manual
TRANSCRIPT
Designing and Managing the Supply Chain
David Simchi-Levi Philip Kaminsky Edith Simchi-Levi
Solutions for Discussion Questions1
Kerem Bulbul
1We would like to thank Shiming Deng for his valuable contributions to the preparation of this manual.
Chapter 1
Introduction to Supply ChainManagement
Discussion Questions
Question 1
Pick any car model manufactured by a domestic auto maker. For example, consider the2002 Ford Thunderbird.
a. The supply chain for a car typically includes the following components:
1. Suppliers for raw materials
2. Suppliers for parts and subsystems
3. Automobile manufacturer (Ford, in this example). Within a company, there are alsodifferent departments, which constitute the internal supply chain:
i. Purchasing and material handing
ii. Manufacturing
iii. Marketing, etc.
4. Transportation providers
5. Automobile dealers
b. Many Þrms are involved in the supply chain.
1. Raw material suppliers. For instance, suppliers for steel, rubber, plastics, etc.
2. Parts suppliers. For instance, suppliers for engines, steering wheels, seats, and elec-tronic components, etc.
3. Automobile manufacturer. For instance, Ford.
4. Transportation providers. For instance, shippers, trucking companies, railroads, etc.
5. Automobile dealers. For example, Hayward Ford.
c. All companies involved in the supply chain want to maximize their respective proÞtsby increasing revenue and decreasing cost. However, companies may employ different
2
strategies in order to achieve this goal. Some of them focus on customer satisfactionand quick delivery, while others may be more concerned about minimizing inventoryholding costs.
d. In general, different parts of the supply chain have objectives that are not aligned witheach other.
1. Purchasing: Stable order quantities, ßexible delivery lead times and little variationin mix.
2. Manufacturing: Long production runs, high quality, high productivity and low pro-duction costs.
3. Warehousing: Low inventory, reduced transportation costs and quick replenishmentcapability.
4. Customers: Short order lead times, a large variety of products and low prices.
Typically, the automobile dealer would like to offer a variety of car colors and conÞg-urations to accommodate different customer preferences, and meanwhile have a shortdelivery lead time from the manufacturer. However, in order to maximize the length ofproduction runs, and utilize resources more efficiently, the manufacturer would like toaggregate orders from different dealers and offer less variety in car conÞgurations. Thisis a clear example of conßicting marketing and manufacturing goals.
Question 2
a. The supply chain for a consumer mortgage offered by a bank may involve various com-ponents:
1. Marketing companies that handle solicitation to potential customers.
2. Credit reporting agencies that evaluate potential customers.
3. The bank that extends the mortgage loans.
4. Mortgage brokers through which the loans are distributed.
b. The marketing companies strive to increase the response rate from homebuyers in orderto maximize their returns. Banks aim at a customer portfolio with a relatively low risk,healthy ßow of payments and low average loan maturity date. The brokers would liketo maximize their sales commissions.
c. Similar to product supply chains, the objective of a service supply chain is to providewhat is needed (in this case a particular type of service, rather than a physical product)at the right location, at the right time, and in a form that conforms to customer require-ments while minimizing systemwide costs. However, there are a number of differencesbetween the two types of supply chains. For instance:
1. In a product supply chain, there is both a ßow of information and physical products.In a service supply chain, it is primarily information.
2. Contrary to a service supply chain, transportation and inventory are major costcomponents in a product supply chain.
3
3. Services typically cannot be held in inventory, so matching capacity with demand isfrequently more important in a service supply chain.
4. In a service supply chain, the (explicit) cost of information is higher than in a productsupply chain. Note that in the mortgage example above, the bank has to compensatethe credit reporting agency for each credit report it obtains.
Question 3
Many supply chains evolve over time. For example, consider a memory chip supply chain.Production strategies may change during different stages of the product life cycle. Whena new memory chip is introduced, price is high, yield is low, and production capacityis tight, and the availability of the product is important. Consequently, production isusually done at plants close to markets, and the management focuses on increasing yield,reducing the number of production disruptions, and fully utilizing capacity. When theproduct matures, however, its price drops and demand is stabilized for a period of time,so minimizing production cost moves to center stage. To reduce costs, production may beoutsourced to overseas foundries, where labor and materials are much cheaper.
Question 4
A vertically integrated company aims at tighter interaction among various business com-ponents, and frequently manages them centrally. Such a structure helps to achieve sys-temwide goals more easily by removing conßicts among different parts of the supply chainthrough central decision making. In a horizontally integrated company, there is frequentlyno beneÞt in coordinating the supply chains of each business within the company. Indeed,if every business specializes in its core function, and operates optimally, an overall globaloptimum may be approached.
Question 5
Effective supply chain management is also important for vertically integrated companies.In such an organizational structure, various business functions are handled by differentdepartments of the company that usually have different internal objectives, and theseobjectives are not necessarily aligned with each other. This may be due to lack of com-munication among departments or the incentives provided by the upper management. Forinstance, if the sales department is evaluated based on revenue only, and the manufacturingdepartment is evaluated based on cost only, the company�s proÞt may not be maximizedglobally. Effective supply chain management is still necessary to achieve globally optimaloperations.
Question 6
The sources of uncertainty in this example include:
1. Factors such as weather conditions, diseases, natural disasters cause uncertainty inavailability of raw materials, i.e., peach crop.
2. Uncertain lead times during transportation of crop from the Þeld to the processingfacility may affect the quality of peaches, e.g., they may get spoiled.
4
3. Processing times in the plant, as well as the subsequent warehousing and transportationtimes are subject to uncertainty.
4. Demand is not known in advance.
Question 7
A small number of centrally located warehouses allows a Þrm to take advantage of riskpooling in order to increase service levels and decrease inventory levels and costs. However,outbound transportation cost is typically higher, and delivery lead times are longer. Onthe other hand, by building a larger number of warehouses closer to the end customers, aÞrm can decrease outbound transportation costs and delivery lead times. However, thistype of system will have increased total inventory levels and costs, decreased economies ofscale, increases warehousing expenses, and potentially increased inbound transportationexpenses.
Question 8
The choice of the particular transportation service depends largely on the types and sizesof products the company wants to transport, the inventory and delivery strategies and theneed for ßexibility:
1. A truckload carrier is better if delivering bulky items or small items in large and stablequantities from warehouses to demand points (stores). A good example is the deliveryof groceries from warehouses to supermarkets. Note that in this case we would like thedemand to be in increments of full truck loads.
2. A package delivery Þrm is more appropriate if relatively small items are delivered fromthe manufacturer/warehouse directly to the customers. Additionally, a package carriercompany offers more ßexibility by different modes of transportation depending on theneeds of the individual customers.
Question 9
1. High inventory levels
i. Advantages: High Þll rate (service level) and quick order fulÞllment.
ii. Disadvantages: High opportunity cost of capital tied in inventory, danger of pricedeclines over time and obsolescence, need for more warehouse space.
2. Low inventory levels
i. Advantages: Low inventory holding and warehousing costs.
ii. Disadvantages: Higher risk of shortages and lower service levels.
5
Case Discussion Questions — Meditech Surgical
Question 1
Meditech experiences poor service levels for new products, and inventory levels higher thannecessary for all products.
Question 2
There are many causes for these problems:
1. Demand is not studied in detail.
2. Information systems that record and monitor demand and inventory are poorly designed.
3. Forecasting errors are not tracked.
4. There is a tendency to shift the blame to the customers, e.g., panic ordering.
5. There are built-in delays and monthly buckets in the planning system.
6. The planning system ampliÞes small variations in demand.
7. Poor communication with customers; Meditech doesn�t typically see end-customer de-mand.
Question 3
The customer service manager is directly exposed to the complaints from the customers.Hence, he is in a good position to gauge the scope of the problems. Other managers donot face the customers, and they do not necessarily focus on their satisfaction.
