case report

aaasss22

EXTERNALCP13322_pcs.pdf

Home >Business & Finance homework help >Operations Management homework help >case report

Supply Chain Application and Policy SCM470-001 and 002

Dr. Carol PRAHINSKI

SCM470 - 002

Michigan State University

Table of Contents

Process Fundamentals.....................................................................................................................5

Earth Buddy....................................................................................................................................23

McLeod Motors Ltd.........................................................................................................................27

Quinte MRI......................................................................................................................................33

Ranger Creek Brewing and Distilling..............................................................................................49

International Decorative Glass........................................................................................................63

The Home Depot, Inc......................................................................................................................79

Industrie Pininfarina: The New Customer Decision......................................................................107

VF Brands: Global Supply Chain Strategy....................................................................................127

Supply Chain Application and Policy SCM470-001 and 002 SCM470 - 002

Dr. Carol PRAHINSKI Michigan State University

9-696-023 R E V : J U L Y 1 3 , 2 0 1 6

Professor Ann E. Gray and Research Associate James Leonard prepared this note as the basis for class discussion, with minor edits by Prof. Michael Toffel. It is a rewritten version of an earlier note by Prof. Paul W. Marshall, “A Note on Process Analysis,” HBS No. 675-038. Copyright © 1995, 1996, 1997, 1999, 2007, 2009 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to www.hbsp.harvard.edu. This publication may not be digitized, photocopied, or otherwise reproduced, posted, or transmitted, without the permission of Harvard Business School.

A N N E . G R A Y

J A M E S L E O N A R D

Process Fundamentals

Imagine that, upon graduation, you take a job managing a business whose operating processes need improvement. Perhaps you need to help management understand how to increase the value that operations provides to customers and/or improve the profitability of the operation. Or imagine that, upon graduation, you take a job in marketing, and you need to understand how the decisions made to improve operations will affect your new marketing programs, or how your new marketing programs will affect the ability of operations to do what they need to do. Or, as an executive in a start-up, you are concerned with both sets of issues.

Operations Management is about designing, managing, and improving the set of activities that create products and services and deliver them to customers. The first half of the TOM course is concerned with the “creating”. We call these activities, the people, the resources (including technology and knowledge), and the procedures that dictate how work is organized the operating system. (In TOM, when we talk about operating systems, we’re usually not talking about DOS or Windows.)

The basic building block of operating systems is the process. Most operating systems consist of multiple processes. A process takes inputs (in the form of raw materials, labor, capital [equipment / technology], knowledge, and energy), and creates outputs that are of greater value to customers (and, thereby, of greater value to the organization itself).

This note is an introduction to process analysis, a set of concepts and tool that will allow you to describe, measure, and ultimately improve operating processes.

As a simple example, imagine that you are in charge of a project for a large bakery supplying supermarket chains with products ranging from breads to pies. Your mission is to improve the baking process.

How will you start? Well, you will first have to develop a good understanding of the current operation, the activities that take place to transform flour, water, yeast, and other ingredients into baked goods, and the effort involved in each activity—such as the labor, materials, and equipment required at each step. You will also need to understand the different products the bakery offers, as well as your business’ competitive priorities, i.e., the reasons that customers buy them from you and not your competitors. Do you have lower prices, faster delivery, higher quality, or a better product line

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

that allows your customers to buy all their bakery needs from one source? Only after understanding the physical process itself, how it links to the performance of the bakery and the level of performance required by customers, can you begin to look for opportunities to improve the bakery’s profitability.

The goal of this overview is to provide tools that can help you understand operations, not just for a bakery, of course, but any type of operation. These tools are important not only for improving operations, but also for the daily management of an operation or for the design of a new operation.

This overview begins by discussing the activities that take place in a process. Analytical tools such as the process flow diagram are provided to help you walk into a new operation, such as your bakery, and understand how each of the process steps fits together. You’ll be introduced to the types of management choices for designing, operating, and improving processes. Next, measures of the performance of a process and basic process analysis, methods used to determine process performance, such as calculating what and how much a process is capable of producing, and how quickly, are introduced. You’ll see how different types of processes can be used to make the same product, and how managers choose which process to use. Finally, the note focuses briefly on the complexity stemming from uncertainty and variability in the process, factors that make managing operations particularly difficult.

Elements of a Process

Again, the basic building block of an operating system, which we will spend much of TOM analyzing, is the process. Consider some examples of processes. An automobile assembly plant takes raw materials in the form of parts, components, and subassemblies. These materials, along with labor, capital, and energy, are transformed into automobiles. The transformation process is an assembly process and the output is an automobile. A restaurant takes inputs in the form of unprocessed or semi- processed agricultural products. To these, labor (a cook and a server, for example), capital equipment (such as refrigerators and stoves), and energy (usually gas and/or electricity) are added, and the output is a meal.

Both of the processes mentioned above have physical products as an output. However, the output of some operating systems is a service. Consider an airline: the inputs are capital equipment in the form of airplanes and ground equipment; labor in the form of flight crews, ground crews, and maintenance crews; and energy in the form of fuel and electricity. These inputs are transformed into a service, namely, a means of transportation between widely separated points. Processes with a service output also include those found in a hospital, in an insurance company, and in a consulting firm. In a hospital, for example, capital, labor, and energy are applied to another input (patients) in order to transform them into healthier or more comfortable people.

In order to understand a process, it is useful to have a simple method of describing the process and some standard definitions for its components. A convenient way to describe an operating system is a process flow diagram.

Returning to our bakery example, depicted in Figure 1, let’s assume that there are two distinct production lines in the bakery for making bread. Flour, yeast, and water enter at the left and are converted into loaves of bread through mixing, proofing (letting the dough rise), baking, and packaging. This is a bit of a simplification, but we’ll use it for illustration. There are two mixers, two proofers, and two ovens organized so that the ingredients mixed on the first mixer are automatically fed into the first proofer, and then sent to the first oven. All of the baked loaves of bread are packaged on the same packaging line. F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

Tasks in this process are shown as small rectangles, flows as arrows, and the storage of goods as inverted triangles. We see two identical parallel lines for mixing, proofing, and baking. Within each line, the tasks of mixing, proofing, and baking are defined as being in a series relationship, because one step cannot start until the previous one is complete. The maximum capacity of the two parallel lines would be found by adding the capacity of each line. Work-in-Process Inventory (WIP) is shown before packaging because, at times, the bakery may produce different types of bread at the same time, one on each line, yet only one type can be packaged at a time. If there were parallel packaging lines, there may not be the need for holding WIP between baking and packaging except, perhaps, to allow the bread time to cool. Once packaged, the bread moves into Finished Goods Inventory, and from there is transported to grocery store customers.

Figure 1 Process Flow Diagram for Bread-Making with Two Parallel Baking Lines

If the mixers, proofers, and ovens were not set up as two distinct lines, and the product could flow from each mixer to either proofer and then to either oven, we would draw the process as in Figure 2. In this case, it is the individual tasks that operate in parallel, instead of two distinct parallel lines. (The distinction between these configurations will become important when performing a more detailed process analysis to determine the capacity of the system).

Figure 2 Process Flow Diagram for Bread-Making with Two Mixers, Proofers, and Ovens

We may also want to show, on a process flow diagram, tasks that are performed in parallel but that must both be completed before the process can continue. For example, our bakery makes filled croissants in addition to breads. For these, the mixing, proofing, rolling, and cutting of the pastry take place in parallel with the mixing of the filling as shown in Figure 3. All these tasks must be completed before the croissants can be filled and baked. Proofing the dough takes longer than any of the other pastry-making steps. Proofing also takes longer than mixing the filling. This means that the rate at which filling and folding takes place is limited by the rate at which the dough, not the filling, is ready. And the rate at which the dough is ready is limited by the rate at which proofing takes place. It is the rate of the proofing step, the longest task, that defines how much bread can be made per hour.

Bake

Raw

Materials

Proo fMix

Pack

Wo rk in

Process

Finished

Good s

BakeProo fMix

Bake

Raw

Materials

Proo fMix

Pack

Wo rk in

Process

Finished

Good s

BakeProo fMix

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

Note that the nature of the parallel activities for making croissants is different from that of the two bread lines working in parallel as in Figure 1. To determine the capacity of the bread-making operation up until the dough is baked, we add the capacity of each of the parallel bread lines. To determine the capacity of croissant-making, however, we would take the minimum of the capacity of the two different parallel processes, in this case, the capacity of pastry making. This is because the output of the two lines must be combined to make the final product. We will revisit this issue in Section 1.2, when we do a formal capacity analysis.

Figure 3 Process Flow Diagram for Croissant-Making

Once a process has been described using a process flow diagram, its components must be analyzed in order to draw some conclusions about its performance as a whole. In the following sections we will discuss each component of the process—the inputs, outputs, tasks, flows, and storage of goods—and begin to develop measurement and analysis methods along the way.

Inputs

As described above, the inputs to a process can be divided into at least four categories: labor, materials, energy, and capital. To analyze an operating system we must measure these inputs and determine the amount of each needed to make some amount of output. Usually we use physical units to measure the inputs – for example, hours for labor and joules for energy. It is sometimes more useful to measure the input in dollars by determining how much it would cost to purchase these units. Thus, in many analyses it will be necessary to consider the economic conditions influencing the cost of labor, materials, energy, and capital. Measuring the cost of inputs becomes more difficult and requires additional care as the time horizon lengthens.

Determining how much of any input is needed to make a given output entails varying degrees of difficulty. Some inputs (e.g., labor and materials) are fully consumed to produce an output and thus are easy to assign to that unit of output. For example, it is easy to measure how much energy the oven uses to bake a batch of bread. Other inputs, however, are utilized in the production of an output, but are not fully consumed—the oven itself, for instance. The capital input is often the most difficult of the four categories to assign to a specific output because it is almost impossible to measure how much capital is consumed at any point in time. Generally accepted accounting rules are often used to allocate fixed costs, such as capital, to each unit of output.

Outputs

The output of a process is either a good or a service. The process flow diagram in Figure 1 shows that the product is stored in Finished Goods Inventory (FGI) before leaving the system. In some organizations, the finished goods inventory is kept apart from the operating system producing the good and is managed separately. In others, the finished goods inventory does not exist at all: the

Raw

Materials:

Doug h

Finished

Good s

Wo rk in

Process:

Doug h

Mix Filling

Raw

Materials:

Filling s

Wo rk in

Process:

Filling

Fill &

Fo ld PackBake

Roll &

CutProo fMix

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

process produces the output directly for distribution. In fact, this is an important characteristic of most processes providing services; it is often not easy (or possible) to store it for later distribution.

Although it is a simple matter to count the number of loaves of bread produced by the bakery, or to count the number of patients served by a hospital, it may not be simple to place a value on this output. The question of valuing the outputs can be approached from an economic point of view if a market will place a value on the output through the pricing mechanism. So, if we know the revenue that can be obtained from selling the good or service, that should serve as a measure of its value. For this reason, we must have a good understanding of the economic environment within which the process exists. “What are the market conditions?” and “What is the competition doing?” are thus important questions to address when analyzing a process.

For a new product, or one that has some improved characteristics, however, the question of what price will be paid for the output is difficult to answer unless some other information is known about the output. Here, we will consider three output characteristics: the cost of providing the output, the quality of the output, and the timeliness of the output. It is often the case that none of these measures is easily obtained, but they can serve as a checklist in our analysis of operating systems. If we are going to consider making a new type of bread, or increasing the quality of the bread, we may not know the price we can get for it. However, we do know that to value the new product, it is important to take into account the new product’s characteristics, market conditions (is there an oversupply of specialty or high-end breads?), and the competitive situation (should we match the price of a competitor’s similar product?)

