purposes of the network study, we assume that customers behave the same no matter where they live in the U.S. The order size, order value, frequency, and return rates all follow the typical customer profile.! The number of daily orders can fluctuate over a wide range. The distribution network must be flexible to allow for this variability.! Sales have been steadily increasing in the U.S. and this trend should continue as PDQ enters new markets. The distribution network must be able to handle this future growth.! Parcel shipments account for over 60% of the customer orders. This has been consistent over time, and is dependent on product type. PDQ must be equally capable of meeting service targets with both parcel and LTL shipments.! More than half of the orders are for a single product only, and 80% of the volume is driven by the fastest one third of products. PDQ should examine its shipping operations to build in efficiency. For example, single line orders could be batch picked to reduce travel time. Even small improvements in order fulfillment methodology could yield significant results. 19

20 ustomer Locations Where are the largest markets? 20

21 Geographic nalysis of Demand PDQ provided PI data which shows the population of office workers in the United States, divided into 3,109 counties.! Office workers are PDQ s target customers.! The data includes workers in large, small, and home offices.! laska and Hawaii were excluded from the study: this was only 0.12% of the total population. 21

22 Demand ggregation To speed calculations, Peach State grouped demand into large metro areas made up of several counties.! 302 Metro areas are defined by the U.S. ensus ureau.! 86 percent of U.S. office workers live in these areas.! Style conscious consumers tend to live in these metro areas also.! Marketing can be easily focused on these dense areas. New York Metro rea 22

23 Demand ggregation The largest 50 metro areas are used in the network study. These areas represent 58 percent of U.S. office workers.! The network model will focus on the largest cities first.! This level of aggregation allows faster modeling results but does not affect validity.! Final results will be reported using the full 3,109 county list and of the population. 23

24 nalysis of Demand ustomer orders and returns should follow the same geographic distribution as the population of office workers in the U.S.!alifornia, New York, and Texas have the largest concentrations of office workers.!the central area of the U.S. is sparsely populated compared to the coasts ustomers by State Percent of Total NY NY TX TX FL FL IL IL P P OH OH MI MI NJ NJ G G M M N N V V IN IN MO MO WI WI MN MN TN TN W W MD MD O O Z Z T T L L L L S S KY KY OR OR I I OK OK KS KS R R MS MS UT UT NV NV NE NE WV WV NM NM NH NH D D ME ME RI RI ID ID DE DE SD SD MT MT ND ND VT VT WY WY

25 Network Model and aseline What service can PDQ provide with its existing distribution center in central NJ? 25

26 ! The network model calculates actual road distances between distribution centers and customers.! ll customers are assigned to the closest D.! LTL shipments travel 500 miles per day on average. Network Model Peach State built a model of PDQ s network based on the following assumptions.! Parcel delivery times are based on UPS Ground service and are quoted in business days.! The model locates each distribution center to minimize average distance to the customers.! Return shipments follow the same pattern as outbound shipments. We assume there is a fixed percentage of all orders that will be returned. 26

27 aseline Network With just one distribution center located in New Jersey, PDQ will not be able to meet its service targets nationwide. This map shows service zones for LTL shipments. Each colored zone is an additional 500 miles away from the closest D. PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 4 Day Service 5 Day Service 6 Day Service 7 Day Service 27

28 aseline Network The first scenario includes only the existing facility in central NJ. This scenario was used as the index to rate all other scenarios. For each scenario, this table gives details for individual D s. This map shows which areas of the country are served by each D. Percent of verage Road verage Parcel D Location ustomers Distance (miles) Days in Transit ranbury, NJ 1, PDQ Locations Top 50 Metros entral NJ Statistics are shown here for the entire network. The LTL efficiency index is proportional to total road mileage. Lower numbers are more efficient. Overall Network 1, LTL Efficiency Index: 100 Parcel Efficiency Index: 100 The Parcel Efficiency Index is proportional to total parcel days in transit. Lower numbers are more efficient. 28