Question 4
1. Recognize that demand is predictable, and establish better forecasting systems andaccountability for forecasts.
2. Institute better planning systems to eliminate planning delays; reduce the size of systemtime buckets.
3. Alternatively, put assembly within the pull system and eliminate bulk inventory com-pletely.
4. Develop and implement better information systems.
5. Improve communications with customers.
6
Chapter 2
Logistics Network Configuration
Discussion Questions
Question 1
The factors that affect the performance of the logistics network are not static, i.e., theychange over time. These factors include demand, product design, various costs in thelogistics network, regulations, contracts, etc. The effects of these dynamics need to beevaluated periodically in order to determine whether the existing conÞguration is stillsatisfactory given the new operating environment.
For instance, service level requirements may change due to increased competition whichtypically means that the lead time to fulÞll customer orders needs to be shortened. Thismay require the Þrm to redesign its logistic network and build new warehouses that arecloser to the end customers.
Question 2
The design of the logistics network is a strategic decision that has long lasting effects andimpacts all functions within the company. For the success of such a project, many levelsof the organization must be involved:
1. Upper Management: The new design must be aligned with the vision and strategic goalsof the company. Additionally, such a project may be costly, so management buy-in isessential to ensure that sufficient resources are devoted to the project.
2. Sales and Marketing: Demand forecasts and anticipated changes in product design andofferings affect the network and need the involvement of sales and marketing teams.
3. Manufacturing and Operations: The logistics network design has obvious impact on day-to-day operation of the Þrm. In order for the implementation to succeed, it is essentialthat the people involved with operating the system on a daily basis are involved in itsdesign.
Question 3
The decision that a single warehouse will be built has been made up-front. Therefore, weonly need to focus on the location and capacity of the warehouse, and determine how much
7
space should be allocated to each product in the warehouse. The main steps of the analysisare outline below.
1. Data collection
i. Location of retail stores, existing warehouses (5 warehouses located in Atlanta,Boston, Chicago, Dallas and Los Angeles), manufacturing facilities (a single man-ufacturing facility in San Jose), and suppliers.
ii. Candidate locations for the new warehouse.
iii. Information about products, i.e., their sizes, shapes and volumes.
iv. Annual demand (past actuals and future estimates) and service level requirementsof the retail stores.
v. Transportation rates by available modes.
vi. Transportation distances from candidate warehouse locations to retail stores.
vii. Handling, storage and Þxed costs associated with warehousing. Fixed costs shouldbe expressed as a function of warehouse capacity.
viii. Fixed ordering costs, order frequencies and sizes by product or product family.
2. Data aggregation. Demand needs to be aggregated based on distribution patternsand/or product types. Replace aggregated demand data points by a single customer.
3. Mathematical model building.
4. Model validation based on existing network structure.
5. Selection of a few low cost alternatives based on the mathematical model.
i. For the Þnal decision, incorporate qualitative factors that were disregarded in themathematical model, e.g., speciÞc regulations, environmental factors, etc.
ii. Optionally, build a detailed simulation model to evaluate these low cost candidatesolutions.
6. Decide where to locate the centralized warehouse.
With the centralized warehouse, service level will increase (less stock-outs) and inven-tory holding costs will decrease due to risk pooling. Also, Þxed costs associated with ware-housing will typically decrease, and inbound transportation costs from the manufacturingfacility to the warehouse should be less than the sum of the previous inbound transporta-tion costs. However, we will incur increased outbound transportation costs from the centralwarehouse to the retailers. In summary, the essential design trade-off is between trans-portation costs on one hand, and inventory holding costs and service level requirements onthe other.
Question 4
a. In automobile manufacturing, cars are usually delivered over land, and demand is con-centrated around major cities. Therefore, we would expect warehouses in this industryto be located near large cities with easy access to freeways and railroads. This wouldhelp to reduce the delivery lead time to dealerships in the cities.
8
b. In the pharmaceutical industry, overnight delivery is common. Therefore, proximityto a major airport is a factor that should be considered when choosing a warehouselocation. Additionally, for raw material warehouses it is important that these are closeto natural resources.
c. In the book industry, supplier warehouse locations would be affected by the availabilityof nearby natural resources.
d. In the aircraft manufacturing industry, sub-assemblies and parts are delivered by thou-sands of suppliers scattered all over the globe to the manufacturing facilities. Therefore,for these supplier warehouses, by far the most signiÞcant consideration is the ability toship parts easily and on-time, i.e., the proximity to railroads, freeways, harbors, etc.In such a capital intensive industry, we would also expect that regulations such as taxbreaks have an impact on potential warehouse locations.
e. With a large customer base shopping for books on-line, short delivery lead times arecrucial. Therefore, in book distribution, we would expect to Þnd large centralized ware-houses on reasonably priced land and where quick transportation modes are available.
f. Furniture manufacturing and distribution depends heavily on manual labor. Therefore,warehouses in this industry should be located close to cities with sufficient labor supply.
g. In PC manufacturing, outsourcing from all around the world is common where labor ischeaper and regulations favor the huge investments associated with high-tech manufac-turing. These considerations should be factored in when choosing candidate warehouselocations.
Question 5
In the pharmaceutical industry, we would expect more warehouses closer to the end cus-tomers for short delivery lead times. On the other hand, in the chemical industry therewould be fewer centralized warehouses in order to consolidate orders and decrease outboundtransportation costs.
Question 6
If we expect that the truck would travel empty on its return route, then TL rate wouldbe higher. Considering the example in Section 2.2.2, the probability that the truck comesback empty from Illinois (industrial heartland) to New York is lower than the correspondingprobability from New York to Illinois which explains the asymmetric cost structure betweenthese two cities.
Question 7
1. Handling Costs
i. Labor cost of workers in material handling.
9
ii. Costs of conveyors, fork lifts, automated guided vehicles (AGVs), etc., used to carrythe goods in the warehouse. Note that these costs have two components: variablecosts that are linearly proportional to the distances the goods are transported over;and purchasing costs of equipment that are proportional to the daily output requiredfrom the material handling system, but in a non-linear way because equipment ispurchased in discrete quantities.
2. Fixed Costs
i. Purchasing or rental cost of land.
ii. Cost of maintaining and operating the warehouse building which includes annualdepreciation and utility costs.
iii. Cost of racks that depend on the capacity of the warehouse.
iv. The cost of insurance for the facility.
3. Storage Costs
i. Opportunity cost of capital tied up in inventory.
ii. Cost of price declines while inventory is sitting in the warehouse. Note that thisincludes the risk of obsolete inventory that needs to be salvaged.
Question 8
An exact optimization technique is guaranteed to provide an optimal solution (if one exists)even if it takes a long time. On the other hand, a heuristic algorithm is a method thatwill Þnd good solutions to the problem in a reasonable amount of time where the terms�good� and �reasonable� depend on the heuristic and the particular problem instance. (SeeBramel and Simchi-Levi (1997).) The choice between an exact optimization technique anda heuristic algorithm for a given problem frequently depends on the trade-off betweensolution quality and solution time. Note that even if a heuristic algorithm (by chance)Þnds the optimal solution to a problem, it cannot conÞrm the optimality of the solution.On the other hand, for many problems there are no known optimal algorithms, so heuristicsmust be used.
Question 9
Simulation is a popular performance evaluation and modeling tool for complex stochasticsystems that cannot be evaluated analytically. A simulation model can closely reßect areal system and mimic its behavior, but it has some drawbacks: simulation is a descriptivetool, i.e., it cannot provide optimal values for system inputs. It generates, for a given setof inputs, sample outputs from the system that are used to compute statistical estimatesof the performance measures. Also, accurate simulation models of large systems requireextensive development effort, and typically take a long time to run. Thus, we advocate atwo-phase approach to solve difficult logistics problems:
1. Use a mathematical optimization model to generate a number of good candidatesolutions, taking into account the most important cost components.
