16915049 Barilla Spa Research PDF

Preview

Summary

Assignments...

Description

BARILLA SpA (A), (C) and (D)

1. Draw a flow diagram for Barilla dry products sold through distributors. Your diagram will have four major sections: Barilla production, Barilla Finished Goods Inventory, Distribution Centers, and grocers. Indicate the flows of both physical goods and information. At appropriate places on your diagram, collect any relevant statistics mentioned in the case, for example on inventories, lead-times, etc.

Please refer to the following page (page 2 of this document) for the process flow diagram for Barilla dry products sold through distributors.

INSERT PROCESS FLOW DIAGRAM

2.

What are the problems and difficulties facing the supply chain? What is their impact? Diagnose their underlying causes.

The supply chain of pasta and pasta-related products in Italy was facing significant problems, including excessive inventory levels, high levels of stock-outs, extreme reliance on promotional activities, inefficient logistics operations and poor communication and cooperation between the supply chain participants. The industry had not leveraged technology to improve communication amongst its members, tracking of inventory or demand forecasting. These problems resulted in higher inventory carrying costs, as supply chain members sought to reduce the frequency and magnitude of stock-outs. Furthermore, industry suppliers competed fiercely on price, resulting in extreme peaks in demand during promotional periods.

The core cause of these problems was unpredictable demand, which resulted in higher inventory levels and carrying costs, in addition to making logistics operations more difficult. Because there was limited demand visibility within the supply chain, any significant fluctuation in demand potentially resulted in stock-outs. Poor communication and cooperation amongst the supply chain participants also contributed to the limited visibility of demand within the supply chain. Because the supply chain participants were not actively sharing information with other members of the supply chain, the participants were unable to react quickly to peaks in demand. As a result, higher levels of safety stock were maintained by all parties in the supply chain to avoid potential stock-outs. Barilla suffered from this

demand variability to an even greater extent than the grocery stores or distributors due to its position in the supply chain and the associated bull-whip effect.

In addition to poor communication, the extensive use of promotions and discounts also contributed to the variability of demand. The grocery and pasta business in Italy had a long history of using promotions and discount pricing to move products. The central distributors expected frequent trade promotions and stocked up on inventory during the promotional periods to benefit from the favorable pricing. As a result of promotions, the demand for Barilla¡¯s products would swing significantly from week to week and was not stable throughout the year. Demand peaked when promotions were offered and dropped significantly when the promotions expired. As a result, higher inventory levels were maintained both by Barilla and the distributors as the distributors purchased sufficient pasta to meet both current and future needs.

An additional cause of the high inventory levels was the manufacturing process. Barilla carried approximately 800 SKUs of dry products, and many of these products were produced in precise conditions (temperature and humidity) to ensure the quality of the product. The manufacturing process used by Barilla focused on producing large quantities of individual products since adjusting the manufacturing conditions to suite a new product was a time-consuming process. As a result, Barilla could not react quickly to changes or peaks in demand as its manufacturing process faced long lead times. This inability to produce a balanced combination of pasta products was illustrated by the fact that the Barilla¡¯s plants were specialized by the type of pasta produced in the plant. This manufacturing system can be contrasted against ¡°Heijunka¡± which provides for a balanced production schedule. Because Barilla¡¯s manufacturing operations could not switch quickly between products, Barilla was forced to maintain overall higher inventory. The long set up times and lack of manufacturing flexibility resulted in higher cycle times to produce the product.

3.

How does the JITD program hope to correct these problems?

The core problem facing Barilla is the high degree of demand variance that it experiences by product and throughout the year. The main causes of the demand variance are limited visibility downstream, and reliance on promotions to push product downstream.

In order to increase visibility downstream, Barilla has developed a program that pulls relevant shipping information from their downstream partners. This allows Barilla to optimize production and transportation decisions given the information. Barilla¡¯s Just-In-Time-Distribution program (JITD) effectively looks at the daily shipping decisions made by distributors and warehouses, combines that information with each distributor¡¯s stock position by SKU and plans future production and replenishment decisions for each distributor.

In order to reduce Barilla¡¯s reliance on promotions to push product downstream, Barilla plans to reduce the frequency and magnitude of its promotions. Barilla will demonstrate to its customers that the gains (reduced inventory carrying cost, reduced stock outs, reduced damages etc¡) experienced by distributors/warehouses using the JITD program outweigh any changes in the average per unit price that a distributor pays. Due to the benefits that Barilla extracts from the JITD program, Barilla may be able to reduce the average price of their product or offer an incentive to adopt the JITD program.