29 If PDQ starts marketing to the entire United States, only 29% of LTL volume and 20% of parcel volume will be within a one day service area from the existing facility. Percent of ustomers LTL Service Levels Within 1 Day Within 2 Days Within 3 Days Within 4 Days Within 5 Days Within 6 Days 99.7% 7 82% 85% 90% 6 73% 80% 5 70% 56% 4 60% 50% 3 29% 40% 2 30% 20% 1 10% 0 0% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. Percent of ustomers % 20% 15% Parcel Service Levels 35% 31% aseline Network PDQ uses both LTL and Parcel shipments. With the existing network, some customers can expect to wait a week for delivery. 66% 12% 78% 22% 1 Days 2 Days 3 Days 4 Days 5 Days usiness Days in Transit ased on quoted duration of UPS Ground service. This chart shows how long LTL shipments take to reach the customer. ll calculations are based on 10 hours at 50 miles per hour, for 500 miles per day. This chart shows how long Parcel shipments can take to reach the customer. The source data assumes UPS Ground shipment. 80% 60% 40% 20% 0% 29

30 Scenario Evaluation How many distribution centers does PDQ need? Where should they be located? 30

31 Scenario Overview Several different scenarios were considered using the model. Each was compared to the baseline to rank its performance.! One day service nationwide: How many distribution centers are needed to reach of the United States in 24 hours?! est 1 D network: If PDQ only has one D, where should it be?! est 2 D network: Where should 2 D s be located? How is service improved?! est network with entral NJ + 1 other D: If PDQ added one new D to its existing facility, where should it be placed?! est 3 D network: Where should 3 D s be located? What are the additional benefits?! est network with entral NJ + 2 other Ds: In addition to the existing facility, where should two new Ds be located? 31

32 Scenario: One Day Service When 10 Ds are placed to minimize cost, one day service nationwide is still not possible. Facilities could be moved into remote areas to achieve true one day service, but this would only increase the overall cost of the network. PDQ D Top 50 Metros 1 Day Service 2 Day Service 32

33 Scenario: One Day Service Distribution centers are located near major metropolitan areas to reduce the overall cost of the network. D Location Percent of ustomers verage Road Distance (miles) verage Parcel Days in Transit Manhattan, NY 23% Richland, TX 10% von Park, FL 5% Stone Park, IL 17% leveland, OH 11% Los ngeles, 9% Snoqualmie, W 3% San Ramon, 5% Golden, O 4% umming, G 13% Overall Network Many of these locations would be appropriate for local showrooms if PDQ wanted to have a presence in key markets. PDQ Locations Top 50 Metros Manhattan, NY leveland, OH umming, G von Park, FL Stone Park, IL Richland, TX Golden, O Los ngeles, San Ramon, Snoqualmie, W LTL Efficiency Index: 19 Parcel Efficiency Index: 44 33

34 Scenario: One Day Service With 10 facilities nationwide, service levels would be very high. The operating costs would also be extremely high. Percent of ustomers LTL Service Levels Within Two Days Within One Day Within Three Days 95% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. Most of the country receives 1 day service. Large geographic areas of the country receive 2 day service, but the number of customers in these areas is very small. Most parcel shipments will reach customers in one business day. Percent of ustomers Parcel Service Levels 99.5% 70% 69% 30% 0.5% 1 Days 2 Days 3 Days usiness Days in Transit ased on quoted duration of UPS Ground service. 80% 60% 40% 20% 0% 34

35 Scenario: est 1 D ecause most of the population lives in the eastern half of the country, a single facility would be located centrally. PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 4 Day Service 5 Day Service 35

36 Scenario: est 1 D Distribution activities are roughly 16-20% more efficient than the baseline scenario. This location was chosen to reduce the total network cost, but some customers will have slower service. Percent of verage Road verage Parcel D Location ustomers Distance (miles) Days in Transit incinnati, OH PDQ Locations Top 50 Metros incinnati, OH Overall Network LTL Efficiency Index: 80 Parcel Efficiency Index: 84 36