10
2. Use a detailed simulation model to evaluate the candidate solutions generated in theÞrst phase.
11
Chapter 3
Inventory Management and RiskPooling
Discussion Questions
Question 1
Companies can cope with uncertainty by
1. keeping safety stock,
2. shortening production and order lead times,
3. using risk pooling strategies,
4. delaying product differentiation in the supply chain as much as possible, i.e., aggre-gating demand for parties upstream of the supply chain, and
5. by installing systems to achieve information sharing between suppliers and buyers,thus enabling collaborative demand forecasting.
Question 2
In general, higher inventory levels make it easier to maintain higher service levels. However,modern inventory management techniques may make it possible to increase service levelswithout increasing inventory levels as much as in the past.
Question 3
The variability in demand increases as the average and the variance of lead time increase.Therefore, for a given service level, inventory levels increase with longer lead times andhigher lead time variance.
Question 4
The target service level depends on the mission-criticality of the product. For instance,consider a service parts vendor for equipment for which every hour of down time is very
12
expensive. (See Example 6-5.) In this case, we would expect the management of the vendorcompany to specify a service level close to 100%.
Market conditions also play an important role in determining target service levels. Forcommodities, we would expect relatively high service levels since customers can switchproducts easily if they do not Þnd the particular product they look for. However, a lowerservice level may be acceptable if the product has a clear value differentiation comparedto its competitors. For instance, customers of a high-end server that is clearly deemedsuperior to the rest of the market may be willing to wait for 1-2 weeks if the manufactureris out-of-stock.
Question 5
The reorder level s = L ∗AV G+ z ∗STD ∗√L has two components. The Þrst componentL ∗ AV G covers the expected demand during lead time, and the second component z ∗ST D ∗ √L is the safety stock that protects against deviations from the expected demandduring lead time. Therefore, immediately before the order arrives, we expect that the Þrstcomponent is depleted completely and the inventory level is z ∗ ST D ∗ √L. Then, whenan order of Q units arrives, the expected level of inventory is Q+ z ∗ ST D ∗ √L.
Question 6
In the base-stock policy, at the time the warehouse places an order, this order raises theinventory position to the base-stock level (r + L) ∗ AV G + z ∗ ST D ∗ √r + L. Similarto the reorder level s in the continuous review policy discussed in Question 5, this base-stock includes two components: the average demand (r+L) ∗AV G until the order arrivesafter r + L periods, and the safety stock z ∗ ST D ∗ √r + L that protects against demanduncertainty during lead time. Thus, just before an order arrives, the expected inventoryon hand is equal to the safety stock z ∗ STD ∗ √r + L.
In order to determine the expected inventory level right after an order arrives at timet + L, note that when inventory is reviewed at time t, the inventory position is raised tothe base-stock level, and an order that was placed at time t− r arrives at time t+L. (SeeFigure 3-12.) Therefore, when an order arrives, the expected inventory level is L ∗ AV Gunits less than the base-stock level, i.e., is equal to r ∗ AV G+ z ∗ ST D ∗ √r + L.
Question 7
Observe that the longer L1, the more time the system has before allocation of inventoryto the retailers need to be made by the cross-dock facility. Thus, the longer L1 the morethe system can take advantage of the risk pooling concept. Hence, the total amount ofinventory is smaller when the cross dock facility is closer to the retail outlet.
Question 8
The answer is not immediately clear because the required safety stock depends both onthe average and the variance of the lead time. The retailer would have to make a decisiondepending on the relative effects of these two factors. See Section 3.2.6. Also, your decisionwould ultimately depend on the requirements of the retailer�s customers.
13
Question 9
For a mature product, it is reasonable to expect that the price and demand are stablein the short term. However, as the time horizon gets longer, and new products are in-troduced into the market, the demand and price for this particular product decrease andexcessive inventories may have to be written off. Thus, inventory holding costs related toobsolescence may be regarded as Þxed in the short term, but not in the long term.
Some storage costs are another example of inventory related costs Þxed in the shortterm, but variable in the long term. For instance, due to large inventories a company mayhave to rent multiple warehouses for a Þxed lease term. However, if inventory policies areimproved and turnover rates are increased in this period of time, then it may be possibleto rent fewer warehouses when renewing the lease contracts. Clearly, similar argumentscan be made for material handling equipment, storage racks, insurance, personnel, etc.
Question 10
Such deterministic models can be used as proxies for the more realistic stochastic modelsif the planning horizon is short, and the parameters of the problem are expected to berelatively stable over this time frame. However, most importantly, simple models canillustrate the basic trade-offs in a given type of problem which also translate into morerealistic and complex situations. For instance, the optimal policy for the economic lotsizing model balances ordering and inventory holding costs which is a general insight formore sophisticated systems as well.
Question 11
There are implicit and explicit penalties associated with a highly variable demand. Forinstance:
1. As discussed in Sections 3.2.5 to 3.2.7, the level of safety stock is proportional to thevariability in demand, i.e., the higher the variability in demand the higher the inventoryholding costs.
2. From a manufacturer�s perspective, highly variable demand means that utilization ofequipment will greatly ßuctuate, and equipment will sit idle when demand is low.
3. From a managerial perspective, high variability makes planning a very complex taskthat requires additional resources, sophisticated models and tools.
On the other hand, if a company is successful at implementing strategies to cope withhigh variability in demand, it may be possible to leverage on these to increase market shareand/or revenue if the competitors are not as successful.
Question 12
The factors that affect the choice of the supply contract type include the following:
1. Business convention: Companies tend to choose the contract form that is most commonin their type of business.
14
2. Information availability: The type of information available may dictate what type ofcontract can be implemented in practice. Depending on the contract type, suppliersand buyers require access to different types of information, and some information maybe difficult for the supplier to acquire but easier for the retailer, or vice versa.
3. Decision making and incentives: To achieve the optimal proÞt for the whole system andto allocate it properly, both suppliers and retailers must understand that decisions mustbe made collaboratively. All parties must be aware that they have to give up part of thecontrol in their individual systems, and the choice of the supply contract type dependson the level of control that parties are willing to share with each other.
1. Buy-back contracts:
i. Advantages:
(a) Commonly used in many businesses.
(b) The coordinating prices are not very sensitive to the demand distribution.
ii. Disadvantages:
(a) The supplier may have to buy back a large quantity of the product when demandis low.
(b) Extra transportation and re-stocking costs for returned items.
2. Revenue-sharing contracts:
1. Advantages:
(a) Easy to understand.
(b) The optimal values of the decision variables are not very sensitive to the demanddistribution.
2. Disadvantages:
(a) Need to monitor the total revenue.
3. Quantity-ßexible contracts:
i. Advantages:
(a) Commonly used in many businesses.
ii. Disadvantages:
(a) The optimal values of the decision variables are sensitive to the demand distri-bution.
(b) Extra transportation and re-stocking costs for returned items.
4. Sale-rebate contracts:
i. Advantages:
(a) It is a direct incentive to the retailer to increase sales.
ii. Disadvantages:
(a) Difficult to track and implement.
15
Question 13
Generally speaking, there is a lag between development of algorithms in the operationsmanagement literature and their adoption in business. There are several reasons for this:
1. Practitioners are not familiar with the literature.
2. It is hard to convince the industry that analytical models are useful. There is resistanceto change.
3. Possible gaps between theory and practice.
4. There are practical problems in implementation. For instance, without the developmentof information technology, it would have been hard for the movie studios to track end-customer demand and the revenue received by Blockbuster.
Also, in this particular case, the increasingly lucrative movie rental business due toadvances in home theater systems, and increased competition from media distributionover the Internet may have forced Blockbuster and the movie studios to re-consider andimprove their business models.