Tasks, Flows, and Storage

So far we have discussed what goes into and what comes out of a process. We must also understand what goes on inside a process. The specifics of every process are different, but there are three general categories for all activities within the process: tasks, flows, and storage.

A task typically involves the addition of some input that makes the product or service more nearly like the desired output. Some examples of tasks are (1) operating a drill press to change a piece of metal; (2) inspecting a part to make sure it meets some standard; (3) flying an airplane; and (4) anesthetizing a patient before an operation. A task quite often takes the form of added labor and capital; in processes with some form of automation, capital and/or material may be substituted for labor in a task.

There are two types of flows to be considered in each process: the flow of goods and the flow of information. Figure 4 depicts a process flow diagram with the flow of information shown explicitly— the flow of physical goods is indicated by solid lines and the information flow by broken lines.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

Figure 4 Information and Physical Process Flow Diagram for the Bread-Making Process

Information flows in the bread-making process depicted in Figure 4 are quite simple; they take the form of recipes and production orders. The list of ingredients and quantities for the type of bread that will be made next must go the operators or material handlers in charge of getting the raw material ingredients to each mixer. Information on mixing times and methods must go to the operators of the mixers, and baking temperatures and times must go to operators of the ovens. We will also have to inform packaging of what types and quantities of breads will be arriving to the packaging area so that they can set up their equipment with the correct bags.

In some types of operations, the information flows take place with the physical flows, often in the form of a routing slip attached to a single product or a batch of products. The analogy here would be the entire recipe and the production order moving with the bread. The oven operator, for instance, would receive baking instructions with the proofed dough as it arrives at the oven. If the operator could not or would not need to adjust the oven in advance, not providing this information in advance would not cause any production delay and would simplify the information flows. Other information that might be included on the routing slip includes the packaging lines that the loaves should be sent to (if there are multiple packaging lines), the appropriate bags to use for packaging, the supermarket name and location, the delivery date and time, and possibly even the truck into which the finished product should be loaded.

When the information does not physically move through the process with the goods, the worker may need to go to a central location to obtain the information before performing the task, or the worker may have the necessary information at the workstation or in his or her head. In analyzing a process, it is often important to consider the information flows in addition to the physical flow of goods or services.

Storage (the holding of inventory) is the last of the three activities within a process. Storage occurs when no task is being performed and the good or service is not being transported. In Figures 1 – 4, we have shown the storage of goods as inverted triangles. While the bakery is operating, there will usually be work-in-process inside the mixers, proofers, and ovens, at the packaging machines, as well as some work-in-process inventory between each step, and raw materials and finished goods inventory in the warehouse. If there is no storage between two connected tasks, there must be a planned continuous

Bake

Raw

Materials

Proo fMix

Pack

Wo rk in

Process

Finished

Good s

BakeProo fMix

P hy sic al Flow

Information Flow

Reco rd s &

Control

Initiatin g Order

or Req uest

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

flow between these tasks to allow the receiving task to operate continuously. Figures 1 and 2 show only one work-in-process storage, whereas Figure 3 shows two. In many processes that are considered continuous, there are at least a few units of work-in-process inventory on a rack or chute waiting to be fed into a machine. Although these units are technically in storage and could be depicted on the process flow diagram as inverted triangles between processing steps, they are often left off of the diagram when they represent just a few minutes of processing time. Similarly, the transport of goods from step to step within the process could be shown as another set of tasks, but unless the necessary times are long, we will generally omit these in process flow diagrams for simplicity.

It is also possible, and in fact necessary, to store information. This storage is shown as a circle in Figure 4, with an arrow coming in from the environment to start the process. In this case, there are two kinds of information: records and control. The term records typically refers to general instructions, such as blueprints and instructions of how a product should be made (i.e., the “recipe”). These records are product-specific. Records may also be machine-specific, tracking repair and preventative maintenance histories, for example. The term control usually refers to information specific to a given order, such as the order quantity, customer name or number, due date and routing procedure for the order, or special instructions that make the order different from the generally accepted procedures outlined in the records.

Measuring the Performance of a Process

So far we have defined the process in general terms and given names to various components of the process, namely the inputs, the outputs, and the tasks, flows, and storage within the process. We have also noted that the process does not exist in isolation. Economic conditions influence the values of inputs and outputs, and the state of technology influences the nature of the tasks and flows. Using these concepts as a base, we can now explore some process characteristics, concentrating on four: capacity, efficiency, flexibility, and quality.

Capacity

Capacity is the maximum output rate from the process and is measured in units of output per unit of time: a steel mill, for instance, can produce some number of tons of steel per year, or an insurance office can process some number of claims per hour. Capacity is easy to define and hard to measure. It is often possible to determine the theoretical capacity of a process—the most output it could generate under ideal conditions over some period of time. For planning purposes and management decisions, however, it is more useful to know the effective capacity of a process – and to measure effective capacity, we must know a great deal about the process, carefully analyzing the particular situation at hand.

Managers often believe that the capacity of a process is an absolute fixed quantity. This is rarely true. The capacity of a process can change for many reasons, and we will encounter several cases where this is a key factor. The steel mill, for instance, may be designed for some ideal capacity, but its effective capacity may be different due to a variety of internal and external factors, as well as management decisions. The nature and availability of the raw materials being utilized, the mix of products being produced, the quantity and nature of the labor input, and the number of shifts of operation will all impact the effective capacity. The yield of the process is also important. In most instances, the rate of good units produced is the relevant capacity measure.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

Efficiency

Efficiency is a measure that relates the amount or value of the output of the process to the amount or value of the input. “Efficiency” is widely used to measure physical processes. Every engine has an efficiency, expressed as a ratio of output energy to input energy. So, an engine with 75% efficiency can deliver 75% of the input energy as useful output energy. The energy efficiency of physical systems cannot exceed 100%; the useful output energy is always less than the energy input. This is not generally true of economic processes, however. For example, if the process is going to generate sufficient resources to support its own continued operation, the value of the output should exceed the value of the input. If we measure the value of output by the revenues it will bring in the market, and if we measure the value of inputs by their costs, the measure of efficiency is profit, i.e., revenue minus cost. Thus, the profit is the value of output minus the value of input. Profit, however, is a very simplistic definition of efficiency; measuring efficiency is generally much more complex.

In some cases, the price received for the product is not a good representation of the economic value of the output. In certain markets, for example, it may be possible initially for a company to sell a product of low quality at a standard price. Over time, however, the company’s reputation may be hurt by doing this, and all of the company’s products, not just the low quality product, might become less desired by the market. The long-term loss in revenue should have been considered when establishing the cost and quality level of the original product. When determining the efficiency of a process as measured by profitability, it is important to look at long-run profits, not just the profit generated from any short-run action.

Utilization is another common measure. Utilization is the ratio of the input the process actually used in creating the output to the amount of that input available for use. In a labor-intensive process, for instance, direct labor utilization is often an efficiency measure. If, say, 100 workers are employed in a given process over an eight-hour shift, and 700 hours of labor were consumed in the actual manufacture of product, then the direct labor utilization during that shift was 87.5% ([700 hours/(100 X 8 hours)] = 0.875). In a similar way, to measure capital efficiency, companies often pay a great deal of attention to machine utilization, which measures the percentage of time the machinery is used. Typically, this includes machine setup time as well as the time the machine is actively producing output.

In a small bakery where workers mix the dough, form loaves, and move the product from one step to another by hand, labor utilization is a critical measure of bakery performance. In an automated bakery, machine utilization may be more relevant.

Flexibility

A third characteristic we want to consider in analyzing a process is its flexibility. This is a measure of how long it would take to change the process so that it could produce a different output, or could use a different set of inputs. Flexibility, which allows a process to respond to changes in its environment, is also the least precise and hardest to define of the characteristics we have considered thus far. Flexibility must often be described in qualitative terms; doing so, however, does not make it any less important to managers.

Returning to our bakery, its flexibility may be described by the different types of bread that can be produced on a given line, or whether pastry products can also be made on the same line as bread

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

products. Another type of flexibility may further be measured by the time required to switch the line from producing one type of product to another.1

Quality

Like flexibility, quality may be described in different ways. Product quality can be evaluated using external measures—comparing the product with others available in the marketplace—or using internal measures—comparing individual units with one another or with the product design specification. External quality measures generally assess how well the product design satisfies the wants and needs of customers. Product performance, features, reliability, durability, serviceability, and design aesthetics may all be components of product quality. Internal measures of product quality generally assess whether individual units meet design specifications.

In addition to designing, measuring and controlling product quality, a manufacturer also designs, measures and controls process quality. In order to produce a product with certain specifications, the process must be operating within certain tolerances. Process measures, such as the temperature in a kiln or the amount of force applied by a punch press, are generally used in assessing process quality. Any piece of processing equipment has specific capabilities defined by the range of process specifications it is able to achieve. A piece of equipment may not be able to perform a certain type of operation, such as grinding a piece of metal to a certain smoothness, if doing so requires operating outside this range or it may not be able to consistently perform the operation properly. In other circumstances, the equipment is capable of consistently operating within certain specifications, but is not operating consistently within these specifications because of poor equipment control. Both the design of the process and the way in which the process is operated are important determinants of process, and thus, product quality.2

Within the plant, the impact of poor quality can be increased scrap, rework, yield losses resulting in lost capacity, downtime, additional testing, and lost management and worker time. If poor quality product leaves the factory, the impact can include a loss of goodwill toward the company and its brands, time and cost responding to customer complaints, and repair costs.

Process Terminology and Process Analysis

As the new manager of the bakery, once you understand its products and the process steps, and you have created a process flow diagram, you may want to determine the capacity of your operation. To do this, let’s further simplify the bread-making example, as illustrated in Figure 5. Here, there are two steps required to prepare bread. The first is bread-making, which includes preparing the dough and baking the loaves, and the second is packaging the loaves. There is only a single line for mixing, proofing, and baking, and it is illustrated by a box representing the entire bread-making line.

1 A more detailed description of different types of process flexibility and how they can be managed can be found in: Upton, David, “The Management of Manufacturing Flexibility,” California Management Review, Winter, 1994, or Upton, David, “What really makes factories flexible?,” Harvard Business Review, July–August, 1995.

2 A more detailed description of different measures of quality can be found in: Garvin, David A., “Competing on the eight dimensions of quality,” Harvard Business Review, November-December, 1987, or Garvin, David A., and Artemis March, “A Note on Quality: The Views of Deming, Juran, and Crosby,” Harvard Business School Note 9-687-001, 1987.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

Figure 5

Based on the size of the mixers in the bakery, bread is made in batches of 100 loaves each. Bread- making completes a batch of 100 loaves every hour; thus, the bread-making cycle time for a batch of 100 loaves is one hour. Although packaging needs only ¾ of an hour to place the 100 loaves in bags (its cycle time), the rate at which the entire process can operate is paced by the bread-making step. Thus, over the course of a day, packaging will incur idle time during the ¼ hour periods in which the next batch of bread is still being made but packaging has already completed bagging the previous batch. Bread-making is the bottleneck of the operation. The cycle time for the entire process is 1 hour, the maximum of the cycle times of the two operations in series. Given the cycle time for the entire process, we can determine its capacity. Simply put, if the cycle time is 1 hour per 100 loaves, the line has a capacity of 100 loaves per hour, the inverse of the cycle time. To determine the daily capacity, we would need to know the number of hours the bakery is in operation per day. With any of these terms of measurement, it is important to be very explicit about the units (i.e., loaves per hour, minutes per loaf), particularly when performing calculations.

To perform our analysis of the capacity of the bread-making line above, we introduced some new terms and concepts. While we have provided formal definitions below, we must also stress that calculating these measures requires close attention to the specifics of a particular process. In addition, different firms sometimes define these terms in different ways for their own internal use. This variation is reflected in some of our case materials. However, for the purposes of class discussion, it makes sense to try to adhere to a common vocabulary.