37 LTL Service Levels Within 1 Day Within 2 Days Within 3 Days Within 4 Days Within 5 Days 8 86% 7 82% 90% 73% 80% 6 70% 5 60% 4 50% 31% 3 40% 30% 2 20% 1 10% 0 0% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. Percent of ustomers Scenario: est 1 D This scenario cannot reach either of the largest markets in one day, but achieves higher overall service than the baseline. The UPS parcel network usually follows the same behavior as the LTL road network. Parcel Service Levels % % 47% % 22% 12% 12% 0 1 Days 2 Days 3 Days 4 Days usiness Days in Transit ased on quoted duration of UPS Ground service. Percent of ustomers 80% 60% 40% 20% 0% 37

38 Scenario: est 2 Ds y adding a second D, the network is able to serve both coasts with one day service, and reach most of the country within 3 days. PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 4 Day Service 38

39 Scenario: est 2 Ds Distribution activities are more efficient than with a single D. ased on customer percentages, the eastern facility is much larger than the second center. PDQ Locations D Location Percent of ustomers verage Road Distance (miles) verage Parcel Days in Transit Fairmont, WV 78% Pearblossom, 22% Top 50 Metros Fairmont, WV Pearblossom, Overall Network LTL Efficiency Index: 55 Parcel Efficiency Index: 77 39

40 Scenario: est 2 Ds Virtually all of the country can be reached in 3 days from the two locations in this scenario. Percent of ustomers LTL Service Levels Within Two Days Within One Day Within Three Days 83% 90% 80% 70% 49% 60% 50% 40% 30% 20% 10% 0% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. lthough there is a small area of southern Texas that receives 4 day service, the population there is minimal. 92% of the nation would receive a parcel shipment within 3 business days. Parcel Service Levels % % % % 10 11% 11% 8% 0 1 Days 2 Days 3 Days 4 Days usiness Days in Transit ased on quoted duration of UPS Ground service. Percent of ustomers 80% 60% 40% 20% 0% 40

41 Scenario: entral NJ + 1 D dding a second D to the current network has very similar results compared with the est 2 D scenario. PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 4 Day Service 41

42 Scenario: entral NJ + 1 D The existing site is not the optimal location for a D, but the efficiency of the network is not reduced significantly. Relocating the original D should be done only if other needs arise. D Location Percent of ustomers verage Road Distance (miles) verage Parcel Days in Transit ranbury, NJ 72% Edwards, 27% PDQ Locations Top 50 Metros entral NJ Edwards, Overall Network LTL Efficiency Index: 61 Parcel Efficiency Index: 79 42

43 Scenario: entral NJ + 1 D small section (3%) of the country receives only 4 day service in this scenario. Houston is the only metro area in this region. Percent of ustomers LTL Service Levels Within One Day Within Two Days Within Three Days 97% Within Four Days 73% 42% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Overall, service is only slightly lower than the scenario with the est 2 D network. Parcel Service Levels % % % 37% % % 23% 5 13% 0.2% 0 1 Days 2 Days 3 Days 4 Days 5 Days usiness Days in Transit ased on quoted duration of UPS Ground service. Percent of ustomers 80% 60% 40% 20% 0% 43

44 Scenario: est 3 Ds dding a third D brings the national service level even higher. PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 44

45 Scenario: est 3 Ds This may be a good long term plan for growing PDQ s distribution network. more detailed second study could identify the expected costs of this scenario. PDQ Locations D Location Percent of ustomers verage Road Distance (miles) verage Parcel Days in Transit Enfield, IL 50% Jersey ity, NJ 31% Palmdale, 19% Top 50 Metros Jersey ity, NJ Enfield, IL Palmdale, Overall Network LTL Efficiency Index: 41 Parcel Efficiency Index: 62 45

46 Scenario: est 3 Ds Most of the nation could receive 2 day service with a 3 D network. Shipments can reach the largest markets in only 1 day. Percent of ustomers Within One Day LTL Service Levels Within Two Days Within Three Days 94% 65% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Of the 50 largest U.S. metro areas only Seattle, Miami, and Fort Lauderdale would not receive at least 2 day service. ll customers could receive parcel shipments within 3 business days. Percent of ustomers Parcel Service Levels % % % 29% % 0 1 Days 2 Days 3 Days usiness Days in Transit ased on quoted duration of UPS Ground service. 80% 60% 40% 20% 0% 46