By adopting the JITD program each step in the supply chain is able to extract benefits that result in reduced costs, improved customer service and can increase and stabilize revenue. After Barilla¡¯s JITD program is up and running a company can expect to fulfill a higher percentage of its orders from stock thus reducing its stock out rate and increasing its revenue. In addition, a company can expect to reduce its inventory that it carries throughout its system. In particular, a company can reduce the amount of safety stock that it carries throughout the year to adjust to demand/delivery fluctuations. Barilla also expects to extract benefits from being able to better plan and organize its production functions around actual product performance. This should also allow Barilla to reduce its costs associated with set-up costs. Barilla is also able to maintain flatter prices throughout the year by reducing its promotions. This should allow Barilla to better estimate its revenue stream and should avoid the wide fluctuations associated with the promotion activity. Lastly, by having Barilla¡¯s customers integrate more closely into Barilla¡¯s supply chain, Barilla has effectively raised its customer¡¯s potential costs associated with changing suppliers and raised the bar that competitors must clear to gain a foothold in Barilla¡¯s business.

However, we need to remember that application of the JITD program to Barilla¡¯s distribution network has several drawbacks (perceived and real) that might slow the expansion/implementation of the program. For example, based on previous trials we believe that the JITD must be implemented in one large event as opposed to a piecemeal approach, which might be preferred by Barilla¡¯s customers. In addition, the JITD program requires a substantial investment in resources to adopt the technology that allows warehouses and distributors to communicate with Barilla each night. One of the necessary components for a successful JITD program is the elimination or severe curtailing of the promotions that seem to be associated with the grocery-industry. This culture shift will require considerable effort to effect the positive change that will allow a JITD program to succeed. In addition, the advocates that need to be selling this program to Barilla¡¯s customers are unhappy with the program because they believe (correctly) that after implementing this type of program they earn less money. One of the final and most important hurdles for Barilla to handle is the requirement of the free flow of information from the distributor to Barilla. This is the most essential part of the program and represents a major departure from current practices.

4. As one of Barilla¡¯s customers, what would your response be to JITD? If you were Barilla¡¯s management, how would you convince this customer to implement JITD?

As a retailer, my initial reaction would be why should I share my sales data upstream to a supplier who also supplies its products to my competitors? Also, I would ask why Barilla believes that it can create better demand forecasts than myself, who is closest to the customer. As a distributor, I would be irritated at the implication that stockouts are a result of my ability (or in this case, inability) to forecast well. Also, I would suspect that Barilla was just trying to reduce its own inventory costs by pushing its products into my warehouses.

JITD is a hard sell because partial implementation (piecemeal approach ¨C only a few firms) is difficult due to the potential negative consequences in the short run (because non-JITD firms can exploit the no promotion constraints of JITD) and the benefits being somewhat longer term in nature. Also, elements of the JITD program are contrary to the existing mentality of Italian pasta distributors and retailers, especially in terms of promotions, discounts, order-placing, and inventory. Barilla also needs to convince non-technology-savvy individuals (retailers and distributors) to adopt this program while also reassuring distributors that it is not the intent of Barilla to bypass them in the supply chain. Barilla needs to establish trust between itself and the retailer/distributor. It needs to prove that the implementation of JITD will lessen the current problems facing the supply chain that have resulted in high costs of carrying inventory and lost sales, which are shared by all members (Barilla, distributor, and retailer) of the supply chain.

The customer has to be convinced that if the upstream site or Barilla has access to demand and inventory information at the downstream site, overall costs would decrease because Barilla is in the best position to create the demand forecast for the entire chain. For JITD to work, stable prices (i.e. no promotions or discounts) and a break in order batches (resulting in more frequent visits possibly with composite distribution of fresh and dry products) must exist. In convincing customers to join, I would first present my ideas that support JITD with graphical examples and lessons similar to the Beer Game. For example, Barilla could do a PV calculation of the reduced costs over a period of time due to lower inventory and reduced lost sales and compare it to the ¡°savings¡± achieved with promotions and discounts, two elements that are not consistent of JITD.

The next step would be to provide ¡°real proof¡± that the JITD concept works. Although we considered a pilot with smaller independent shops, we realized that this may not convince the larger supermarket chains that JITD will reduce overall costs. Rather, we chose to try to convince supermarket chain customers to try the JITD concept on a select number of products. Barilla would provide the technology, training and ¡°extra¡± work on the customer end. Over several months, these firms would not participate in the promotions of those particular products. Although firms may be hesitant to accept this, Barilla could offer to promise other compensation if those ¡°savings¡± from the promotions were not recovered by the reduced inventory carrying costs and lost sales of those particular products (money-back guarantee).