Cycle Time (CT): The cycle time of a process is the average time between completion of successive units. In other words, cycle time answers the question, “How often does a unit complete the process?” Cycle time can be similarly defined for an individual task or for portions of a process.

Often a process or portion of a process is not operated at its theoretical capacity. In those instances, you may need to distinguish between the minimum amount of time that could elapse between the completion of successive units (the minimum cycle time of the process) and the amount of time required for the process to actually complete successive units (the actual cycle time). For example, in the process depicted in Figure 5, because bread-making, the process bottleneck, requires an hour to finish a batch, packaging will only receive batches of bread every hour. Thus, while its task time is ¾ of an hour, it could be operated more slowly. As long as it is operated so that it can package a batch of bread in no more than an hour, there will be no loss of capacity.

Bottleneck: The bottleneck of a process is the factor which limits production. Usually, we will speak of the task with the longest cycle time as a bottleneck, such as bread-making in Figure 5. In other situations, the available labor may be the bottleneck. In some settings, information, raw materials flow, or even a specific order may be a bottleneck. Just as the neck of a bottle limits the rate at which the liquid inside can be poured, a process bottleneck limits how quickly products can move through the process, and thus determines the process cycle time. The bottleneck may shift depending on what products are being produced or what labor or equipment is available at any point in time. Because bottlenecks pace a process and limit its capacity, they are important focal points for management attention.

Bread -Mak in g Pack agingWIP FGI

Cycle Time :

1 hou r/10 0 lo aves

Cycle Time :

3/4 hou r/10 0 lo aves

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

Idle Time: Idle time refers to the time when useful work is not being performed; the term can be applied to a worker or to a machine. Time spent waiting to receive or deliver a unit is idle time unless there is some other useful task to be performed in the interim. Idle time can be present even in a perfectly balanced process. A worker in the packaging department, for example, may merely load twenty loaves of bread on a machine and then stand by while it bags the bread. This time might be idle time for the worker (unless he or she is needed to monitor the equipment’s performance), while it is not idle time for the packaging machine. In Figure 5, the packaging machine will be idle ¼ of the time.

Capacity: Capacity is a measure of how much can be produced or serviced in a specified period of time, e.g., tons per day, parts per minute, customers per hour. The capacity of a process is determined by the output rate of the process bottleneck. Capacity utilization is a measure of how much output was actually achieved relative to capacity (how much output could have been achieved in an ideal situation). If the capacity of a process is 500 units per day and on a given day 480 are produced, then on that day capacity utilization was 96% (480 units/500 units = 0.96). We may speak of the capacity of a task, a machine, a worker, a work area, or of an entire process.3 Cycle time and capacity are closely related. If the cycle time of a task is 30 minutes per unit, then the capacity of that task is 2 units per hour.

Capacity seems a straightforward measure, and for a specific task producing a specific product it often will be. But finding relevant capacity measures for an entire process can be complicated. In many cases, the system capacity will depend on the size and mix of products and order sizes. And because the capacity of an entire process is affected by product mix, staffing, labor contract issues, maintenance time, etc., the effective capacity and capacity utilization will depend upon the way the process is managed.

Suppose we have two lines for bread-making, as shown in Figure 6:

Although the cycle time for each bread-making line is 1 hour/100 loaves, the cycle time of the two lines together is ½ hour/100 loaves or, equivalently, 0.005 hour/loaf. Because the packaging line takes

3 Companies such as Toyota Motor Manufacturing ensure that capacity exceeds customer requirements by calculating the takt time. This is determined by taking the time available to produce a certain product and dividing by customer demand for that product. The result, the takt time, is the production rate necessary to meet demand, and, if possible, a process should be designed with this takt time in mind. (For example, if an assembly line is available for 16 hours (960 minutes) per day and customer demand is 1,000 cars per day, the takt time is 0.96 minutes per car, and in order to fully meet demand, the process must be designed so that the process cycle time is no more than 0.96 minutes.)

Bread -Mak in g

Pack agingWIP FGI

Cycle Time :

1 hou r/10 0 lo aves

Cycle Time :

3/4 hou r/10 0 lo aves

Bread -Mak in g

Cycle Time :

1 hou r/10 0 lo aves

Cycle Time :

1/2 hou r/10 0 lo aves F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

696-023 Process Fundamentals

¾ hour to bag 100 loaves (or 0.0075 hour/loaf), it becomes the bottleneck of the process. If both bread- making and packaging were operated for the same number of hours each day, we would not make bread at its maximum cycle time rate because we would not have the capacity to package it. There is, therefore, an imbalance in the cycle times of the two parts of the process. If, however, we could operate packaging for three shifts and bread-making for two shifts each day, then the daily capacity of each would be identical. To do this requires building up a shift’s-worth of inventory each day as work-in- process that packaging would then bag during the third shift.

Balance/Imbalance: If every step in a process had the same cycle time (and performed consistently at that precise cycle time, with no variability), then the process would be in perfect balance. This, however, is virtually never achieved in practice. The processes shown in Figures 5 and 6 are both imbalanced. If a system is not perfectly balanced there will be potential idle time at the non-bottleneck parts of the process. Although the cycle times are imbalanced in Figure 6, we can balance the daily capacity at each of the two steps by adjusting the hours worked per day in each step.

For the processes above, we might want to know the throughput time-—how long it takes to actually make and pack a batch of 100 loaves. For the process in Figure 5, the throughput time is 1¾ hours, as long as packaging begins immediately once the 100 loaves are made. The throughput time for the process in Figure 6, however, depends upon how the process is managed. Let’s assume that we operate both steps in the process (bread-making and packaging) for the same number of hours per day. If we start a new batch every ¾ hour, alternating between bread-making lines, each of these batches will be able to proceed directly to packaging, resulting in a throughput time of 1¾ hours. We could start a new batch on both bread-making lines at the same time, every 1 ½ hours (any more rapidly would mean that we would be producing more bread than we could package). While one batch would move immediately to packaging, for a throughput time of 1¾ hours, the other would have to wait the ¾ hour for the first batch and would then take ¾ hour to be packaged, for a total throughput time of 2½ hours. Thus, the average throughput time would be 2.125 hours. As seen by this example, throughput time is a function of the way in which a process is managed.

Now, let’s assume that our bakery operation consists of three steps, as in Figure 7.

Figure 7

Bread-Making Packaging WIP FGI

Cycle Time :

1 hour/ 100 loaves

Cycle Time :

¾ hour/ 100 loaves

Cycle Time :

1 hour/ 100 loaves

WIP Pricing &

Palletizing

The throughput time for this process would be 2¾ hours, assuming that the batches of bread do not wait at all between steps. As in the last example, management decisions regarding scheduling could affect the lead time. If, for example, each step began operation at the same time, every hour on the hour, then packaged bread would wait for ¼ hour before pricing and palletizing, and the throughput time would be 3 hours. There would probably be no reason to follow this policy in this situation, but for many assembly lines the units are transferred by conveyor from one step to the next, and movement of the conveyor must be paced by the slowest step in the process.

So far, we have not considered the direct impact of work-in-process on throughput time. In the bakery described by Figure 7, bread may not flow immediately from one step to the next. If, occasionally, a batch of bread does not bake properly, management may choose to keep an inventory buffer of one batch between steps to minimize disruptions. If there is a bad batch of bread, then, the

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

packaging line would not have to shut down and, an hour later, be restarted. It would simply use up the batch in the buffer. Suppose we manage the process so that every step begins operation every hour on the hour. A policy of keeping one batch of work-in-process as a buffer between steps would add 2 hours to the usual throughput time. To see this, imagine following a small amount of flour through the entire process. After being made into bread, it now waits 1 hour for the batch ahead of it before moving into packaging. After packaging, it then waits 1 hour until the batch ahead of it is completely priced and palletized.

Another reason why there may be work-in-process between steps is because it may take some time to move the bread between steps, either manually or on a conveyor. This time would also add to the throughput time.

Throughput time: Throughput time refers to the length of time spent in the process. For a single task the cycle time and throughput time may be equal if only one station is performing the task. When a process involves multiple tasks or steps, the concepts of cycle time and throughput time are quite different. Cycle time refers to how often a unit “drops off” the end of the process whereas throughput time refers to how long that unit takes between entering and leaving the process, including any in- process storage or transport time. Process cycle time determines a process’ capacity, which limits the volume of product that the process is able to produce, and which, in turn, given products’ prices, determines the maximum revenue that a process can generate. In contrast, throughput time is an important determinant of the speed of a process. For a process that produces customized product to order and carries little finished-goods inventory, throughput time may be an important competitive performance measure, since throughput time may be a major determinant of how long a customer will have to wait after placing an order. Similarly, for many services, throughput time determines how long it takes for a customer to be served.

If units must wait between steps, the throughput time for a process may be far greater than the sum of the processing times of its individual tasks. Units may wait as work-in-process inventory between tasks, either while other batches are being processed, or while other units in their own batch are processed. Idle time may also add to the throughput time. This often occurs when a line is imbalanced but is paced by a conveyor at the speed of the slowest task, resulting in idle time between most steps of the process.

Batch Size (also called Lot Size): Most processes produce more than one product type. Suppose a process produced three products: P1, P2, and P3. The process could produce one unit of P1, then one unit of P2, then one unit of P3, then one unit of P1 and so on until we had 100 units each of P1, P2, and P3. Alternatively the process could produce 100 units of P1 before beginning production on 100 units of P2. In the first case, the lot size is equal to one unit; in the second case, the lot size is 100 units. If time must be expended setting up the equipment to make the transition from producing P1 to producing P2, then these two different lot sizes will result in quite different throughput and cycle times. The batch size, then, is number of units of a particular product type that will be produced before beginning production of another product type. Different product types in the same plant may have different lot sizes.

The size of a lot, or batch, may be constrained by physical limitations such as, in our bakery example, the size of the mixers or ovens. It may be determined by the size of an order (e.g., a customer orders 300 units of a special part). Or, the lot size may be strictly a management decision. A company making three colors of telephones, for instance, may choose to mold 1,000 plastic casings of each color before changing over to another color.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

Setup Time/Run Time: Setup time refers to the time spent arranging tools, changing dyes, setting machine speeds, cleaning equipment, etc., in preparation for the beginning of work on a specific type of product. Depending on the type of process, it might be necessary to spend from a few minutes to a few hours setting up to make the process ready for the transition from producing one product type to producing another. Setup time does not necessarily mean idle time for the task or process, however. It may be possible for a worker to do much of the setup for the production of a second product type “off- line” during the production of the first product type. This minimizes the amount of machine capacity lost due to setups. For our purposes, setup time refers to any time that is necessary for production but is independent of the number of units to be produced. In this respect, for production of a given lot, the setup time is fixed and the run time is proportional to the batch size. This is a useful distinction. Long setup times, for example, may make it attractive to produce in large lot sizes because the fixed cost of the set up can be spread over a larger production volume.

Run time per unit is the amount of time actually spent manufacturing the item (or performing the service) independent of the time required to set up the equipment. If units in a lot are processed sequentially, the run time per lot is the run time per unit multiplied by the number of units in a lot (i.e., the time the units in the particular lot are actually “running” on the equipment).

Returning to the concept of throughput time, if we were to follow a unit from when it first enters a process until it leaves, we would see that during some of this time it would be worked on (run time), it may spend some of the time waiting for a machine to be set-up (setup time), and some of the time waiting due to imbalance in the processing steps, machine downtime, or time spent as WIP inventory waiting for other products to be completed.

In addition to measures of cycle time, capacity, and throughput time, to evaluate a process you may also want to know the direct labor content and the direct labor utilization.