47 Scenario: entral NJ + 2 Ds This scenario is almost identical to the est 3 D scenario since the existing D is within 50 miles of the optimal location. PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 47

48 Scenario: entral NJ + 2 Ds The performance of this scenario is statistically identical to the est 3 D solution. PDQ Locations D Location Percent of ustomers verage Road Distance (miles) verage Parcel Days in Transit ranbury, NJ 30% Lancaster, 20% Henderson, KY 50% Top 50 Metros entral NJ Henderson, KY Lancaster, Overall Network LTL Efficiency Index: 41 Parcel Efficiency Index: 59 48

49 Scenario: entral NJ + 2 Ds gain, this solution provides the the same service as the est 3 D scenario. Percent of ustomers LTL Service Levels Within Two Days Within One Day Within Three Days 93% 90% 80% 66% 70% 60% 50% 40% 30% 20% 10% 0% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. The 2 day service level is only 1% less than the previous scenario. Percent of ustomers Parcel Service Levels 89% % % 32% % 1 Days 2 Days 3 Days usiness Days in Transit ased on quoted duration of UPS Ground service. 80% 60% 40% 20% 0% 49

50 Summary and Recommendation 50

51 Scenario aseline (ranbury) 1 Day Service est 1 D est 2 Ds ranbury + 1 D Scenario Summary - LTL This chart shows the performance of the LTL distribution network under each scenario in the study. est 3 Ds ranbury + 2 Ds These scenarios are located in the appendix ranbury & Reno ranbury & Las Vegas Number of Ds LTL Service Levels 1 Day 29% 95% 31% 49% 42% 65% 66% 35% 39% 2 Days 56% 73% 83% 73% 94% 93% 74% 73% 3 Days 73% - 82% 97% 94% 4 Days 82% - 86% Days 85% Days 99.7% Days verage miles from D to ustomer LTL Efficiency Index LTL Shipments 1,

52 Scenario Summary - Parcel This chart shows the performance of the parcel distribution network under each scenario in the study. These scenarios are located in the appendix Scenario aseline (ranbury) 1 Day Service est 1 D est 2 Ds ranbury + 1 D est 3 Ds ranbury + 2 Ds ranbury & Reno ranbury & Las Vegas Number of Ds Parcel Service Levels 1 Day 20% 70% 12% 11% 27% 30% 32% 22% 21% 2 Days 35% 99.5% 58% 67% 49% 84% 89% 51% 50% 3 Days 66% 78% 92% 86% 86% 86% 4 Days 78% % % 5 Days Parcel Shipments verage Parcel Days in Transit Parcel Efficiency Index

53 Recommendations Peach State recommends the following implementation plan for PDQ during their period of growth in the U.S. market.! Existing Facilities: The existing facility was placed in an important strategic location. It is not necessary to relocate this facility to increase customer service. However, relocation may be necessary for operational reasons, such as outgrowing the space.! New Facilities: dding a second distribution center near Los ngeles would achieve three day service levels for 97% of the target customers. ased on the population, shipments from this facility would only represent 25-30% of total U.S. sales volume.! dditional Expansion: If higher service levels are required or if sales volumes increase, a third facility could be added to the network. With a third facility, PDQ could provide 2 day service to 93% of the country.! Detailed ost nalysis: more in depth study would provide a view of PDQ s actual distribution costs. This analysis should consider the transportation, inventory, and operating costs of PDQ s distribution network. PDQ should consider all relevant costs to determine whether a new facility can be justified economically.! Inbound Shipments: Without actual cost data, it is not possible to identify the best ports to bring products into the U.S. Peach State can provide a framework for these decision for PDQ to use in the future. 53

54 ppendix! Scenario: entral NJ & Reno! Scenario: entral NJ & Las Vegas! PI Data: 50 Top Metro reas 54