Question 14
(a) Risk pooling across locations: combining several warehouses into a single central ware-house.
(b) Risk pooling across time: using quarterly demand forecasts instead of monthly forecaststo do capacity planning.
(c) Risk pooling across products: designing products with maximum commonality anddelaying product differentiation in the supply chain as much as possible.
Question 15
If pricing strategies, service levels and quality of service in two stores are similar, then wewould expect the demand in these two stores to be positively correlated. However, assumethat while the overall market demand is relatively stable, one of the stores is running apromotion. In this case, we would expect that the promotional campaign would steal salesfrom the other store, so that the demand in the two stores would be negatively correlated.
Question 16
In the absence of historical data, judgment and market research methods would be mostuseful at the beginning of the life cycle of the Þrst Sony WalkmanTM. As more data be-comes available during the product adoption phase, time-series methods could be employedsuccessfully. Then, as the product matures and the manufacturer has a better understand-ing of the factors that affect demand, both time-series and causal methods could proveeffective.
When introducing a more recent Walkman model, we would expect Sony to rely muchless on judgment and market research methods compared to the very Þrst model. Years ofexperience, knowledge and data can be used to develop accurate quantitative time-seriesand causal models.
16
Question 17
a. The beneÞts of risk pooling increase as the correlation between demands decrease.Therefore, similarity of demand across the Þve regions makes the proposed system lessappealing.
b. The total cost of the decentralized system is $9,272 per week. In the centralized system,LA is the best location with the total cost of $6,545 per week. Please refer to theattached spreadsheet �Chapter 3 Question 17.xls� for details.
c. In this case, the minimum total cost is $8,808 per week, and the central warehouseis located in LA. In other words, the decrease in the inventory holding costs due tothe decreased lead time between the manufacturing facility and the warehouse is morethan offset by the increase in the transportation costs. Please refer to the attachedspreadsheet �Chapter 3 Question 17.xls� for details.
Case Discussion Questions — Sport Obermeyer
Please refer to �Harvard Case Notes: Sport Obermeyer Ltd., Teaching Note (5-696-012)31p, Janice H. Hammond, Ananth Raman� for solutions of the case discussion questions.
17
Chapter 4
The Value of Information
Discussion Questions
Question 1
a. Barilla experiences wild ßuctuations in pasta demand while variability in end-customerdemand is quite small. This ampliÞcation in demand variability in the supply chainis known as the bullwhip effect, and it strains Barillas manufacturing and logisticsoperations. Several factors contribute to this effect:
1. Transportation discounts, which induce distributors to order larger quantities lessfrequently
2. Trade promotions and volume discounts that create demand ßuctuations.
3. Delivery lead times of an average of 10 days from Barilla to the distributors.
4. Product proliferation, which makes forecasting more difficult.
5. Poor communication between parties in the supply chain.
6. Sequential decision making process in the supply chain, i.e., no collaboration.
The JITD program transfers decision-making authority for determining Barilla-distributorshipments from the distributor to Barilla. Rather than simply Þlling orders speciÞedby the distributor, Barilla would monitor the ßow of its product through the distribu-tors warehouse, and then decide what to ship to the distributor and when to ship it.This system alleviates many of the problem listed above, and enables Barilla to makemanufacturing and logistics decisions that beneÞt the entire system.
b. The most signiÞcant internal barrier to JITD is raised by the sales reps, who feel thatJITD would diminish their role in managing inventory and setting up promotions, poten-tially threatening their job security. Giorgio Maggiali needs to explain and demonstrateto the sales force that the proposed program would in fact increase customer servicelevel by reducing stock-outs, and potentially lead to cost savings. Ultimately, JITDwould help the sales reps to manage the orders more efficiently by increasing visibilityof the demand process. JITD is not a substitute for the sales force; it is a tool thatis made available to them for better customer service. Also, Maggiali needs to explainthat JITD is a company-wide effort, essential for Barilla�s long term success.
18
c. As a customer, JITD would at Þrst be disconcerting because I would be losing con-trol of my inventory. In order for me to agree to JITD, Barilla needs to convincinglydemonstrate the speciÞc beneÞts that JITD will have for me.
d. The proposed system will be effective if it can be implemented correctly, and indeed,subsequent results showed that JITD was very effective. In order to show value, itwould be useful to demonstrate that JITD beneÞts the distributors (lowering inventory,improving their service levels, and increasing their returns on assets) by running exper-iment at one or more of Barillas 18 depots. If customers will not agree to JITD, theymay at least agree to collaborative forecasting or increased supply chain visibility.
Question 2
a. E-commerce and the Internet allow upstream parties, e.g., suppliers, to have accessto more accurate demand information. It mitigates the bullwhip effect by preventingdistortion and miscommunication of demand information, and reducing the lead timein order processing.
b. Express delivery reduces lead times, and the associated demand variance. Note that inthe formulas in Sections 4.2.1 and 4.2.2, the variability of demand is proportional to thelead times in the system.
c. Collaborative forecasts help all stakeholders in the supply chain to arrive at a common,agreed-upon forecast of end-customer demand and reduce the bullwhip effect.
d. Periodic promotions create artiÞcial demand peaks and bottoms and increase the vari-ance in customer demand which ampliÞes the bullwhip effect. By everyday low pricing,these demand ßuctuations can be prevented, alleviating the bullwhip effect partly.
e. Vendor-managed inventory allows the supplier to monitor downstream demand and tomake a well-informed decision about how much to keep on-hand and how much to shipto its customers. Thus, the supplier does not have to rely on order data to forecastdemand and thus reduces the bullwhip effect.
f. Supply contracts align incentives in the supply chain, and reduce the uncertainty indemand by determining agreed-upon supply limits, thereby reducing the bullwhip effect.
Question 3
Sharing inventory is a form of risk pooling because it reduces the amount of safety stockneeded at each of the retail stores without decreasing the service level. Thus, sharinginventory reduces inventory holding costs or increases service level. Additionally, trans-portation costs may be reduced because goods are delivered from closer sources in case ofstock-outs.
Proper incentives must be in place to encourage retailers to share inventory. For in-stance, consider a case in which retailer one holds relatively less safety stock than retailertwo. Retailer one will experience stock-outs more frequently, and will order those itemsfrom retailer two. However, retailer two may not be willing to share its inventory unless itis compensated by retailer one for the additional holding costs it incurs, because retailer
19
two�s additional inventory is a (typically expensive) competitive advantage. For example,think about a small Chevy dealer and a large Chevy dealer in the same town. The largernumber of cars on the large dealer�s lot is an advantage�if he constantly supplies cars tothe smaller dealer, he loses that advantage.
Question 4
Several strategies can be employed to reduce the lead times in the supply chain:
1. Use EDI (Electronic Data Interchange) that reduces the information lead time in theorder process.
2. Use cross-docking to reduce/eliminate the time items spend in inventory.
3. Share inventory with nearby retail stores to reduce the lead time during stock-outs.
4. Share demand information throughout the supply chain to be able to respond to demandßuctuations rapidly.
5. Use delayed differentiation, which pushes generic products down the supply chain asmuch as possible, and allows the supply chain to more easily accommodate demand fora variety of related products.
6. Implement VMI (Vendor Managed Inventory) strategies so that suppliers can be moreresponsive to changes in inventory levels.
7. Use an express delivery service such as UPS to reduce the transportation lead time.
8. Adopt new technologies such as machine automation to reduce the manufacturing leadtime.
Question 5
1. The farmers prefer large and stable demand over time so that they can dedicate largeportions of land to a particular product, e.g., wheat, corn, etc.
2. The manufacturing division likes long production runs with minimal number of setupsand a low mix of cereal products which keep the production costs low.