Barilla could offer additional incentives, such as a general discount, to convince retailers to share information but this discount would have to be spread across the entire year or in such a manner not to cause additional variability.

JungleBooks.com

1. First consider a decentralized analysis. Estimate the optional parameters for a (Q,R) inventory policy for each of the 4 SKU types detailed in the case for each proposed DC. Estimate the total annual inventory holding plus ordering costs for each of the 4 item types for each DC.

Decentralized Analysis:

Soft Cover I

(Q, R) Analysis

c = $2.00 i = 0.2 h = ic = $0.4/ unit-yr L = 2 weeks Weekly demand is normal with mean 25 and and variance 52 = 25 s = $50.00

Therefore, Q* = [2(25)(52)(50)/0.4]1/2 = 570

Cs = $6.00 - $2.00 = $4.00 Ce = h(Q/¦Ë) = .4[570/(25)(52)] = 0.18

Hence, SL* = 4/4.18 = 0.958

Z0.958 = 1.73 from the normal distribution table.

Hence, R*

= L¦Ë + ZSL¦ÒL1/2 = 2(25) + 1.73(5)(2)1/2 = 50 + 12.23 ¡Ö 62

Cost Analysis

INV = Q*/2 +SS = 570/2 + 12.23 = 297 Annual Inventory cost = 297($.4) = $119 Annual ordering cost = (25)(52)(50)/570 = $114

Total annual cost per DC = $119 + $114 = $233

For all DCs, total annual cost = 6($233) = $1,398 per title.

Soft Cover II

(Q, R) Analysis

c = $2.50 i = 0.2 h = ic = $0.5/ unit-yr L = 3 weeks Weekly demand is normal with mean 50 and variance 502 = 2,500 s = $50.00

Therefore, Q* = [2(50)(52)(50)/0.5]1/2 = 721

Cs = $6.00 - $2.00 = $4.00 Ce = h(Q/¦Ë) = .5[721/(50)(52)] = 0.14

Hence, SL* = 4/4.14 = 0.966

Z0.966 = 1.83 from the normal distribution table.

Hence, R*

= L¦Ë + ZSL¦ÒL1/2 = 3(50) + 1.83(50)(3)1/2 = 150 + 158.48 ¡Ö 308

Cost Analysis

INV = Q*/2 +SS = 721/2 + 158.48= 519 Annual Inventory cost = 519($.5) = $260 Annual ordering cost = (50)(52)(50)/721 = $180

Total annual cost per DC = $260 + $180 = $440

For all DCs, total annual cost = 6($440) = $2,640 per title.

Hard Cover I

(Q, R) Analysis

c = $10.00 i = 0.2 h = ic = $2.00/ unit-yr L = 3 weeks Weekly demand is normal with mean 100 and variance 202 = 400 s = $50.00

Therefore, Q* = [2(100)(52)(50)/2.00]1/2 = 510

Cs = $11.00 - $4.00 = $7.00 Ce = h(Q/¦Ë) = 2[510/(100)(52)] = $0.20

Hence, SL* = 7/7.20 = 0.972

Z0.972 = 1.915 from the normal distribution table.

Hence, R*

= L¦Ë + ZSL¦ÒL1/2 = 3(100) + 1.915(20)(3)1/2 = 300 + 66.16 ¡Ö 366

Cost Analysis

INV = Q*/2 +SS = 510/2 + 66.16= 321 Annual Inventory cost = 321($2.00) = $642

Annual ordering cost = (100)(52)(50)/510 = $510

Total annual cost per DC = $642 + $510 = $1152

For all DCs, total annual cost = 6($1152) = $6,912 per title.

Hard Cover II

(Q, R) Analysis

c = $10.00 i = 0.2 h = ic = $2.00/ unit-yr L = 2 weeks Weekly demand is normal with mean 10 and variance 122 = 144 s = $50.00

Therefore, Q* = [2(10)(52)(50)/2.00]1/2 = 161

Cs = $11.00 - $4.00 = $7.00 Ce = h(Q/¦Ë) = 2[161/(10)(52)] = $0.62

Hence, SL* = 7/7.62 = 0.919

Z0.919 = 1.40 from the normal distribution table.

Hence, R*

= L¦Ë + ZSL¦ÒL1/2

= 2(10) + 1.40 (12)(2)1/2 = 20 + 23.76 ¡Ö 44

Cost Analysis

INV = Q*/2 +SS = 161/2 + 23.76 = 104 Annual Inventory cost = 104($2.00) = $208 Annual ordering cost = (10)(52)(50)/161 = $161.50

Total annual cost per DC = $208 + $161.50 = $369.50

For all DCs, total annual cost = 6($369.50) = $2,217 per title.