55 Scenario: entral NJ & Reno Instead of placing a second D in alifornia, PDQ could locate in Reno, NV to share space with its sister company. efore: entral NJ + 1 D scenario PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 4 Day Service 55

56 Scenario: entral NJ & Reno This network is less efficient than other scenarios with 2 distribution centers. However, it is still a significant improvement over the baseline network. PDQ Locations Top 50 Metros D Location Percent of ustomers verage Road Distance (miles) verage Parcel Days in Transit ranbury, NJ 79% Reno, NV 21% entral NJ Reno, NV Overall Network LTL Efficiency Index: 64 Parcel Efficiency Index: 78 56

57 LTL Service Levels Within One Day Within Two Days Within Three Days 7 94% Within Four Days 6 74% 90% 80% 5 70% 4 60% 3 35% 50% 40% 2 30% 1 20% 10% 0 0% Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. Percent of ustomers Scenario: entral NJ & Reno Service levels are lower than the entral NJ +1 D scenario, but this may be an easier scenario to implement in the short term. 97% of customers would still receive parcel shipments within 4 days, compared to 99.8% for the entral NJ + 1 D scenario. Percent of ustomers % 22% 27% Parcel Service Levels 51% 34% 86% 97% 11% 3% 1 Days 2 Days 3 Days 4 Days 5 Days usiness Days in Transit 80% 60% 40% 20% 0% ased on quoted duration of UPS Ground service. 57

58 Scenario: entral NJ & Las Vegas ased on distribution alone, Las Vegas is not the best city for a second D. However, other factors may outweigh transportation. efore: entral NJ + 1 D scenario PDQ D Top 50 Metros 1 Day Service 2 Day Service 3 Day Service 58

59 Scenario: entral NJ & Las Vegas PDQ s facility costs may be less expensive in Las Vegas than in the Los ngeles area. Residents of alifornia would not have to pay sale tax, possibly providing an increase in sales. D Location Percent of ustomers verage Road Distance (miles) verage Parcel Days in Transit ranbury, NJ 71% Las Vegas, NV 29% PDQ Locations Top 50 Metros entral NJ Las Vegas, NV Overall Network LTL Efficiency Index: 61 Parcel Efficiency Index: 81 59

60 Scenario: entral NJ & Las Vegas Service levels in this scenario are similar to other options. Within One Day Within Two Days Within Three Days % % Hours from D to ustomer Drivers travel 500 miles per day, based on a 10 hour shift at 50 mph. Percent of ustomers LTL Service Levels 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Los ngeles is just 275 miles from Las Vegas, and could receive LTL shipments in one day. UPS would guarantee a parcel shipment within 2 business days. Percent of ustomers % 21% Parcel Service Levels 29% 50% 36% 86% 13% 1 Days 2 Days 3 Days 4 Days usiness Days in Transit 80% 60% 40% 20% 0% ased on quoted duration of UPS Ground service. 60