3. The logistics division wants to ship full truck loads to reduce transportation costs.
4. The marketing division prefers a high product mix with short lead times. For marketingpurposes, the ability to offer different types of cereal is crucial.
5. The distribution arm of the grocery chain wants to replenish stocks as quickly as possiblewhile keeping inventory and transportation costs low.
6. The grocery store manager wants on-time delivery of a high variety of products withshort notice in order to be able to quickly respond to a highly diversiÞed end-customerdemand.
Note that the upstream parties in the supply chain typically prefer stable demand witha low mix while the more downstream parties prefer shorter lead times and a higher mix.
20
Chapter 5
Supply Chain Integration
Note: The solutions for this chapter are incomplete at this time. For morecomplete solutions, please see the textbook web site.
Discussion Questions
Question 1
A push-based supply chain has following advantages: 1) It utilizes long-term forecast andplanning; 2) It stores inventory in the system. If the products are generic, custom demandcan be satisÞed from inventory right away and customers waiting time is reduced.
A pull-based supply chains has following advantages: 1) It has minimal inventory inthe system. 2) It can better match supply and demand when products are differentiated;3) It reduces the reliance of production planning on forecast.
Question 3
The advantages of moving the push-pull boundary are as follows: 1) It increases the degreeof risk pooling and therefore reduces the inventory level in the system; 2) It reduces thereliance of production planning on forecast.
The advantage of moving the push-pull boundary later are as follows: If the productsare generic, custom demand can be satisÞed right away from inventory. It, therefore,shortens the customers� waiting time.
Question 4
1. Amazon: Amazon switched form pure pull to push and pull, because the sale volume hasgrown large. The boundary is at the their regional warehouses. This change is becausethe large volume allows Amazon to aggregate demand across geographical areas.
2. Peapod: Similar to Amazon, Peapod moved form pure pull to push and pull by estab-lishing warehouses which serve demand aggregated over an area larger than traditionalsupermarkets. Delivery time constraint and service level are also reason cause Peapodmoving to push-pull mode.
21
3. Dell: The push-pull boundary in Dell supply chain is at the assembly. Computercomponents are standard, their demands are aggregated over all Þnished products thatrequire those components. They can be managed using push mode to take advantageof risk pooling and economy of scale. Computers are customized according to theconÞguration speciÞed by buyers, the assembly of computers should be managed in apull mode.
4. Furniture companies: The production is based on pull but delivery is push, becausethe transportation cost for furniture is high and furniture companies need to aggregatedemand over some period to take advantage of risk-pooling and economy of scale.
Question 10
Consider Wal-Mart.
1. Cross DockingSuppliers for nonperishable products such as beer, rice, and shampoo, for which thetotal sale volume is large but demand at each individual stores is low, are suited toa cross-docking strategy. These products can be cross-docked with other low-volumegoods to create full-loaded truck for each store.
2. Direct ShipmentSuppliers for perishable and time sensitive products such as fruits and milk are suitedto a direct shipment strategy. Goods that are ordered by retailers in full truck loadscan also be shipped directly.
Case Discussion Questions — The Great Inventory Correction
22
Chapter 6
Strategic Alliances
Discussion Questions
Question 1
a. Consider a company for whom logistics is a core competency, and that has sufficentresources to implement the strategy.
b. Consider a company whose products are perishable, and for which environmental condi-tions must be tightly controlled during warehousing and transportation. (For instance,a biotechnology or a pharmaceutical company.) In this case, the company may notwant to give up control of products during logistics operations in order to ensure theirquality, and it may choose to acquire a logistics company with appropriate expertise.
c. Consider a company that has internal logistics expertise, and that is trying to expandits operations into new markets. If the company is already resource-constrained, itmay make sense to develop the strategy in-house, but delegate speciÞc operations tothird-party logistics providers.
d. Consider a new company that established its manufacturing operations recently, and isstruggling to ramp-up its output. In this case, it may make sense for this company tofocus on design, manufacturing and sales operations of its products and outsource allactivities related to transportation and warehousing of raw materials, and distributionof Þnished goods.
Question 2
Logistics is a complex set of tasks that requires vast resources, analytical expertise, relatedsoftware and a signiÞcant amount of investment in information technology which goes outof date very rapidly. Thus, unless logistics is a core competency of the company and it hasresources to dedicate to it, it is difficult to design and implement effective logistics oper-ations. Therefore, companies are increasingly willing to outsource logistics operations tothird parties who have the analytical expertise, necessary software and sufficient resourcesto keep their information systems up-to-date. In addition, since these third party logisticsproviders have several (or many) customers, they have sufficient economies of scale to makethe necessary capital investments to operate an advanced logistics system.
23
Question 3
1. Quick response: In a quick response system, the retailer determines the order quantitiesand replenishment times, while suppliers analyze POS data to improve their forecastingand production scheduling. This system could be preferred when the retailer-supplierrelationship is new, and trust between the two parties has not been fully developed yet.In this strategy, the retailer has complete control on its inventory, but helps suppli-ers improve operations by providing POS data. Additionally, this type of partnershipcould be preferred if Þnancial and personnel resources to develop a more integratedrelationship are not available.
2. Continuous replenishment: In a continuous replenishment system, vendors receive POSdata and use these data to prepare shipments at previously agreed-upon intervals tomaintain speciÞc levels of inventory. This type of partnership is a system between quickresponse and VMI, because suppliers and buyers together agree on target inventory andservice levels. It involves less risk for retailers than VMI, and typically leads to a morestable and long-term relationship between suppliers and retailers than quick responsedoes.
3. In a vendor-managed inventory system, suppliers decide on the appropriate inventorylevel for each product and the appropriate inventory policies to maintain these levels.The retailers give the vendors full authority to manage inventory replenishment. Thissystem is more integrated than the previous two systems, and requires a high levelof trust between the supplier and the buyer. If implemented properly, VMI can leadto more overall system savings than the other two types of partnerships. However,VMI requires more commitment, and initially, signiÞcant investment in informationinfrastructure, time and personnel.
Question 4
Information sharing in this case is most beneÞcial if the production capacity of the supplieris tight. The supplier can use the weekly POS data to forecast the order size that will arriveat the beginning of next month, to adjust its production schedule accordingly given thetight capacity, and to reduce the risk of not being able to meet the demand.
Question 5
Initial VMI schemes in which ownership of goods transferred to the retailer when goodswere delivered beneÞted suppliers by giving them complete control of their system. Ad-ditionally, it also provided an incentive to the suppliers to ship as much inventory to theretailer as permitted under the contract. This was an important disadvantage of thesecontracts because the overall system proÞt was not necessarily maximized.
The more recently developed consignment schemes have clear inventory and managerialcost reductions for the retailer since the supplier owns the goods until they are sold.However, this strategy still allows the supplier to manage its operations more efficientlybecause it makes all the production and distribution decisions. Note that inventory holdingand personnel costs at the supplier increase, and these issues must be addressed in thesupply contract by sharing cost savings of the retailer with the supplier.
24
In general, both of these types of inventory ownership policies require signiÞcant in-formation sharing and large technology investments. Also, note that in these policies theretailer does not have any control on its inventory any more, and trust issues must beresolved completely for this type of partnership to work out.
Question 6
Some of the following steps could have been taken to prevent the failure of the program:
1. In a VMI system, trust issues should be resolved at the very beginning during contractnegotiations. Additionally, communication channels must be open at all times to discussany additional concerns that may arise. Clearly, these issues were not addressed insufficient detail at Spartan.
2. Spartan could have speciÞed a test period for its suppliers, and after the test period theprogram could have been discontinued with unsuccessful suppliers.
3. Spartan could have collaborated with its suppliers to determine agreed-upon forecastsand help them resolve forecasting issues due to promotions.