Direct Labor Content: Different organizations and disciplines use the term “direct labor content” in different ways. For our purposes, “direct labor content” refers to the actual amount of work “contained” in the product. Returning to the process in Figure 5, let’s say that for a batch of 100 loaves, while the packaging equipment has a cycle time of ¾ hour, the packaging operator spends only 40 minutes in activities such as loading the loaves onto the machines, setting up the right bags on each machine, and making any necessary machine adjustments. The direct labor content in packaging would be 40 minutes/100 loaves or 0.4 minutes/loaf. For the total direct labor content of the bread, the direct labor content of bread-making would need to be included. Indirect labor hours (maintenance, materials handling, management, etc.) are not included in the calculation of direct labor content. Setup time may or may not be included in direct labor content. Exactly what goes into direct labor content varies by firm, but, generally, setups performed by dedicated setup workers are not included because those workers are usually classified as indirect labor, and setups done by operators on their own machines are included, because these workers are classified as direct labor.

Note that direct labor content is not the same as direct labor cost. “Content” refers to the work done in actually manufacturing the product or performing the service (or setting up to do so), not to the wages paid. Labor cost differs from labor content due to imbalance, vacation pay, paid breaks, etc. Recall that the process cycle time in Figure 5 is 1 hour/100 loaves. Even if the packaging operator is busy for only 40 minutes of every hour, the operator is paid for an entire hour. Thus, idle time adds to labor cost but does not affect labor content.

Direct Labor Utilization: Rather than measure idle time or direct labor content in minutes, it is often more useful to talk in percentage terms. Direct labor utilization is a measure of the percentage of time that workers are actually working on a product or performing a service, i.e.:

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

Direct Labor Utilization = Direct labor content

Total available labor time

Total available labor time consists of both direct labor content and any idle time. In the example above, direct labor utilization in packaging is 67% ([40 minutes/60 minutes] X 100). Part of the reason that the utilization is not 100% is due to the imbalance between packaging and bread-making (this accounts for 15 minutes of idle time per hour), while part (5 minutes of idle time per hour) is due to the work design (driven by the level of mechanization) in the packaging operation. If there are as many packaging operators as bread-making operators, and if the bread-making operators are 100% utilized, the overall direct labor utilization for the process would be the average of the two numbers, or 83%.

All of this discussion has assumed that yields are 100%, i.e., that every unit that starts through the process goes all the way through every step of the process without mishap. The presence of rejects and/or rework complicates the analysis, of course. The impact of defects resulting in yields of less than 100% will be explored in subsequent modules.

When calculating process performance measures, carefully consider the details of the process and the managerial question you are seeking to answer. A good conceptual understanding of the terms in this overview will guide you in applying them to new environments and new types of decisions.

So far we have described some of the physical attributes of operating systems and how to measure their performance. We now turn to the decisions made in the management of operating systems.

Management Decisions

There are three general categories of decisions that must be made to manage an operation. First is the design of the operation, and of the distinct processes that constitute the operation. Second is the ongoing set of decisions—the operating decisions—that determine what is made, by whom, using what equipment, and at any point in time. Finally, there are process improvement decisions, which may be made, for example, to increase the output, lower cost, or improve the range of products that can be produced. You will be exposed to an extensive range of management decisions through the cases in this course, going far beyond the list below.

Design Choices. To start a new operation, the type of process must be selected. Although a high- volume assembly line might be chosen for a large bakery that makes only breads, a smaller, more flexible set of mixers and ovens might be chosen for a bakery making a wide variety of small batches of specialty breads, pies, cookies, cakes, rolls, and croissants. After selecting a general process type, the actual technology must also be chosen. The choice of process is likely to limit somewhat the choice of technology. When purchasing a high-volume line, a firm is likely to have the option of selecting a much higher degree of automation than if purchasing individual machines designed for lower volumes and a wider variety of products. Other design choices include specifying the capacity of the operation, with an eye to future sales forecasts and the cost of adding capacity later to meet those forecasts. The way that the equipment is situated in the space available—the process layout—must also be determined in conjunction with the pattern of material and information flows.

The choices made in designing a particular process determine in large part the inputs needed to provide the process outputs. It also determines in large part the range of outputs possible. As a simple example, consider two alternatives for providing copies of some information, one using a printing press and the other using a copier. The press involves several tasks to produce the information such as typesetting, proofreading, and press operation. The copier, on the other hand, requires only an operator Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

and an original document containing the information and requires no setup activity. The cost of providing copies is different for each type of process. The printing press has a much higher setup cost but each additional copy is very cheap. Thus, for large numbers of copies the press has an economic advantage.

Ongoing Operating Decisions. Once an operation is up and running, the main management tasks are often order selection, scheduling, setting batch size, and inventory management. The bakery’s manager(s) would need to determine, for example, when to make whole-wheat bread and when to make raisin bread over the coming week, how much of each to make, whether to take an order from a new grocery store, and how much inventory to hold as raw materials, work-in-process, and finished goods. Although these decisions may seem relatively straightforward for a bakery, the complexity increases with the number of different products, the lack of excess capacity available, the number of different customers and their requirements, and the cost of holding inventory.

Process Improvement Decisions. Managers can decide to physically alter the process by changing technology or adding machines or workers. They can redesign the physical or information flows, change the design of individual tasks, or improve the methods by which the process is managed. These decisions are much like design decisions, but the existing system puts additional costs and constraints on the changes that can be made. The objective of process improvements may be to improve the cost, quality, or timeliness of the output. It may also be to make the process more flexible, allowing entirely new outputs to be made.

Management Complexity

Variability and uncertainty in the inputs, in the transformation process itself, in the outputs, or in demand increase the complexity of managing an operation. This can be illustrated using our bakery example. The simplest bakery produces one type of bread on a single line. No routing or product mix decisions need to be made, and very little information needs to be managed. The primary management tasks involve fixing any problems that arise, scheduling overtime if necessary, and looking for ways to improve the efficiency of the process or the quality of the bread.

Once there is variability in the process, it becomes more difficult to manage. One source of variability may be the inputs. If the flour purchased from different vendors is slightly different, methods must be in place to make the necessary adjustments in the quantities used, the mixing and proofing times, or the baking time. This may require more sophisticated control of equipment and additional quality control activities. It also requires an information system in place to inform operators (and the machines, directly, if the system is automated) of the composition of the flour being used and how to adjust for it. The bakery may also want to track the use of the different flour in such a way that they could determine if the differences in the flours lead to any differences in customer satisfaction. This would place an additional burden on the information system.

Also, the more types of bread that the bakery makes, i.e., the higher the variability in outputs, the more difficult it is to manage. A bakery making 16 different types of bread has to schedule the process so that each type is made in the right amount, and demand is met. In addition, a set-up is involved in switching the machines from making one type of bread to another. The mixers will have to be cleaned and the oven temperature may need to be reset. Moreover, variability in demand also adds to the complexity of managing the bakery. If bread demand is seasonal, even if that demand is perfectly predictable, many more loaves will have to be produced during certain months of the year—and that may entail increasing the number of workers for that period or scheduling overtime.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Process Fundamentals 696-023

In our discussion so far, the factors adding to management complexity have all been sources of variability, that is, changes that are known and anticipated. But uncertainty in inputs, in the process itself, in outputs or in demand also leads to management complexity. Thus, in the bakery, it may happen that, owing to incomplete mixing or to fluctuations in the temperature in different parts of the oven, for example, not all the bread that comes out of the oven is good. Uncertainty in the subsequent yield of the process—not knowing exactly how many good loaves you will get from a given amount of dough—makes the scheduling task much more difficult. It also will probably require additional labor to sort out the bad bread, and may require that yield information be kept so that scheduling can be adjusted as necessary to ensure that demand is met. Uncertainty in demand also adds to management complexity. Inventory is often used to make it easier to fill demand when uncertainty exists in yield and/or demand. But inventory may be costly to hold, and can also increase management complexity because it, too, must be managed.

As you analyze new case situations, consider what it is that makes the management task complex in the different environments presented. Look for ways to eliminate the sources of complexity or to simplify the task of managing it.

Summary

In the bakery example presented here, we have not tried to address a management problem. In any real case it is clear that the nature of the problem should guide your analysis. Instead, we have tried to show the first two of several steps that might be useful in an analysis designed to address a management issue. These steps are summarized below.

Define the process—determine the tasks and the flows of information and goods. Also, determine where inventory is kept in the process. This effort can be recorded in a process flow diagram.

Determine the capacity or range of capacities for the process. This will require an analysis of each task and a comparison of how these tasks are balanced. In addition, determine the effect of inventory in the system on the capacity of tasks and flows. Inventories may allow the process to operate out of balance for some time, but in the long run the capacity of the process is limited by the capacity of its slowest task.

In many instances you will also need to determine the cost of inputs and relate these costs to the value of the output in some market by comparing the cost, quality, and timeliness of this output to the needs of the market.

We’ve seen that we cannot fully describe a process simply by the physical tasks, but that we need to consider management decisions ranging from how information is used to how work is scheduled in the process.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

696-023 Process Fundamentals

Glossary of TOM Process Analysis Terms

Throughput Time The amount of time each unit spends in the operating process. This includes time during which the unit is actively being worked upon at each step of the process, as well as any time spent waiting between steps. In a manufacturing process, throughput time is sometimes referred to as “manufacturing lead time”. The concept of a lead time applies to the total time spent in any process in which the start and finish are well-defined events. We will talk about lead times, for example, in the entire order-to-delivery process.

Cycle Time (CT) The average time between completion of successive units. It is directly related to the output rate. A process with an output rate of 4 units per hour has a cycle time of 15 minutes.

Work-in-Process (WIP) The number of units in the process at any point in time. If the process includes buffer inventories between steps, then the work-in-process is the total number of units being worked upon as well as waiting in inventory between steps. The units in inventory are usually referred to as work-in-process inventory, to distinguish them from raw materials inventory or finished goods inventory.

Bottleneck The production resource that limits the capacity of the overall process. This is often the production equipment at the step with the lowest overall capacity, i.e., the longest cycle time. In some situations, the bottleneck resource may be labor available at a particular step or steps.

Capacity The maximum rate of output of a process, measured in units of output per unit of time. The unit of time may be of any length: a year, a day, a shift, or an hour.

Utilization The ratio of the input actually used over the amount of the input available. Labor utilization is the ratio of the actual labor time spent processing to the total amount of labor time available. Differences between the two can be due to inefficiencies in the process that lead to lost working time, as well as to imbalances in the cycle times at each step of the process that lead to idle time of workers at some steps while those at others are working. Capacity utilization is the ratio of the capacity actually used (i.e., the output of the process) to the total capacity available.

Process Flow Diagram The diagram depicting the activities in a process and the flows between them. While most process flow diagrams focus on the physical processing of goods, information processing may also be depicted.

Batch Size (also called Lot Size) The number of units of a particular product type that is produced before beginning production of another product type.

Note: In all of the above definitions, a “process” may refer to the complete production process, such as the making of bread from start to finish, or to a segment of the complete process, such as the packaging process.