2. Now consider a centralized analysis (their current system), using the same service level derived for the decentralized system. Estimate the optional parameters for a (Q,R) inventory policy for each of the 4 SKU types detailed in the case for a centralized DC. Estimate the total annual inventory holding plus ordering costs for each of the 4 item types for the centralized DC.

Centralized Analysis:

Soft Cover I

(Q, R) Analysis

Centralized demand is normally distributed with mean 6(25) = 150 and variance 6(52) = 150 = 12.252

Therefore, Q* = [2(150)(52)(50)/0.4]1/2 = 1,396

R*

= L¦Ë + ZSL¦ÒL1/2 = 2(150) + 1.73(12.25(2)1/2 = 300 + 30 = 330

Cost Analysis

INV = Q*/2 +SS = 1,396/2 + 30 = 728 Annual Inventory cost = 728($.4) = $291 Annual ordering cost = (150)(52)(50)/1,396 = $279

Total annual cost = $291 + $279 = $570 per title

Soft Cover II

(Q, R) Analysis

Centralized demand is normally distributed with mean 6(50) = 300 and variance 6(502) = 15000 = 122.472

Therefore, Q* = [2(300)(52)(50)/0.5]1/2 = 1,766

R*

= L¦Ë + ZSL¦ÒL1/2 = 3(300) + 1.83(122.47(3)1/2 = 900 + 388 = 1,288

Cost Analysis

INV = Q*/2 +SS = 1,766/2 + 388 = 1271 Annual Inventory cost = 1271($.5) = $636 Annual ordering cost = (300)(52)(50)/1,766 = $442

Total annual cost = $636 + $442 = $1,078 per title

Hard Cover I

(Q, R) Analysis

Centralized demand is normally distributed with mean 6(100) = 600 and variance 6(202) = 2400 = 48.992

Therefore, Q* = [2(600)(52)(50)/2]1/2 = 1,249

R*

= L¦Ë + ZSL¦ÒL1/2 = 3(600) + 1.91(148.99(3)1/2 = 1800 + 162 = 1,962

Cost Analysis

INV = Q*/2 +SS = 1,249/2 + 162 = 787 Annual Inventory cost = 787($2.00) = $1,574 Annual ordering cost = (600)(52)(50)/1,249 = $1,249

Total annual cost = $1,574 + $1,249 = $2,823 per title

Hard Cover II

(Q, R) Analysis

Centralized demand is normally distributed with mean 6(10) = 60 and variance 6(122) = 864 = 29.392

Therefore, Q* = [2(60)(52)(50)/2]1/2 = 395

R*

= L¦Ë + ZSL¦ÒL1/2 = 2(60) + 1.40(29.39(2)1/2 = 120 + 58 = 178

Cost Analysis

INV = Q*/2 +SS = 395/2 + 58 = 256 Annual Inventory cost = 256($2.00) = $512 Annual ordering cost = (60)(52)(50)/395 = $395

Total annual cost = $512 + $395 = $907 per title

3.

Estimate the difference in annual expenses between the centralized and decentralized system over the 150,000 titles classified in the table.

Difference in Annual Expense:

Soft Cover I

Number of titles = 50,000 Total annual cost in decentralized system = 50,000($1,398) = $69,900,000 Total annual cost in centralized system = 50,000($570) = $28,500,000 Hence, cost difference = $69,900,000 - $28,500,000 = $41,400,000

Soft Cover II

Number of titles = 70,000 Total annual cost in decentralized system = 70,000($2,640) = $184,800,000 Total annual cost in centralized system = 70,000($1,078) = $75,460,000 Hence, cost difference = $184,800,000 - $75,460,000 = $109,340,000

Hard Cover I

Number of titles = 10,000 Total annual cost in decentralized system = 10,000($6,912) = $69,120,000 Total annual cost in centralized system = 10,000($2,823) = $28,230,000 Hence, cost difference = $69,120,000 - $28,230,000 = $40,890,000

Hard Cover II

Number of titles = 20,000 Total annual cost in decentralized system = 20,000($2,217) = $44,340,000 Total annual cost in centralized system = 20,000($907) = $18,140,000

Hence, cost difference = $22,140,000 - $18,140,000 = $26,200,000

By decentralizing additional cost incurred will be $41,400,000 + $109,340,000 + $40,890,000 + $26,200,000 = $217,830,000

4.

In what ways is this analysis approximate? In what ways...