4. The retailer-supplier partnership could have started with a continuous replenishmentsystem which is less risky than a vendor-managed inventory system. Then, the VMIsystem could have been established based on this experience.
Case Discussion Questions — Audio Duplication Services, Inc.(ADS)
Question 1
ADS�s customers� customers, i.e., the national retailers, want to reduce inventory holdingcosts and expenses related to managing inventory by moving towards VMI agreements.
Question 2
The most signiÞcant impact on ADS�s business will be on its distribution functions be-cause it will have to ship directly to the individual stores instead of distribution centers.Transportation costs will increase due to the increase in the number of destinations. Also,depending on its agreement with the record companies, ADS may incur additional inven-tory holding costs because compact disks and cassettes will spend more time in its system.
ADS should take this situation as an opportunity to re-design and streamline its op-erations and prepare itself for the general trend towards VMI in the industry. This mayhelp ADS to gain a competitive advantage, and potentially increase its market share.
Question 3
Logistics is not a core competency of ADS. Additionally, the management team of ADSmust address more crucial issues such as the emerging media distribution business overthe Internet. Therefore, ADS should outsource its distribution functions to a third-partylogistics provider in order to focus on issues threatening its core business.
25
Question 4
Under the proposed consignment scheme, inventory belongs to the suppliers until theproducts are sold. Under this scheme, it would be inefficient and unreasonable if productswere Þrst delivered to distribution centers that belong to the retailers. This would add anunnecessary complexity to the system as to how goods in the distribution centers wouldbe handled and ultimately delivered to the individual stores. Additionally, by encouragingdirect shipments, retailers can reduce the warehouse space they rent or own, and reducecosts further.
26
Chapter 7
Procurement and OutsourcingStrategies
Discussion Questions
Question 1
In order to assess the effect of product life cycles on the make/buy framework developedin Section 7.3, we need to understand how product knowledge and capacity requirementsevolve over the life cycle of a product. A typical product life cycle has several stages:
1. The product is newly introduced, and bought initially by a relatively small number ofcustomers (early adopters). At this stage of the product life cycle, production level islow and capacity is usually not an issue. Also, quite possibly, there are problems inthe design and manufacturing of the product such that yield is low, and changes in theproduction processes are frequent. Under these conditions, the buy/make frameworkimplies that a new product should not be immediately outsourced.
2. In the rapid adoption phase, demand for the product increases at an increasing rate. Ifthis leads to tight production capacity, and the production processes are already stable,then outsourcing should be considered. It may even be the only choice to keep up withthe increasing demand.
3. When the product reaches the maturity phase, i.e., when demand stabilizes over time,then the outsourcing decision should be re-considered under the framework discussed inSection 7.3. For instance, if earlier an integral product was outsourced due to capacityconstraints, then the Þrm should consider installing additional capacity to meet thestabilized demand, and manufacturing the product in-house.
4. When the product is approaching its end-of-life, outsourcing is a good option for bothintegral and modular products because the company should focus on new products.
Question 2
This question is relevant for modular products for which only a subset of the componentsdrive the customer�s purchasing decision. For instance, consider the PC market. In general,
27
the customers are primarily concerned with the CPU speed, the amount of memory, andthe hard disk capacity when buying a PC. Typically, all other components of a PC, e.g.,the motherboard, the sound card, the CD drive, cables, etc., are of secondary importanceto most of the buyers.
In other words, the value proposition of a modular product depends on a few of itscomponents, and we can expect that manufacturers focus their innovative efforts and newproduct introductions on this small set. Thus, it is crucial that the knowledge for thesecomponents is kept in-house. Under the buy/make framework developed in Section 7.3,this implies that the manufacturer should refrain from outsourcing key components of amodular product.
Question 3
In this case, the risk is high, but also the potential return is high. If the base commitmentlevel is low, the option level is high, and the demand turns out to be higher than what theretailer expects, then the return of the supplier may be substantially more than what itwould have been under a long-term contract. This is due to the differential between theoverall unit price (prepayment plus the exercise price) in an option contract and the unitprice in a long-term contract. In addition, the supplier can pool the risk among severalcustomers.
Question 4
The private and consortia-based e-marketplaces play different roles in IBM�s supply chain.IBM beneÞts from both marketplaces in different ways.
1. In the private exchange, the focus is on better supply chain collaboration. Sensitiveinformation such as product design, customer orders, cost, production plans, etc., areshared between IBM and its suppliers to improve operations, decrease costs and increaserevenues.
2. The objective in the consortia-based e-marketplace is to increase purchasing power byaggregating buyers, to Þnd new and high quality suppliers, and to allow these suppliersachieve cost efficiencies by dealing with standardized systems.
28
Chapter 8
International Issues in SupplyChain Management
Discussion Questions
Question 1
a. Domestic production with international distribution is typical for small businesses thatattempt to expand to serve global markets. For instance, Mavi Jeans manufactures itsproducts in two facilities in Istanbul, Turkey, but sells them in more than 3,000 locationsworldwide. (See Mavi Jeans (2003).)
b. The product(s) may be so complex that the necessary expertise and resources to man-ufacture different components are spread across the world. For instance, Boeing hasmore than 15,000 suppliers in 81 countries. (See The Boeing Company (2003).)
c. Typically, high labor cost in labor-intensive industries causes manufacturing operationsto shift offshore. For instance, consider Nike, which subcontracts its manufacturing toapproximately 350 factories in the Asia-PaciÞc region employing nearly 400,000 workers.(See Nike (2003).)
d. A fully integrated supply chain is crucial for companies that implement ßexible globalsupply chain strategies in a highly variable environment, e.g., for companies that intendto shift production from country to country in order to reduce costs or risk. The PCmanufacturing industry is a good example of this.
Question 2
Although the speciÞc answer depends on the management style, resources and long-termstrategies of the Þrm, a hedge strategy seems more appropriate for a small Þrm. Followinga speculative strategy may be disastrous if the company does not have the size and Þnancialstrength to absorb potential losses associated with a speculative strategy. On the otherhand, following a flexible strategy may require a signiÞcant amount of additional investmentin infrastructure to ensure the necessary coordination, and for the re-design of productsand processes. The resources necessary to implement a ßexible strategy successfully maynot be available to a small Þrm. Clearly, for a large Þrm, the Þnancial and resource
29
constraints may be much less relevant than for a small Þrm, and the global strategy to beemployed may be any of the three strategies mentioned above depending on the speciÞcsof the situation.
Question 3
Some of the factors that affect whether a product can be sold globally without modiÞcationare discussed below:
1. Brand recognition. Some brands have developed a global image of quality and desir-ability that they do not have to be customized to regional and national tastes. This isespecially true for luxury brands, such as Ferrari that sells the same cars globally.
2. Cultural differences. Some products reßect long-term traditions and/or a certain re-gional taste developed over time, and Þt into the regional category. A lot of foodproducts fall here, and it is much harder for these products to get accepted globally.For instance, it is hard to think of kabobs that are very popular in the Middle East tobe sold in global chain restaurants. An exception in this category is pizza, for instance.
3. Standardization. In consumer electronics and computer manufacturing there are plentyof examples of global products. The complexity of such products requires standard-ization throughout the world, and makes them global products. However, note thatexceptions exist. For instance, TV systems in the US and Europe are different, andthe same TV set cannot be sold in both markets. Medical products also Þt into thiscategory as long as they satisfy differences in health regulations in different countries.
Question 4
In Belgium and Canada, the most important issue would be keeping the costs down. Bakeryproduction is labor-intensive and requires relatively unskilled labor. However, unskilledlabor costs in the First World countries are typically very high compared to the rest of theworld.