Source: Adapted from Professor W. Bruce Chew, “A Glossary of TOM Terms,” HBS No. 687-019.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

S w

9A94D019 EARTH BUDDY® Professor Chris Piper prepared this case solely to provide material for class discussion. The author does not intend to illustrate either effective or ineffective handling of a managerial situation. The author may have disguised certain names and other identifying information to protect confidentiality. Ivey Management Services prohibits any form of reproduction, storage or transmittal without its written permission. Reproduction of this material is not covered under authorization by any reproduction rights organization. To order copies or request permission to reproduce materials, contact Ivey Publishing, Ivey Management Services, c/o Richard Ivey School of Business, The University of Western Ontario, London, Ontario, Canada, N6A 3K7; phone (519) 661-3208; fax (519) 661-3882; e-mail [email protected]. Copyright © 2003, Ivey Management Services Version: (A) 2010-02-23 Earth Buddy® was rapidly becoming the hit novelty item of the summer. Although it was only mid-July, Seiger Marketing had already moved and expanded its Earth Buddy division’s factory and warehouse twice since production began in mid-April. Even so, current production levels were straining the physical limits of its latest facility in Toronto, Ontario. Nothing was certain, however, and Anton Rabie and Ronnen Harary, recent Ivey Business School graduates and Earth Buddy’s co-owners, were reluctant to give their production director and business school classmate, Ben Varadi, any production advice except: “Remain flexible. We could get an order for 100,000 units, but if the order doesn’t arrive, we would have to put the workforce on hold. We can’t afford to carry large inventories.” Against this background of uncertainty, Ben was looking for ways to increase his capacity and stay flexible at a minimum of expense. THE PRODUCT When the Earth Buddies’ owners removed them from their boxes, they found a bald, but cute, humanlike head about eight centimetres in diameter. After soaking in water and sitting in a moist environment for a few days, the Earth Buddy sprouted a beautiful head of green hair. See Exhibit 1 for the before and after look. The owner’s creativity could be expressed through the hair’s styling. Earth Buddy sales had originally been through Toronto-area flower shops and gift stores, but as the product’s wide appeal began to be realized, distribution spread nationally through stores such as K-Mart, Toys R Us, and Wal-Mart. By mid-July, over 100,000 units had been sold in Canada, and exports had begun to the United States. THE PROCESS Earth Buddies were produced in a hybrid batch-flow process illustrated in Exhibit 2. Six filling-machine operators working in parallel produced the basic rounded shapes by filling pieces of nylon stocking with sawdust and grass seed. The operators placed the heads in plastic tote boxes that held batches of about 25 F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 2 9A04D019

heads. In another batch operation, an operator shaped the Earth Buddies’ eye glasses by wrapping plastic- coated wire around a simple jig composed of two short, vertically-mounted dowels. The remainder of the process was a worker-paced flow. Three moulding operators removed the heads from the tote boxes, and formed the nose and ears with the help of elastic bands. Next, two people working between the moulders placed the previously-formed eye glasses over the nose, and glued small plastic eyes inside the rims. Each shaped and assembled Earth Buddy was placed in a bin for the painter, who fashioned a red mouth with fabric paint before placing the head on shelving to dry. Drying usually took about five hours, but could take as long as seven hours during humid summer days. After drying, two packers placed the Earth Buddies in boxes, and then into cartons ready for shipment. Work in process inventory (WIP) prior to drying was not large. Typically about 250 heads were at various stages of completion between filling and painting, but sometimes WIP seemed to grow much larger. In an effort to analyse his capacities, Ben and his day supervisor, Bob Wakelam, estimated the time it took an operator to process and move an Earth Buddy through each step. The times were: filling — 1.5 minutes; moulding — 0.8 minutes; eyes — 0.4 minutes; eye glass fabrication — 0.2 minutes; painting — 0.25 minutes; and packing — 0.33 minutes. After allowing for unavoidable delays and rest periods, Ben figured that he could count on seven hours of production from each eight-hour shift. With weather forecasts calling for more hot, humid summer days, Ben wondered how his production capacity and WIP levels might be affected.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 3 9A94D019

Exhibit 1

THE EARTH BUDDY Source: Company Files

Exhibit 2

THE EARTH BUDDY

Mould, eyes and paint

Drying shelves

Pack

Paint

R aw

m at

er ia

l i nv

en to

Fill Eye glasses

Mould Eyes Mould Eyes Mould

Fill

Fill F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

S w

9A95D008 MCLEOD MOTORS LTD. Professor John Haywood-Farmer prepared this case solely to provide material for class discussion. The author does not intend to illustrate either effective or ineffective handling of a managerial situation. The author may have disguised certain names and other identifying information to protect confidentiality. Richard Ivey School of Business Foundation prohibits any form of reproduction, storage or transmittal without its written permission. Reproduction of this material is not covered under authorization by any reproduction rights organization. To order copies or request permission to reproduce materials, contact Ivey Publishing, Richard Ivey School of Business Foundation, c/o Richard Ivey School of Business, The University of Western Ontario, London, Ontario, Canada, N6A 3K7; phone (519) 661-3208; fax (519) 661-3882; e- mail [email protected]. Copyright © 1995, Richard Ivey School of Business Foundation Version: 2012-06-05

Sue, I just got the fourth quarter inventory reports. When we standardized those end shields last September you told me it would cut both our manufacturing costs and our inventories. Manufacturing costs are down all right, although not as much as I would like, but our inventory has gone up. Sue, it costs us 25 per cent more to keep that stuff around for a year and you know the warehouse guys are really cramped for space. I want you to look into it and let me know what is going on and what we should do about it by Monday. I leave on Tuesday and won’t be back until early February.

Sue Reynolds, plant manager at the McLeod Motors Ltd factory in Chilliwack, British Columbia, put down the phone and thought back to the decision implemented four months earlier in September and to which John Ingram, vice president of manufacturing had referred. MCLEOD MOTORS The Market In its plant McLeod made over 40 models of electric motors ranging from one-quarter to 10 horsepower. The plant was divided into three parts. One half was devoted to production; the other half was divided into a small office, and a warehouse used to store raw materials, work in process, and finished goods inventories and supplies. Production was in small batches. The production area of the plant was laid out by process; that is, the machines were grouped together by function into departments (drilling, milling, turning, assembly, etc.). The plant operated five eight-hour days per week. The company had a number of customers in the original equipment manufacturer (OEM) market, who used the motors as components in larger products, and also in the replacement market. Naturally, McLeod’s product mix changed over time as its OEM customers phased products in and out and made annual supply decisions. In general, however, the size of the OEM market was known well in advance through the annual F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 2 9A95D008

order process and was fairly easy to forecast beyond that. In contrast, the replacement market was less stable and harder to forecast, especially for relatively short time periods. McLeod had recently experienced a welcome sales boom which was affecting all areas of the company. Warehouse space was scarce and the machines were usually busy. Consequently, company personnel would be certain to question any change that would take machines “out of circulation.” As a start to the bidding process for new contracts, the company calculated the direct costs of its products, added overhead based on direct labour hours (at 300 per cent), and added a 20 per cent profit margin. The recent sales boom had led to profits which management considered to be “decent for a change,” but the business tended to be boom or bust. In addition, labour negotiations, that management thought might be difficult, were scheduled to begin within a few months. Consequently, control of costs was important to the company. The BN-88-55 End Shield During the previous year, a McLeod engineer had suggested that the company could reduce the number of different end shields (see Exhibit 1) it made and used as motor components from 36 to about five by making minor modifications. Ms Reynolds decided to propose a test with BN-88-55, a new end shield for the shaft end of some motors (see Exhibit 1). It would replace 15 others in one motor product line accounting for about 20 per cent of McLeod’s sales. Ms Reynolds expected that this move would allow longer production runs, lower inventories, and better service to and by the manufacturers, wholesalers, retailers, and repair shops that used and carried McLeod products. The BN-88-55 was machined from an aluminum die casting which came complete with several indentations and holes, including one to accommodate the shaft, and a hardened steel ring in the centre which acted as a bearing housing. The BN-88-55’s inner surface had various ribs to add strength, particularly around the holes and indentations. McLeod had to thread (tap) eight of the holes (two sizes), and finish the surfaces in contact with the bearings and the frame (see Exhibit 1) and the convex (outer) face which abutted on machines using the motors. When finished, the BN-88-55 was bowl-shaped (16.6 centimetres in diameter, 4.8 centimetres deep, and 0.5 centimetres thick) and weighed about 800 grams. Because of its function, almost all of the BN-88-55s were sold as part of complete motors to either the OEM or the replacement market. Each BN-88-55 casting cost $5.28 (net of recovered scrap) and took a total of 3.1 minutes ($0.77) of direct labour time to process. With the added overhead allocation of $2.31, the total cost of a BN-88-55 was $8.36. McLeod sold them for $10 to replacement market customers but, as noted above, usually sold them in completed motors priced at $150 to $300. The Production Line All production in the plant was in lots; parts were carried in standard bins by forklift trucks. The standard bins held a maximum of 1,248 BN-88-55s. Demand for the 15 end shields replaced by BN-88-55 had totalled 2,500 units per week, ranging from 58 to 375 per week with an average of 165. BN-88-55 was very similar to the 15 end shields it replaced and their production required the same basic steps. The differences were in the structural features of the casting, particularly hole size and location. Before production of BN-88-55 had begun, the end shields had been made in batches of about two weeks’ demand. Although there had been some variation between products, the direct labour processing time for the 15 original end shields had also averaged 3.1 minutes.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 3 9A95D008

Each machine had space for only two bins near it. On one side was a supply bin from which the worker took parts as he or she worked on a batch; on the other was a receiving bin into which the worker placed completed parts. Although the batches could not be divided (that is, only a single machine could work on a given batch at any one time) if necessary, more than one machine in a department could be scheduled for BN-88-55 production simultaneously. For each operation the inventory in a supply bin slowly decreased and that in a receiving bin slowly increased as work on the batch progressed. Each of the production operations (milling, drilling, etc.) involved a number of small steps (locating a piece in the supply bin, picking it up, inspecting it, aligning it in the machine, activating the machine, inspecting the part, putting it in the receiving bin, etc.). When working on a batch, a production worker carried out these steps on each part in the batch. After the worker finished the batch, he or she informed the supervisor, the supervisor informed the production control department, and the production control department issued a move order. Subsequently, a materials handler moved the full receiving bin to a work-in-process storage area in the warehouse, a materials handler brought a new supply bin of parts, and the two-person set-up crew set the machine up to do the next operation. A materials handler later moved the full receiving bin to the next operation in the sequence where it then became a supply bin for the next stage. After this series of activities had been completed, production could continue. It took only about 10 minutes to move a batch to or from the warehouse, and about 30 minutes on average to set-up and test a machine. However, based on their experience, the production control staff knew that, on average, a batch spent a total of 17 working days between operations: three working days in the warehouse between each pair of operations, and an additional working day for each operation waiting while the material handling and set- up personnel became available and completed their work, and moving between the work centres and the warehouse. After McLeod’s executive committee approved Ms Reynolds’ proposal, the engineering and production control staff drew up manufacturing plans for BN-88-55. The company would require about 2,500 BN-88- 55s per week to meet motor assembly and replacement part demand. This demand was the same as the total for the 15 end shields it replaced. BN-88-55 would require five separate operations normally performed in the sequence shown in Exhibit 2. Although tapping had to precede turning, the two tapping operations and/or the two turning operations could be interchanged. The production control staff decided on a lot size of 1,248 to maximize run length using full standard bins.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 4 9A95D008

Exhibit 1

SIDE VIEW DIAGRAM OF AN ELECTRIC MOTOR

1 Fan cover 5 Centrifugal switch 9 Frame 2 Fan 6 Capacitor protector 10 Rotor 3 End shield 7 Stator 11 Shaft 4 Bearings 8 Winding 12 Base

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 5 9A95D008

Exhibit 2

OPERATIONS SHEET FOR MOTOR END SHIELD BN–88–55

Operation

Department

Machine

Pieces per Machine

Hour1

Machines Available in Department

1. Tap four holes, concave face 2. Tap four holes, convex face 3. Turn convex face 4. Turn concave face 5. Inspect and finish

Drilling Drilling Turning Turning

Inspection

Drill press Drill press

Lathe Lathe

Work bench & hand tools

75 75 100 120 150

9 9 5 5

1Assuming a 40-hour week, an output of 2,500 units per week requires 62.5 units per hour on average, or just under 20 hours per work centre for each batch of 1,248. McLeod could vary the output rate by assigning two machines and/or workers in a department to work on more than one batch at a time. However, because each department also worked on jobs for other motor lines and for other parts of motors in addition to end shields, it was extremely unlikely that all of a department’s capacity would be working on BN-88-55s at any one time. 2The number of inspectors — no machinery involved.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