61 PI Data: Top 50 Metro reas Peach State used the top 50 metro areas to design the U.S. network. Rank by Size % US Office Workers (Sales Volume) umulative Percent of U.S. Population Metro rea entral ity State Zip ode Miles to D Est. LTL Travel Time (Days) Parcel Days in Transit U.S. Population Served by ranbury U.S. Population Served by West oast D losest D New York-Newark, NY-NJ-P % PM Manhattan NY ranbury, NJ % 0.0% % hicago, IL PM Stone Park IL ranbury, NJ % 0.0% Los ngeles-long each, % PM Los ngeles West oast D % 3.5% oston-rockton- Nashua, M-NH % NEM Waltham M ranbury, NJ % 3.5% % Philadelphia, P-NJ PM Philadelphia P ranbury, NJ % 3.5% % Washington, D-MD-V-WV PM McLean V ranbury, NJ % 3.5% % Detroit, MI PM Franklin MI ranbury, NJ % 3.5% % tlanta, G M tlanta G ranbury, NJ % 3.5% % Dallas, TX PM Dallas TX West oast D 1, % 5.1% % Houston, TX PM Houston TX West oast D 1, % 6.7% % Minneapolis-St. Paul, MN-WI M Minneapolis MN ranbury, NJ 1, % 6.7% % Orange ounty, PM Irvine West oast D % 7.9% % St. Louis, MO-IL M St Louis MO ranbury, NJ % 7.9% % Phoenix-Mesa, Z M uckeye Z West oast D % 9.0% % Seattle-ellevue-Everett, W PM Fall ity W West oast D 1, % 10.0% % leveland-lorain-elyria, OH PM leveland OH ranbury, NJ % 10.0% % Pittsburgh, P M Pittsburgh P ranbury, NJ % 10.0% % San Francisco, PM Daly ity West oast D % 11.0% % altimore, MD PM altimore MD ranbury, NJ % 11.0% % Tampa-St. Petersburg- learwater, FL M Tampa FL ranbury, NJ 1, % 11.0% % Denver, O PM urora O West oast D % 11.9% % San Diego, M Ramona West oast D % 12.7% % Miami, FL PM Miami FL ranbury, NJ 1, % 12.7% % Oakland, PM San Ramon West oast D % 13.6% % San Jose, PM San Jose West oast D % 14.4% Distances, travel times, and D assignments are from the entral NJ + 1 D scenario. Indicates metro areas that would be assigned to a D on the West oast under most scenarios. 61

62 PI Data: Top 50 Metro reas Peach State used the top 50 metro areas to design the U.S. network. Rank by Size % US Office Workers (Sales Volume) umulative Percent of U.S. Population Metro rea entral ity State Est. LTL Travel Time (Days) Parcel Days in Transit U.S. Population Served by ranbury U.S. Population Served by West oast D Zip Miles to ode losest D D % Kansas ity, MO-KS M Kansas ity MO ranbury, NJ 1, % 14.4% % Portland-Vancouver, OR-W PM Portland OR West oast D % 15.1% % New Haven-ridgeport-Stamford- Danbury-Waterbury, T NEM Shelton T ranbury, NJ % 15.1% % Milwaukee-Waukesha, WI PM Milwaukee WI ranbury, NJ % 15.1% % incinnati, OH-KY-IN PM incinnati OH ranbury, NJ % 15.1% % Indianapolis, IN M Indianapolis IN ranbury, NJ % 15.1% % olumbus, OH M olumbus OH ranbury, NJ % 15.1% % Orlando, FL M Orlando FL ranbury, NJ 1, % 15.1% % harlotte-gastonia- Rock Hill, N- S M harlotte N ranbury, NJ % 15.1% % Riverside-San ernardino, PM Ludlow West oast D % 15.8% % Nashville, TN M Nashville TN ranbury, NJ % 15.8% % Fort Lauderdale, FL PM Weston FL ranbury, NJ 1, % 15.8% % Fort Worth-rlington, TX PM Fort Worth TX West oast D 1, % 16.3% % Hartford, T NEM Hartford T ranbury, NJ % 16.3% % San ntonio, TX M San ntonio TX West oast D 1, % 16.9% % Las Vegas, NV-Z M Henderson NV West oast D % 17.4% % Greensboro-Winston-Salem- High Point, N M Greensboro N ranbury, NJ % 17.4% % Salt Lake ity-ogden, UT M Salt Lake ity UT West oast D % 17.9% % New Orleans, L M New Orleans L ranbury, NJ 1, % 17.9% % Norfolk-Virginia each- Newport News, V-N M Norfolk V ranbury, NJ % 17.9% % Raleigh-Durham- hapel Hill, N M Raleigh N ranbury, NJ % 17.9% % Memphis, TN-R-MS M Memphis TN ranbury, NJ 1, % 17.9% % Sacramento, PM Rancho ordova West oast D % 18.4% % uffalo-niagara Falls, NY M uffalo NY ranbury, NJ % 18.4% % Louisville, KY-IN M Louisville KY ranbury, NJ % 18.4% Distances, travel times, and D assignments are from the entral NJ + 1 D scenario. Indicates metro areas that would be assigned to a D on the West oast under most scenarios. 62