In the emerging nations, such as Russia, Singapore and Argentina, logistics issues wouldbe most important since the necessary infrastructure may not be in place. On-time deliveryof bakery products is crucial because they are perishable items, and there is a time windowthey need to reach the bakeries in order to be sold. For instance, if the fresh productsarrive at noon, instead of early in the morning, most of these items may be left on theshelves at the end of the day.
In a Third World country, such as Nigeria, in addition to the logistics problems men-tioned above for the emerging nations, environmental conditions may not be sufficient toensure a healthy production of bakery products. For instance, water may not be clean,etc.
Question 5
a. Possible reasons for Wal-Mart to start international operations include:
1. Domestic revenues do not increase at a pace Wal-Mart wants them to. Therefore,Wal-Mart intends to achieve a high growth rate in foreign countries to hedge againstthe slow business in the domestic markets.
30
2. A rival European chain, Carrefour SA, has already a strong presence in the SouthAmerican market. Wal-Mart could lose the market totally if it did not respondquickly.
3. Foreign operations would also help Wal-Mart identify potential new low-cost suppliersfor domestic operations.
b. Having many suppliers in different countries would enhance Wal-Mart�s �everyday lowpricing� formula by allowing it to hedge against cost increases due to economic and/orpolitical conditions in a given country.
c. Strong centralized control is helpful for aligning local operations with the overall cor-porate objectives. Additionally, Wal-Mart�s years of domestic expertise in retail opera-tions, e.g., distribution systems, can best be leveraged to improve distribution in SouthAmerica if strong centralized control is implemented. On the other hand, it may beimpossible to analyze customer preferences and understand cultural differences unlesslocal managers control certain aspects of the business.
d. In general, political instability is a major concern in South American countries. Forinstance, due to the recent downfall of the Argentine economy, Wal-Mart may experiencedifficulties if the Argentine government starts to employ policies that favor domesticproducers rather than foreign companies. Also, if the dollar continues to gain in valueagainst major South American currencies such as the Argentine Peso and BrazilianReal, American exports into these countries will be expensive compared to their localcompetitors, which will negatively impact Wal-Mart�s margins.
31
Chapter 9
Coordinated Product and SupplyChain Design
Note: The solutions for this chapter are incomplete at this time. For morecomplete solutions, please see the textbook web site.
Discussion Questions
Question 1
1. IKEA Furniture
2. ßoor lamps
3. Tupperware made by Rubbermaid
4. Collapsible chairs
5. Fedex package box
Question 2
1. Offering more products, models, and options requires more customization and morecomplicated production scheduling.
2. Product proliferation makes it difficult for manufacturers to take advantage of economyof scale, because products are processed in small batches.
3. Offering more products adds difficulty to inventory tracing.
4. Non-aggregate demand usually has higher variability than aggregate demand.
5. Great demand variability makes forecast inaccurate.
6. When demand variability is high, a Þrm needs to prepare more stock to maintain thesame service level than when demand variability is little.
32
Question 4
Standardization must be a positive force to eliminate the cost and inconvenience causedthe possible incompatibility between products offered by different vendors, and still allowcompetition and innovation. Excessive standardization could reduce design options andstiße the development of a product or industry.
Microsoft windows is an example of excessive standardization. The large populationof MS windows users and the wide acceptance of this operating system among varioussoftware vendors make many MS window technologies industrial standards. Microsofttakes advantage of this to block competition and manipulate market rather than improvethe performance of its products.
Case Discussion Questions — Hewlett-Packard Company:NetworkPrinter Design for Universality
33
Chapter 10
Customer Value and Supply ChainManagement
Discussion Questions
Question 1
In traditional retailing, quality is frequently the primary determinant of price: the higherthe quality of the physical product, the higher is the price. However, in on-line retailing,Þrms are able to leverage on value propositions other than product quality, and thus gainmore ßexibility in pricing. For instance, the ability to browse and search an on-line catalogor the availability of an order tracking system may lure the customer to a site which hashigher prices for the same product or to a site that offers lower quality products for thesame price. In other words, in on-line retailing the price versus quality trade-off curve isaffected by other factors.
Question 2
a. As the available capacity decreases with respect to the average demand, it may beimpossible to satisfy the realized demand. Therefore, by changing prices to affect thedemand distribution, the Þrm would be able to increase revenues. For instance, increas-ing price may decrease the overall demand, but if the realized demand for the companyis still more than the available capacity, the Þrm is better off by increasing price.
b. As the standard deviation of demand increases with respect to its mean, it becomesmore likely that the realized demand over time is not close to the demand estimateassociated with a Þxed-price policy. Therefore, adjusting prices over time increasesproÞts when demand variability is high.
c. Demand is not only a function of price. Seasonality and macroeconomic conditions,e.g., changes in the size of the whole market, affect the demand distribution that anindividual Þrm faces. Clearly, as the demand distribution changes over time, the optimalprice changes as well.
34
Question 3
a. This dimension of customer value refers to the ability to match supply and demand. Asdiscussed in Section 10.2.1., if the supply chain is not designed by taking into accountthe speciÞc characteristics of the market, then the consequences can be severe. Forinstance, if the demand is highly variable, then the focus of the supply chain shouldbe on ßexibility and reducing lead times such that the product can be delivered tothe market before the conditions change. Otherwise, customer requirements cannot besatisÞed, sales are lost, and obsolete inventories accumulate.
b. The decision where to draw the push-pull boundary in the supply chain has a strongimpact on what product portfolio can be offered. For instance, if Þnished products aremade-to-stock and kept in a central inventory until a customer order arrives, then itmay not be possible to meet the demand for all types of product conÞgurations. Onother hand, if the push-pull boundary is at the assembly operation, as it is in the Dellcase, it is possible to offer a greater variety of products.
c. An important component of price is cost. Supply chain decisions determine production,inventory and transportation costs, and affect the ability of a company to offer lowprices. Furthermore, supply chain decisions impact the brand recognition of a product.Note that it would be very hard to build a prestigious brand name if the supply chainproduces defective products, does not meet delivery due dates, and does not providesupport for its products after the sale.
d. Consider, for example, service and maintenance support. Clearly, in order to performthese activities, the supply chain must satisfy certain requirements. For instance, ifon-site support is provided, then it may be necessary to keep component inventoriesclose to the major customer locations in addition to the necessary personnel to performthese activities. Also, if a piece of hardware needs to be sent back to the factory to beÞxed, then reverse logistics operations must be planned, e.g., pick-up from the customersite, transportation back to the factory, etc.
e. In order to ensure that the learning process enhances customer value, it is critical toinstall the necessary information loops in the supply chain. For instance, if databasesare used to track customer preferences, then this data must be fed back to the productdesign process, and eventually a superior product must reach the customer. Clearly,this may require additional investment in information technology infrastructure.
Also, consider the Dell example, in which Dell manages the entire purchasing process oflarge customers including special custom features. In this case, the supply chain musthave enough ßexibility in production and distribution to accommodate customer-speciÞcdemands.
Question 4
a. Starbucks emphasizes customer access along with experience by creating a pleasantambiance in its stores.
b. The Gap creates customer value by offering affordable products along with a brandimage built around stylish designs.
35
c. Expedia.com brings customer value by offering a large selection of products, i.e., consoli-dating information about different travel rates from many companies for easier shopping,and enabling one-stop shopping for ßights, hotels and cars.
Question 5
Examples of experiences enabled by the Internet include:
1. Newsletters sent regularly by e-mail.
2. On-line customer service.
3. On-line birthday cards sent based on birthday data collected from customers.
4. User groups created by companies that bring together people that share the same en-thusiasm and interest about the company�s products.
Question 6
1. Possible performance measures for Amazon.com include:
i. The average interval between re-visits (of the same customer) to the site.
ii. Average order size in dollars, average number of items in an order.
iii. Growth rate of hits on the site.
iv. Links to the site from other web sites.
v. Classical performance measures, such as sales, total revenue, costs and proÞts,earnings per share, etc.