S w

9B02D024 QUINTE MRI David Wright and Kevin Saskiw prepared this case under the supervision of Professors Carol Prahinski and John Haywood-Farmer solely to provide material for class discussion. The authors do not intend to illustrate either effective or ineffective handling of a managerial situation. The authors may have disguised certain names and other identifying information to protect confidentiality. Ivey Management Services prohibits any form of reproduction, storage or transmittal without its written permission. Reproduction of this material is not covered under authorization by any reproduction rights organization. To order copies or request permission to reproduce materials, contact Ivey Publishing, Ivey Management Services, c/o Richard Ivey School of Business, The University of Western Ontario, London, Ontario, Canada, N6A 3K7; phone (519) 661-3208; fax (519) 661-3882; e-mail [email protected]. Copyright © 2002, Ivey Management Services Version: (A) 2009-11-30 On June 12, 2002, David Wright and his colleague, Kevin Saskiw, business development co-ordinators at Quinte MRI in Belleville, Ontario, were trying to decide what to propose regarding the magnetic resonance imaging (MRI) facility at Benton-Cooper Medical Center (BCMC) in Palmer, New York. Both men were frustrated and confused. Although the BCMC facility was only six weeks old, it already had a waiting list of 14 days for MRI scans. Because of this backlog, physicians had begun to refer their patients to competing MRI clinics. Dr. Syed Haider, Quinte MRI’s chief executive officer, expected Wright’s and Saskiw’s recommendations and action plan in two days. QUINTE MRI Quinte MRI, Inc. was a small (annual revenues of $1.5 million),1 but growing, international service provider specializing in medical diagnostic technologies, including MRI, nuclear medicine, ultrasound, computerized tomography (CT) scanning, bone densitometry, mammography and teleradiology services. The company helped design, install and operate scanning centres, and provided continued training and support for data interpretation. It maintained a variety of exclusive or partnership business arrangements with both fixed-site and mobile service turnkey operations. Quinte MRI’s equipment and components were from many leading manufacturers. Quinte MRI’s founder, Dr. Syed Haider, received his PhD in electron spin resonance and nuclear magnetic resonance from the University of Wales. After a short time as professor at the University of Guelph, he became a physics and chemistry teacher at Centennial Secondary School in Belleville, Ontario, in 1968. When he retired 30 years later, he started Quinte MRI. Haider firmly believed that the residents of small communities deserved the same level of health services as residents of large urban centres. However, MRI systems in small communities were rare. Haider’s first attempt to establish an MRI facility (in Belleville) was unsuccessful because Canadian regulations prohibited private-sector MRI. Thus, he turned to the Caribbean and the United States.

1All currency in this case is expressed in United States dollars. In June 2002, the Canadian dollar traded at about US$0.63.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 2 9B02D024

Quinte MRI had established facilities in five locations: the company headquarters in Belleville; a partnership arrangement with a radiologist in Laval, Quebec; and private MRI clinics in St. Louis, Missouri, the Cayman Islands, and Palmer, New York. With the exception of the Palmer facility, Quinte MRI held an interest of less than 20 per cent in each clinic. In June 2002, the company employed a total of about 20 people. Quinte MRI served three distinct client groups: 1. Hospitals seeking to outsource their diagnostic imaging services were particularly interested in service

reliability, access to the diagnostic equipment 24 hours per day, seven days per week and reasonable cost.

2. Physicians wanting to be partners in an independent diagnostic imaging centre saw cash flow, accessibility to the equipment and the strength of the relationship with their diagnostic imaging partner as major criteria.

3. Individuals wanting to operate their own diagnostic imaging centre, using Quinte MRI as a consultant in developing and carrying out the necessary steps to establish the clinic, wanted freedom from the hassles involved with establishing the business and were willing to pay a 10 per cent project development fee.

SCANNING TECHNOLOGY2 Various scanning technologies produced high quality images of the human body. The most obvious imaging technique was to use a camera to capture a visual image on photographic film. Although this technology was simple, it could be invasive, as surgery or probes were required for images of internal tissues, and it was normally limited to the wavelength range of visible light. Modern scanning began in 1895 with the discovery that tissues absorbed X-rays. Although X-ray technology was relatively easy to use and gave high-resolution scans, the rays were penetrating and potentially dangerous,3 and gave unclear images of some body features. They were particularly suited for examining tissue abnormalities, such as fractures, malignant tumors and respiratory diseases. The 1970s saw the first of an explosion in imaging techniques, all of which relied on computers to help gather and analyse scanning data in electronic form. Computerized tomography (CT) relied on a series of X-rays from various angles that were combined to provide a three-dimensional picture from which two- dimensional images from any angle and at any depth could be derived. In positron emission tomography (PET), the patient ingested a positron-emitting radioactive substance that could be monitored as it proceeded through the body. In the closely related technique known as single-photon emission computed tomography (SPECT), the ingested active component emitted high-energy photons. In ultrasound (US), sound waves were bounced off tissues or objects inside the body; the reflected sound waves were converted into an image. MRI relied on the fact that diamagnetic nuclei (those with magnetic moments) interacted with strong magnetic fields to create their own small magnetic fields. The induced fields were studied using variable

2Much of the material in this section was adapted from the Web site:

www.whitaker.org/94_annual_report/over.html, September 20, 2002. 3Although X-rays were potentially dangerous, the low intensity of the radiation and the short duration of typical scans had effectively eliminated the danger to patients. However, medical personnel faced a much higher risk, as they received repeated exposure to this radiation.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 3 9B02D024

frequency electromagnetic signals. At a certain frequency, the induced field resonated with the electromagnetic signal; this resonance was measured. Water comprised some 70 per cent of the human body, making hydrogen, which was diamagnetic and thus gave an MRI signal, the most common atom in living tissue. Although MRI did not involve the radiation danger of many other scanning techniques, it could heat up the tissues if the radio frequency was too intense. Also, because ferromagnetic materials — those containing iron, nickel or cobalt — interacted strongly with magnetic fields, people with screws, plates or other ferromagnetic materials such as pacemakers or metal fragments in their bodies could not be scanned with MRI. Doing so gave poor resolution scans and could be dangerous to the patient. The many types of scans were valuable and complementary because they relied on different physical phenomena and gave different information. Although X-ray scans differentiated among tissues based on their density, MRI differentiated based on the tissues’ water content. Whereas X-rays and MRI gave information about internal structures, PET, SPECT and US could be used to observe biochemical processes, such as metabolism and fluid flow, as they occurred. Active research continued in scanning technology and techniques. Although conventional MR machines were multi-purpose and expensive, many newer ones were smaller, cheaper, tailored for a particular part of the body and more patient-friendly, with reduced noise and open on one side to reduce the patient’s feelings of claustrophobia. Other research streams aimed to (1) improve the MRI image and scanning depth capabilities by modulating the frequency of the electromagnetic signal; (2) broaden the scanning technique to other diamagnetic atoms, such as carbon, sodium and phosphorus; (3) develop ways to monitor body processes with MRI, (4) combine two or more scanning techniques; and (5) expand the ways in which these technologies could be applied. Image quality depended critically on the strength of the magnet and the time required to produce the image. In 2002, the newest generations of magnets approved for clinical use were 3.0 Tesla, whereas the previous standard had been 1.5 Tesla and 0.7 Tesla for the closed and open MRI unit, respectively.4 Exhibit 1 shows a photograph of a 1.5 Tesla short-bore MRI system. A typical exam took from 30 to 45 minutes, although some exams could be completed in 10 minutes. MRI had become increasingly popular within the medical profession. In 1998, an estimated 11.9 million MRI procedures had been performed in the United States; by 2001, this number had risen to 18 million procedures. In addition to growth in the number of scans, the number of hospital and non-hospital scanning sites had risen from 4,490 in 1998 to 5,550 in 2001.5 MRI equipment represented a significant investment. In 2002, the approximate cost of an MR machine was $1.5 million to $3.5 million. In addition, the facility required space6 and the equipment required shielding from magnetic fields. Installation, including shielding, cost $250,000 to $500,000 depending on the extent of renovations required. The typical reimbursement from United States insurance companies was $700 per scan. Exhibit 2 shows operating costs, which Quinte MRI’s managers believed were conservative, for a typical MRI facility.

4By way of comparison, a 1.0 Tesla magnet had a magnetic field about 20,000 times stronger than the Earth’s natural magnetic field. 5Van Houten, Ben, IMV Census Shows MRI Growth, Decisions in Imaging Economics, 15 (8), August 2002, page 8. 6For example, a model facility proposed by General Electric occupied 167 square metres gross and 154 square metres net.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 4 9B02D024

BENTON-COOPER MEDICAL CENTER Benton-Cooper Medical Center was a private, not-for-profit, 144-bed community hospital and regional cancer centre that provided primary care to the nearly 16,000 residents of Palmer and regional services to the 118,000 people in Adelaide County, which was in a largely rural area. BCMC had an active medical staff of more than 40 physicians, representing over 20 specialties. Although Creston, another Adelaide County community of 19,000, about 40 kilometres from Palmer, had two 200-bed hospital facilities with MR machines, BCMC’s administrators believed that there was an opportunity to compete successfully with a third MR machine. The primary reason for this view was that there appeared to be enough demand — in the United States the annual scan rate was approximately 68 per 1,000 people and the cancer rate in Adelaide County was somewhat higher than the national average. Second, the administrators anticipated that overall demand for MRI scans in Adelaide County would continue to grow at approximately 15 per cent per year. However, they recognized that the number of scans depended critically on the number of doctors practising various specialties. Because the MRI centre would get referrals from the hospital doctors and promotional support for advertisements with the local print and radio stations, the administrators believed that they would be able to generate sufficient volumes for their own fixed MR systems. In conjunction with the hospital administrators, Quinte MRI staff developed the monthly demand forecasts shown in Exhibit 3, which reflect seasonality owing to doctor vacation schedules and statutory holidays. And finally, the administrators were concerned that if they did not have an MR machine, BCMC would become a second-rate hospital. During the winter of 2001-02, BCMC decided to replace its MRI service provider because the medical centre wanted to increase the number of days of operation beyond the current two days per week. As they searched for a replacement, the administrators became aware of Quinte MRI’s impressive capabilities, such as availability for 24 hours per day and seven days per week, and Haider’s integrity and personal attentiveness. In February 2002, BCMC’s chief executive and board approved the outsourcing of MRI services to Quinte MRI. The agreement specified that Quinte MRI would own 100 per cent of the MRI centre, including imaging equipment, and would be responsible for most of its operation and management, including the hiring and salary of MR technologists to conduct the actual procedures. Quinte MRI would bill the hospital on a fee per scan basis. In the negotiation process, the anticipated average revenue was adjusted based on the expenses that would be covered by BCMC. The hospital would pay the salary and expenses of the radiologist, who would analyse the MRI scan and report the results. The hospital would also schedule the MRI clinic. It would charge Quinte MRI $140 and $5 per scan, respectively, for these two activities. The imaging suite was housed in a trailer connected to a hospital corridor. The other required functions were housed inside the hospital, some distance from the scanning suite. Exhibit 4 shows a layout of the radiology department. The MRI clinic began operations on May 1, 2002. At the hospital’s request, Quinte MRI leased one 1.5-Tesla General Electric (GE) short-bore high-speed MRI system, as shown in Exhibit 1. Although the rated capacity of the machine was two patients per hour, the actual number of scans in any period of time would depend on the types of exams being performed. For example, as shown in Exhibit 5, an abdominal MRI scan without contrast was projected to take 30 minutes, whereas an abdominal scan with contrast was projected to take 45 minutes. Contrast, which provided a more detailed image, was usually required in about 25 per cent of scans. F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 5 9B02D024

