Strategic Partnership Case: Barilla SpA

OPERATNS 476

Distribution and Supply Chain Management 2007

Jeannette Song

Session 1B
1

Learning Objectives
A supply chain example:
Supply chain composition:
products, players, facilities technology value added activities

Supply chain dynamics:

Product life-cycle, demand pattern production/distribution leadtimes replenishment cycles

Bullwhip effect: causes Incentive conflicts and resolutions

Strategic partnership and implementation

Quick review of base-stock model

2

The Barilla Supply Chain

Plant

CDCs

Grande Distribuzione

Chain supermarkets

Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

- Need to study physical flow of goods i.e. distribution processes - Need to study information flows i.e. flow of orders - Need to study coordination of various stages / parties
1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

The Physical Production-Distribution Process

Plant CDCs Grande Distribuzione Chain supermarkets Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

- 6 Pasta Plants, 25 across product categories size of pasta shells determined production allocation to plants -Fully automated, 120 meter long production lines, continuous flow manufacturing -Cyclical production (i.e. produce one product type for a while, and then changeover to another product type
1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

The Physical Production-Distribution Process

Plant CDCs Grande Distribuzione Chain supermarkets Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

- Barilla dry products 800 SKUs -Pasta 470 packaged SKUs -CDCs held a months worth of dry goods, 3 days worth of fresh products inventory -Cortese DO stocked 100 Barilla SKUs and overall carried 5000 SKUs

1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

The Physical Production-Distribution Process

Plant CDCs Grande Distribuzione Chain supermarkets Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

- GDs and DOs acted independently; purchased, and maintained their own inventory -Distributors warehouse held 2 weeks worth of supply -Supermarkets typically placed orders with DOs daily; deliveries took place in 24-48 hours

1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

The Physical Production-Distribution Process

Plant CDCs Grande Distribuzione Chain supermarkets Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

- GDs and DOs acted independently; purchased, and maintained their own inventory -Distributors warehouse held 2 weeks worth of supply -Supermarkets typically placed orders with DOs daily; deliveries took place in 24-48 hours -DOs and GDs placed weekly orders; lead time for replenishment was 8-14 days
1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

Information Flows in Barillas Supply Chain

Plant CDCs Grande Distribuzione Chain supermarkets Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

- Trade Promotions & Incentives signals from Barilla to distributors to purchase more -Volume discounts; 2 to 3% for orders in FTL quantities; 4% for 3FTL -Advertising campaigns to stimulate end-customer demand -DOs channel end-customer demand information to Barilla!

1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

Incentive Conflicts in Barilla Supply Chain

Plant CDCs Grande Distribuzione Chain supermarkets Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

Variability in Demand / Orders

- Trade promotions give incentives to batch ordering -Other promotions lead to spikes in demand, and in general erratic ordering patterns

1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

Key characteristics of current operations

Variability in demand, specifically in the order quantities, propagates (bullwhip effect) as we move upstream in the supply chain. Exhibits 12 and 13 show:
-Ordering patterns are extremely volatile (i.e. with spikes followed by periods of small orders -Inventory levels are high even as stock-outs occur!

Lead-time variability makes buyers nervous; shortage gaming Stockouts lead to larger batches; in turn erratic order patterns worsen lead time variability! Vicious cycle
10

Exhibit 12 Weekly Demand for Barilla Dry Products from Corteses Northeast Distribution Center to the Pedrignano CDC, 1989.

11

Sources of Demand Fluctuations

Distributed Inventories Promotions (4 to 5 weeks) Full truck load / less than truck load Poor communication Process (production) inefficiencies and variability No inventory information
12

Sources of Demand Fluctuations (cont.)

Lead time and lead time variability Too many products (SKUs) No forecasting Lack of sophistication of distributors Sales force incentives Quantity Discounts Local vs. global optimization Inflated orders No miminum or maximum order quantities
13

Costs attributable to demand fluctuations

Inventory costs Order costs Lost customers (stock-outs and their fallout) Overcapacity Increased space requirements More trucks (increased transportation requirements) Quality issues (for example, low quality batches produced earlier, that are not discovered before time of consumption) Overtime Perishability
14

Review: The Base-Stock Model

Warehouse Retailer

We are going to consider situations in which inventory is managed over a long time horizon and demand is stochastic Orders are received L time units after they are placed (lead time) Unsatisfied demand is backlogged
15

Inventory Position
What information do you need when deciding how much to order?