2. Possible performance measures for the supply chain include:
i. On-time delivery, lead time, Þll rate.
ii. Return rate.
iii. Value added per employee.
iv. Inventory turnover.
36
Chapter 11
Information Technology for SupplyChain Management
Discussion Questions
Question 1
The ERP industry is fragmented, and there are many ERP vendors (see TechnologyE-valuation.Com (2002)); a number of them are also mentioned in this chapter: Baan Co.,Oracle, QAD Inc., System Software Associates, PeopleSoft, SAP, etc. The reason for thisfragmentation seems to be the lack of standardization in the large portfolio of businessfunctions, e.g., human resources, Þnance, purchasing, production, that these ERP systemssupport. All companies have some subset of these functions, and they all have differentneeds which lead to different software requirements and a number of ERP vendors.
Question 2
In addition to enabling access to data that currently reside in closed systems, the Webservice technology would ultimately remove the barriers to adopting a �best-of-breed� ap-proach to supply chain management by alleviating the challenging application integrationand version upgrade issues.
For instance, a 3PL provider that handles inbound logistics of a company may installa tracking system for its trucks that can be integrated to its customer�s systems over theWeb service technology. This would provide better information to the vendor when toanticipate deliveries in order to improve production schedules.
Question 3
The most important advantage of buying a Decision Support System (DSS) from an Ap-plication Service Provider (ASP) is that it reduces initial implementation costs, both forhardware and services. The company does not have to buy the IT infrastructure that isneeded to support the DSS, and many issues related to the setup of the DSS system aretaken care of automatically by the ASP. Thus, in general, we would expect that the initialimplementation project would take less time and effort compared to an on-site DSS imple-
37
mentation. Additionally, no maintenance costs related to version upgrades are incurred.(See the Starbucks case at the beginning of the chapter.)
On the other hand, there are many issues that need to be addressed before a companydecides to run a DSS in this mode. First, sensitive data, such as production costs, need tobe transferred to a remote site over which the company has no control. Most companieswould be very reluctant to agree to this due to security concerns, especially if the same ASPwould be hosting the DSS for a competitor at the same time. Second, there may be seriousintegration issues involving the DSS and the internal systems of the company, e.g., the ERPsoftware that feeds the input data into the DSS. In other words, cross-enterprise integrationcould be substantially more challenging than intra-enterprise integration. Third, a DSS istypically a complex system, and changes to modeling are certain to occur in a dynamicbusiness environment. In the ASP mode, it is much harder for the company to develop skillsand expertise to make these changes happen, and the company will always be dependenton the ASP. Note that this becomes a very critical issue if the ASP is not a big and stablecompany and eventually goes out of business.
Question 4
The purpose of the second generation of ERP systems (ERP II) is to achieve integrationacross the supply chain, and the Internet technology is a major tool for the ERP vendorsto achieve this objective. Three of the four goals of the supply chain information technol-ogy, i.e., collection and access of information and collaboration, can be facilitated greatlythrough standardization by the use of Internet. (See the topics “Communications” and“Collaboration” in Section 11.3.) Along these lines, we are beginning to see the architectureof ERP systems evolve toward a modular and Web-based environment.
Question 5
a. Network design has the longest planning horizon (up to a few years) with a strategicfocus on how to optimize the overall supply chain. Supply chain master planning isa tactical decision with the objective of coordinating different functions in the supplychain, and the planning horizon ranges typically from a week to a few weeks. Operationalplanning systems have short planning horizons and the focus is on a single system,e.g., production planning. The objective is to generate feasible strategies under theconstraints imposed by the decisions made previously in strategic network design andsupply chain master planning. In operational execution systems, the focus is on real-time execution.
b. The execution systems use very detailed data about the state of the supply chain, andthe level of data aggregation increases as you move from execution to strategic networkdesign. For instance, strategic network design problems aggregate demand typically onan annual basis.
c. Complexity and duration of implementation increases as you move from strategic decision-making toward execution due to increasing level of detail.
d. Only a few people need to be involved in a strategic network design project, but thenumber of users increases dramatically toward execution.
38
Chapter 12
Decision-Support Systems forSupply Chain Management
Discussion Questions
Question 1
Some of the factors that need to be considered when selecting a DSS for supply chainmaster planning are:
1. Analytical capabilities.
i. Consider whether the algorithms are heuristics or provide optimal solutions. If theyare heuristics, is it possible to quantify the quality of the heuristic solution?
ii. Scalability. How is the performance affected by the length of the planning horizon,number of products, etc.?
iii. Robustness and sensitivity analysis. Do the master plans change signiÞcantly withsmall changes in demand forecasts? Is the DSS able to perform sensitivity analysis?
iv. Can multiple solutions be generated easily?
2. Data requirements. Is there enough data of sufficient quality to support the needs of theplanning system? Does the planning tool integrate easily into existing ERP software orother pre-existing systems?
3. User interface. Is the DSS easy to use? Are there import/export capabilities based oncommon PC formats?
4. Output analysis. Is the output presented in a format that is easy to understand andinterpret? Can the results be exported into common PC formats?
5. Integration with operational planning systems. Is the integration possible from an ITperspective? Is it possible to generate feasible solutions in the operational planningsystems based on the output of the supply chain master planning tool?
6. Collaboration. Does the DSS facilitate collaboration with other parties involved in thesupply chain, e.g., suppliers?
39
7. Hardware and software requirements.
8. Price.
Question 2
For instance, if a DSS is used by many users spread across different geographic locations, itis important that the DSS has a Web-based client/server technology in order to facilitatecollaboration. Clearly, a similar argument can be made if the input data reside on manydifferent systems across different locations. However, the Web technology may not beimportant for a strategic network design DSS that is used every few years by a few users.
Question 3
Planning does not add much value unless the plan can be executed properly. Therefore,the interface between supply chain planning tools and execution systems is critical. Ofcourse, this is true from an IT perspective, but it is also crucial from an analytical point ofview. For instance, consider an execution system to which a large order from an importantcustomer arrives, and assume that this is a rush order. In this case, a relatively short leadtime is quoted to the customer, and production and delivery schedules must be updatedby disrupting the schedules of the rest of the orders minimally. In other words, the DSSis required to propagate the effect of this rush order back to operational planning, and tosupply chain master planning if necessary.
Question 4
a. By benchmarking, a company can assess its performance against its competitors alongcommon dimensions in the speciÞc industry. This helps the company to determine itsstrengths and weaknesses, and prioritize the areas in which improvements are needed.
b. For instance, benchmarking can provide insight into what the customers value the most.In other words, it may be possible to attribute differences in sales, revenue and cus-tomer service among different companies to the speciÞc performance measures based onbenchmarking results.
c. We would expect benchmarking to have an immediate impact on the performance at thefront-end of the supply chain, e.g., order fulÞllment, customer service, etc., because thisis highly visible to the customer, and affects sales and marketing signiÞcantly. Of course,the front-end must be supported by the back-end, and the effect of benchmarking wouldeventually propagate back toward the supply chain planning level.
40
Bibliography
Bramel, J. and Simchi-Levi, D. (1997). The logic of logistics : theory, algorithms, andapplications for logistics management, chapter 1. Springer, New York.
Mavi Jeans (2003). Mavi.com.http://www.mavi.com/story.asp?s=4.
Nike (2003). Nike Biz::About::Company Overview:: The Facts.http://www.nike.com/nikebiz/nikebiz.jhtml?page=3&item=facts.
TechnologyEvaluation.Com (2002). ERP Evaluation Center - Technology Evaluation.Com.http://www.erpevaluation.com/index.asp.
The Boeing Company (2003). Quick Facts.http://www.boeing.com/companyoffices/aboutus/quickfacts.html.
41