THE SCANNING PROCESS To receive an MRI scan at BCMC, patients first had to receive a referral from their doctor. The scanning process commenced when the patient or doctor’s assistant contacted the MRI scheduling department to arrange an appointment. Although the expected lead time for referred patients was 48 hours, some patients, called “walk-ins,” required a scan that day. When the scheduling department received a call, the receptionist wrote the patient’s name and type of procedure on a form with eight time slots, each for a one-hour increment. Exhibit 6 gives the schedule for June 12. Upon arrival at the MRI clinic for their appointment, patients checked in with the receptionist at the front desk and waited for the MR technologist to escort them to the MR machine in the magnet room. Some patients had difficulty walking or were confined to stretchers or wheel chairs. As the patients were escorted, the MR technologist asked questions to determine whether there were any health reasons that would prevent the patients from having an MRI. Patients who indicated possible health risks were further tested. The technologist took approximately five minutes to pick up the patient and determine if there were health conflicts. Patients not fit for the MR test were sent home. In such cases, the machine sat idle. During the first month of operation, an average of 1.2 patients per day were rejected for these reasons. In addition to checking possible health risks, the MR technologist ensured that patients were not wearing clothes with metal components. If the clothes had metal, the patient was required to change into a hospital gown at the change room, which took an additional four minutes, on average. Approximately 25 per cent of patients were in this category. Once in the magnet room, the MR technologist took one minute to provide a brief orientation and verify the paperwork. Patients would lie on a movable bench protruding from the bore, or tunnel, of the MR machine. A surface coil was positioned around the part of the patients’ anatomy of interest, such as the head, and the patients were then moved into the bore where the scanning began. It took approximately four minutes to position the coil and move the patient into the bore. The MR tunnel was relatively small, dark and noisy, which caused a feeling of claustrophobia in some patients. In addition, during scans it was important for patients to remain as motionless as possible. The MR technologist was responsible for conducting a set number of procedures to obtain the images requested by the referring physician. These procedures took a specific amount of time that was easily measured and consistent. For a 30-minute scheduled MRI scan, the actual time in the MR tunnel was 16.5 minutes. While the scans were in progress, the MR technologist sat in the tech room and entered the patients’ information into the hospital information system so that the patients could be tracked. Data entry took one minute, on average. Upon finishing the MRI scan, the MR technologist printed the MRI films and removed the patient from the machine. The technologist then took two minutes to escort the patient back to the front desk, stopping at the change room, if needed, for approximately four minutes. At the receptionist’s desk, the MR technologist checked off the patient’s name on the log to confirm that the task had been completed. Then, the technologist greeted the next patient. Throughout the day, the receptionist printed the confirmations and reports for billing purposes. Because each patient required between four and 16 sheets of film per MRI scan, averaging eight sheets, and it usually took 45 seconds to print each sheet, the MR technologist waited until after the fifth or sixth patient before collecting, sorting, labelling and then transferring the film to the radiologist’s office on his or her way to pick up another patient. The radiologist took approximately five minutes to read the patient’s film and dictate a diagnosis into a recorder. The dictation was transferred electronically to the transcription F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 6 9B02D024

department, where it was typed. The transcription department was located in a building adjacent to the hospital. One to three hours after they received the transcription, the transcription department returned the typed diagnosis to the radiologist for final approval. About every two hours, the radiologist verified and signed a group of transcriptions as a break from reading images. Once approved, the signed transcriptions and MRI films were transferred to the scheduling department, where a copy of the signed diagnosis was printed. The original transcript report and the MRI films were attached to the patient’s files, and together they were sent to the basement for filing and storage. The copy of the transcription report was sent to the referring physician. IMPLEMENTATION ISSUES Now that the BCMC MRI clinic had been in operation for six weeks, Haider was becoming increasingly concerned about its performance. The MRI clinic was not meeting promises made by Haider and GE to scan patients at a rate of two per hour. The hospital’s administrators continued to complain about the MR machine’s low productivity, the strain resulting from the MR technologist’s heavy overtime schedule, and the loss of patient referrals from doctors within the hospital and in the surrounding community. Doctors expected to receive the transcription report within two days of their request. BCMC, Quinte MRI’s customer, was dissatisfied because the backlog had exceeded 14 days by early June and doctors had begun to refer patients to competing clinics to obtain more timely MRI scan results. On June 11, 2002, Haider asked Wright and Saskiw to address the problems. Wright and Saskiw were halfway through the two-year honors business program at the Richard Ivey School of Business, at The University of Western Ontario, London, Ontario. Both of them were seeking challenges in entrepreneurial environments and wanted to avoid positions in large corporate environments, which limited business exposure and responsibility. They viewed the opportunity of summer jobs at Quinte MRI not only as being consistent with this career goal, but also as an opportunity to assist Wright’s long-time family friend, Haider, by applying some of the tools they had learned. Although none of Quinte MRI’s employees had a job description, Wright and Saskiw understood that, as business development co-ordinators, their job was to establish new relationships with doctors and investors, review existing operations and make and implement recommendations to improve operations. MR TECHNOLOGIST Before operating an MRI machine, most MR technologists had earned a two-year degree in radiological technology. If the technologist planned to work solely with MRI, the minimum education requirement was a one-year MR technician diploma. In upstate New York, MR technologists earned approximately $32 per hour; MR technicians earned about $25 per hour.7 Employee benefits typically added an additional 20 per cent to salary figures. After earning a degree and finding employment, new MRI technologists were typically trained by their employer on its MR systems for about three weeks. Jeff Sinclair, BCMC’s sole MR technologist, was scheduled to work 40 hours per week, Monday through Friday, 7:30 a.m. to 4:30 p.m. The first half hour of each day was occupied with setup and debugging of the equipment, called “phantom scanning.” During May, Sinclair had worked an additional 40 hours at a rate of 1.5 times his regular hourly wage. Although the MR machine was scheduled for one scan per hour, 7As a comparison, in 2002, the United States Department of Labor established the minimum wage rate at $5.15 per hour. In upstate New York, an assistant for an MR technologist would earn about $10 per hour.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 7 9B02D024

it was not meeting that rate (see Exhibit 7). When Wright asked Sinclair about his productivity, Sinclair responded:

Due to poor communication between the patient and the scheduling department, many patients fail to show up on time or cancel their appointment at the last minute. At the other extreme, patients experience frequent delays at the clinic. Some wait as long as an hour before I can start the MRI scan. I alternate between sitting on my butt for an hour or two to running around frantically attempting to placate angry and impatient customers.

I’ve got to deal with a lot of mistakes in the scheduling department. Patients are booked at the wrong times and they aren’t being screened properly. I’m getting patients that shouldn’t receive an MRI — but they are scheduled and I have to deal with them. Since they had to take a day off work, they get angry when I send them home. And I’m sitting here twiddling my thumbs! The scheduling department really causes me a lot of headaches. They write down that I’m supposed to do scan A, but when the patient gets here, the form says do A and B. Another time there were only three appointments scheduled for a day, and the scheduling department thought the day was full because they couldn’t understand what other people had written on the form. The previous MRI provider handled all of the scheduling. Now, however, the scheduling department is expected to buck up and cope with the additional workload.

In addition to the scheduling department, I’ve got to put up with the radiologist. He wants the images right after each patient is scanned. There is no way I can do that. It takes way too much time. I do it when I have a slow moment.

I’ve been putting in a lot of overtime since I started here and, to be frank, I am getting sick of it. The money is nice, but I have a family and my son is experiencing some medical problems. I need to be there for him. I really don’t want any more hours.

Things are improving a bit, though. I was originally trained on equipment from GE, but during May, the clinic used a Siemens unit. It took me a while to get used to it. Now, we’ve got our GE equipment and I’m much happier with it.

Monica Zimmerman, manager of the radiology department, was concerned that Sinclair was working too hard and for too long. She was pressuring Wright and Saskiw to hire another MR technologist to alleviate Sinclair’s load and improve the lead-time. She believed that the most appropriate move would be to add a partial second shift in the late afternoon and early evening hours. In considering this option, Saskiw said:

Hiring another MR technologist is a big decision for Quinte MRI. While Jeff worked a lot of overtime in May, he hasn’t worked much overtime yet in June — even though we are doing more scans. Hiring another MR technologist would increase our costs, since we would have to pay $38 per hour plus benefits for someone to come in for the second shift, and it might mean more idle time. It would alleviate the problem of allowing Jeff vacation time, or leave for illness or other extenuating circumstances. As it is now, it would take a while to get an MR technologist through a temporary employment agency specializing in medical personnel, and we would have to pay at least $60 per hour. In addition, using that source would eliminate our control of quality. The bottom line, though, is that we lose F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 8 9B02D024

over $6,000 in revenue for every day we are down. Whatever the decision, two things need to be considered — Quinte’s relationship with BCMC, which is getting fragile due to the backlog, and providing quality patient care. There is a shortage of good MR technologists, especially in rural areas, so I think it will be virtually impossible to find someone competent enough who would be willing to work part time.

In attempting to solve the problems at BCMC, Wright was focused on trying to find the bottleneck. From his reading of The Goal, he remembered a boy named Herbie, who hindered his boy scout troop’s ability to reach its destination quickly during a hike. Wright commented:

Finding Herbie is our first order of business. I know that if we can find out where he is, then we can take the appropriate action and make him more efficient. We are committed to keeping things simple and moving quickly because we are working within such a short time frame. I know decisions have to be made and action taken, and we can’t sit around waiting for Herbie to find us. We need to hunt him down. But, where is he?

We are tossing around the idea of developing a pay-for-performance system for Jeff so that he has an incentive to work harder and faster. Jeff is our most valuable asset at the clinic and he needs to be treated that way. We need to find a way to maximize his consistency so we are able to maximize the number of scans performed. I know we can grow the number of patients that we can handle in an eight-hour shift. But, if we continue to follow the same process, we won’t be able to continue to grow.

I am worried since we have only two days to provide a detailed action plan on how to solve the problem. Haider expects us to identify the problem and outline, in detail, the steps that should be taken to solve the capacity issues. He also expects us to make any additional recommendations to improve the performance of the MRI clinic. I wonder what would make sense here.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 9 9B02D024

Exhibit 1

AN IMAGE OF A 1.5-TESLA SHORT-BORE, HIGH-SPEED MRI SYSTEM

Source: GE Medical Systems.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 10 9B02D024

Exhibit 2

TYPICAL ANNUAL OPERATING COSTS OF AN MRI CENTRE

Leases Equipment $240,000 Building 50,000 Salaries and wages Radiologists 140 per scan MR technologist 60,000 Support staff 30,000 Office manager 45,000 Other Medical supplies 50 per scan Insurance 15,000 Leasehold improvements 10,000 Utilities 15,000 Advertising 15,000 Maintenance 110,000 Miscellaneous unforeseen expenses 100,000 Total annual operating expenses $690,000 plus variable costs per scan

Assumptions: Revenue per scan $700 Number of referred scans per year 1,600 Number of walk-in scans per year 600 Operating days per year 250 Effective tax rate 25%

Source: Quinte MRI files.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 11 9B02D024

100

150

200

250

300

May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr

Exhibit 3

FORECAST OF SUSTAINABLE DEMAND FOR MRI SCANS AT BENTON-COOPER MEDICAL CENTER

Source: Quinte MRI files.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 12 9B02D024

Exhibit 4

LAYOUT OF THE RADIOLOGY DEPARTMENT

Note: This diagram is approximately to scale. Source: Hospital files.