Inv. Position = Net Inventory1 + Inventory On Order

Inv. On Hand - Backorders

1. Also referred to as Inventory Level

16

The Base-Stock Policy

A base stock is a target level for the firms inventory position Orders are placed only when demand lowers the firms inventory below the base stock level Any incoming demand triggers an order, so this policy is generally appropriate when ordering costs are negligible
17

Demand Distribution
Assume demand is continuous Demand is Normally distributed with mean = standard deviation = Lead time demand is also Normally distributed with mean = L standard deviation = L
18

Average Cost
fill rate = 1 A stock-out frequency = A average backorder level = B average amount of inventory = I average cost rate per unit of time = C C=hI+pB
holding cost rate penalty cost rate
19

Base Stock Level

Base stock level = s

s = L + z
average lead time demand

L
safety stock

20

10

Base Stock Level (Cont.)

To compute the base stock level that minimizes average cost (C = h I + p B), find z* such that:

( z*) =

p p+h

When p is unknown, but a high fill rate or service level is set, then choose z* such that:
predetermined stock out

A = 1 ( z*)

21

Standard Loss Function

We will use the standard loss function

G ( x ) = x (1 ( x )) + ( x )
standard normal distribution standard normal density

22

11

Base Stock Model

The corresponding base stock level s* is found from:

s * = L + z *

And the average inventory per unit of time is:

I * = ( z * + G ( z * )) L

A = 1 ( z* )
23

Example
Tool-kits at a hardware store
Cost/unit $40 Holding cost rate 22.5% per year Mean Demand: 10/week Variance of Demand: 25/week Leadtime: 1 week Backorder cost: $15/toolkit-week Holding cost: 40*(0.225/52) =$0.173/toolkit-week
24

12

Solution
Units
Toolkits per week

Parameters
= 10 , h = 0.173 , 2 = 25, p = 15 L=1

Solution
ratio = p/(p+h) = 15/(0.173 + 15) = 0.9886, z* = 2.28 s* = (10)(1) + (2.28)(25)(1) = 21.38
25

JITD Proposal
Plant CDCs Grande Distribuzione Chain supermarkets Customers

Distribuzione organizzata

Independent supermarkets

Customers

Physical Flow of Goods

Barilla Run depots Signora Maria Shops Customers

Proposed Flow of Information

- Barilla takes over ordering decisions at DOs & GDs -Problem is how to forecast end-customer demand?

1 Based

on Simchi-Levi et al (2002), Designing and managing the Supply Chain: Concepts, Strategies and Case Studies

26

13

JITD Proposal: Ordering Mechanisms

Information on intermediate stage demand, is available to Barilla; but Barilla has to work with end-customer demand forecasts to make replenishment decisions at DCs

Hope is that this will stabilize ordering patterns; reduce strain on plants; decrease lead time variability, and ultimately decrease stock-outs as well as inventory levels.

27

JITD Proposal: Demand Smoothing

Barillas actions aim to remove bias in demand as a result of supply uncertainty Barilla aims to control flow of goods based on its own best available information regarding its own capacities and process capabilities. Essentially, Barilla wants to take on the role of a central supply chain planner.

28

14

JITD Proposal: What is Just-in-Time?

Is this a Push or a Pull system? Replenishment based on usage, inventory, and forecasting Joint or global optimization, as opposed to local or sequential optimization

29

JITD Proposal: Stakeholders Involved

Internal Stakeholders Manufacturing / Production Sales Logistics External stakeholders CDCs Distributors Retailers

30

15

Criticisms of JITD Proposal

Risk sharing? Sales compensation in new mode of operations? Removal of inventory would expose previously unknown (or hidden) inefficiencies in some parties policies and organization Loss of control, and removal from decision-making authority from the hands of the distributors Loss of shelf space for Barilla products at both the DC as well as the retail level: was a problem feared by Sales
31

Criticisms of JITD Proposal (Cont.)

Result in more power to Barilla? Trust? Will Barilla share any profits / benefits from the efficient mode of operations, with the other parties? Will Barilla share the risks of loss and failure in the projects and ventures that would result from this program? Will Barilla demonstrate the competence to make this new mode of operations a living reality?
32

16

What Happened? Barilla (B) Case

Demonstrate that JITD benefits the distributors (lowering inventory, improving their service levels and increasing their returns on assets); Run experiment at one or more of Barillas 18 depots Maggiali needs to look at JITD not as a logistics program, but as a company-wide effort; Get top management closely involved Trust
33

34

17

35

Summary
Bullwhip effect is a fundamental phenomenon in supply chains (first law) Supply chain management strive to mitigate this effect
Identify causes Develop mechanisms to eliminate/reduce the causes Reallocation of decision right, partnerships
Common goal

Implementation difficulty Pilot programs Realign incentives

36

18