Computer Room

Tech Room Hydraulic Ramp

Magnet Room

Change Room

Radiologist's Office

Scheduling Department

Waiting Room Reception Desk

Entrance

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 13 9B02D024

Exhibit 5

TIME SCHEDULED FOR TYPICAL PROCEDURES, IN MINUTES Procedure Time Procedure Time

Abdomen with contrast 45 Lumbar spine with contrast 30 Abdomen without contrast 30 Lumbar spine without contrast 30 Abdomen with and without contrast 45 Lumbar spine with and without contrast 60 Bone marrow 60 Magnetic resonance angiogram (MRA), chest 60 Brain with contrast 60 MRA, abdomen 60 Brain without contrast 30 MRA, head with contrast 30 Brain with and without contrast 60 MRA, head without contrast 30 Breast, bilateral 60 MRA, head with and without contrast 45 Breast, unilateral 60 MRA, lower extremity 90 Heart 60 MRA, neck with contrast 30 Chest-mediast with contrast 60 MRA, neck without contrast 30 Chest-mediast without contrast 30 MRA, neck with and without contrast 45 Chest-mediast with and without contrast 60 MRA, pelvis 60 Cervical spine with contrast 60 MRA, upper extremity 90 Cervical spine without contrast 30 3-D reconstruction 15 Cervical spine with and without contrast 60 Orb, face, neck with contrast 60 Lower joint extremity with contrast 30 Orb, face, neck without contrast 30 Lower joint extremity without contrast 30 Orb, face, neck with and without contrast 60 Lower joint extremity with and without contrast 60 Pelvis with contrast 30 Lower extremity with contrast 30 Pelvis without contrast 30 Lower extremity without contrast 30 Pelvis with and without contrast 45 Lower extremity with and without contrast 60 Tempero mandibular joint 60 Upper joint extremity with contrast 30 Thoracic spine with contrast 30 Upper joint extremity without contrast 30 Thoracic spine without contrast 30 Upper joint extremity with and without contrast 60 Thoracic spine with and without contrast 45 Upper extremity with contrast 30 Upper extremity without contrast 30 Upper extremity with and without contrast 60 Source: Jeff Sinclair’s estimates.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 14 9B02D024

Exhibit 6

THE SCHEDULE FOR WEDNESDAY, JUNE 12

Note: The patients’ names, phone numbers and doctors have been omitted for reasons of confidentiality. Source: Company files.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 15 9B02D024

Exhibit 7

DATA ON PERFORMANCE SINCE STARTUP ON MAY 1, 2002 Date Number of Scans Number of Patients Rejected Hours Worked1

May 1 (Wednesday)2 2 8.0 2 (Thursday)2 5 2 10.0 3 (Friday)2 5 8.0 4 (Saturday) 0 5.0 5 (Sunday) 6 (Monday) 8 14.0 7 (Tuesday) 4 4 9.0 8 (Wednesday) 10 11.0 9 (Thursday) 6 2 9.0 10 (Friday) 4 2 7.5 11 (Saturday) 12 (Sunday) 13 (Monday) 7 2 9.5 14 (Tuesday) 9 9.0 15 (Wednesday) 11 1 11.5 16 (Thursday) 9 2 7.5 17 (Friday) 6 1 8.0 18 (Saturday) 19 (Sunday) 20 (Monday) 10 2 9.0 21 (Tuesday) 12 1 11.5 22 (Wednesday) 11 1 10.0 23 (Thursday) 13 2 11.0 24 (Friday) 10 2 9.5 25 (Saturday) 26 (Sunday) 27 (Monday)3 28 (Tuesday) 10 8.0 29 (Wednesday) 16 12.0 30 (Thursday) 7 6.0 31 (Friday) 10 2 12.0 June 1 (Saturday) 2 (Sunday) 3 (Monday)4 0.0 4 (Tuesday) 7 3 7.5 5 (Wednesday) 12 12.0 6 (Thursday) 12 2 12.0 7 (Friday) 6 5.5 8 (Saturday) 9 (Sunday) 10 (Monday) 14 1 8.5 11 (Tuesday) 14 11.0

Source: Company files.

1Overtime was calculated based on weekly, not daily, hours worked. 2From May 1 to May 4, Sinclair was conducting application and hospital safety training, in addition to his MRI duties. 3May 27 was Memorial Day, a national holiday. 4On June 3, the clinic was closed to allow for the removal of the Siemens MRI equipment and the installation, testing and training on GE MRI equipment.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

9B14D011

RANGER CREEK BREWING AND DISTILLING Jorge Colazo wrote this case solely to provide material for class discussion. The author does not intend to illustrate either effective or ineffective handling of a managerial situation. The author may have disguised certain names and other identifying information to protect confidentiality. This publication may not be transmitted, photocopied, digitized or otherwise reproduced in any form or by any means without the permission of the copyright holder. Reproduction of this material is not covered under authorization by any reproduction rights organization. To order copies or request permission to reproduce materials, contact Ivey Publishing, Ivey Business School, Western University, London, Ontario, Canada, N6G 0N1; (t) 519.661.3208; (e) [email protected]; www.iveycases.com. Copyright © 2015, Richard Ivey School of Business Foundation Version: 2015-02-03

In September 2013, the owners of Ranger Creek Brewing and Distilling (Ranger Creek) — Mark McDavid, Dennis Rylander and T.J. Miller — met at their manufacturing facility in San Antonio, Texas. Miller was going to present some potential operational scenarios for the company for the 2014 to 2019 period. Now in its third year of operations, Ranger Creek had so far been a success. Its beers and bourbons had won several tasting awards and the financials were meeting expectations, but the managing team was starting to feel acute growing pains. Before the owners could even focus on growth plans, there were some unanswered questions they needed to tackle: Considering the current production records, what was the maximum attainable capacity of the plant? Would brewing capacity need to be expanded and, if so, what equipment should be purchased? Should the focus be placed on automating the brewing or filling process? And what were the operational consequences of the recently passed beer laws in Texas? Miller opened the meeting by announcing that he was going to present a few different operational scenarios for the 2014 to 2019 capital and manpower budget that would address all these concerns. Since moving the entire operation to a bigger facility had been ruled out as too costly, Miller was going to present other alternatives to increase capacity, including buying additional or bigger equipment, redesigning the plant’s layout and renting outside storage space. RANGER CREEK BREWING AND DISTILLING In the early 2000s, Miller, Rylander and McDavid met while working at the same employer. They shared a passion for beer making, and soon they started homebrewing together. They quickly realized that San Antonio was the seventh-largest city in the United States, with a rich brewing heritage but no sizeable microbrewery. Drawing from their time at business school, they started writing a microbrewery business plan.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 2 9B14D011 They also figured that proud Texans would love to drink a whiskey made in their home state. Plus, they argued, “The craft distilling movement is about to explode.” Thus, they started writing their microdistillery business plan too. All three founders of Ranger Creek were passionate homebrewers and had attended numerous workshops, trade fairs and brewing competitions, winning several awards, including a gold medal at the Dixie Cup, one of the most prestigious competitions for homebrewers in the United States. McDavid had a BA in marketing from Texas A&M and an MBA from the University of Texas at Austin. He had developed his business skills in the financial services area and founded several small businesses before joining Ranger Creek. Rylander hailed from Sweden, where he had commanded a military communications unit and had been a FIDE master, having won several regional and national chess competitions, including two Swedish junior chess championships. He held a BA in marketing from the University of Texas at Dallas and a MBA from the University of Texas at Austin. Prior to co-founding Ranger Creek, Miller served eight years in the U.S. Army, including two combat tours, and got a degree in political science and a MBA from Vanderbilt, with a concentration in marketing and entrepreneurship. With their own savings, funds from friends and relatives, the help of a few angel investors, and a bank loan, the three owners raised about US$1,000,0001 for startup costs. McDavid would manage marketing, distribution and sales, Rylander would deal with accounting and regulatory issues and Miller would be in charge of operations (see Exhibit 1). In October 2010, they started brewing and distilling in a 7,500- square-foot facility located in a light industrial sector of northeast San Antonio. Manufacturing used 6,000 square feet of brewing equipment (see Exhibit 2), most of it imported from Europe (see Exhibit 2). About 1,000 square feet of the facility was reserved for a “front room” where the company hosted tastings and educational and promotional functions for the public. PRODUCTS Beer In 2012, 200 million (M) barrels of beer were sold in the United States, of which about 12.5 M barrels were sold in Texas. Craft beer sales in the United States were 11 M barrels in 2011 and 13 M barrels in 2012. National brewers included Anheuser-Busch (the largest brewer in the United States), SABMiller and the Molson Coors Brewing Company. Imports had made significant inroads into the U.S. beer market, with a market share of about 15 per cent. In 2012 alone, 310 microbreweries and 99 brewpubs opened for business in the country. As of March 2013, there were 2,360 U.S. craft breweries in the United States, of which 1,124 were brewpubs.2 Fewer than 100 companies in the craft beer segment retained the status of regional microbrewer (sales of more than 15,000 barrels per year). This segment included Samuel Adams, Yuengling, New Belgium and

1 All currencies are in US$ unless otherwise stated. 2 The New Brewer, May/June 2013, magazine of the Brewers Association. Distributed to members only.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 3 9B14D011 Shiner. Samuel Adams, the largest of this group, commanded a market share of 0.8 per cent.3 The craft brewer segment had been growing at double-digit rates (15 per cent in 2012) and benefited from increased public awareness and a more favourable tax environment. Off-premises consumption of beer accounts three out of every four pints sold, but contributes to only half of the sales. The wholesale price for a keg of beer (15.5 gallons) was $130, while a pint of draft beer retails at about $4.50. Specialty or seasonal beers command a 20 per cent premium over regular ones.4 Ranger Creek’s main line of beers had four products (see Exhibit 4). Oatmeal Pale Ale was an American- style pale ale with a sweet, toasty flavour imparted by oats and had an alcohol by volume (ABV) of 5.8 per cent. Lucky Ol’ Sun was named after a traditional working man’s song and was a light “patio beer” brewed with pilsner malt. Light-bodied and dry, this brew had a 5.5 per cent ABV. La Bestia Aimable (“the friendly beast”) was a strong, Belgian-style dark beer made with local Texas honey, with a hefty 9.4 per cent ABV. Mesquite Smoked Porter was a bottle-conditioned dark beer with a 6.4 per cent ABV. For this beer, the grains were smoked before they were mashed. In addition to the main line, Ranger Creek offered seasonal and experimental brews, such as Strawberry Milk Stout, which was made with strawberries from Poteet, Texas, that were sliced by hand and added at two separate times during the brewing process to ensure that the taste and aroma of the strawberries could be appreciated in the finished product. Some of these beers were eventually moved to the main line such as had been the case with Lucky Ol’ Sun (see Exhibit 4). Spirits Following a constantly increasing trend (see Exhibit 5), in 2011, sales of spirits in the United States were 38 M gallons, of which 4 per cent was sold in Texas. Of the 38 M gallons, 16 M were bourbon. Roughly 60 per cent of spirits were sold on-premises (in bars, restaurants, etc.), while the rest was for off-premises consumption, such as sales at liquor stores for home consumption. The American whiskey market was dominated by two players: Beam Global recently acquired by Suntory of Japan (Jim Beam, Maker’s Mark) and Brown-Forman (Jack Daniel’s). They accounted for 55 per cent of premium whiskey sales. Stranahan’s and Anchor Distilling were some of the biggest microdistilleries, a segment rapidly growing but too small to make a dent in the market share. A handful of microdistilleries had been opening recently in Texas, such as Balcones Distilling in Waco, with its award-winning single malt. Ranger Creek believed that an important opportunity existed to create the “Texas Whiskey” category, following the very successful story of Tennessee Whiskey. In the spirits area, Ranger Creek’s main product was the award-winning Ranger Creek .36 Texas Bourbon Whiskey, a traditional sour mash bourbon with a flavour that had hints of vanilla, brown sugar and maple syrup. Ranger Creek .36 took its name from the 0.36 calibre pistol that the Texas Rangers used to carry. This product accounted for half the whiskey production at Ranger Creek. Ranger Creek Rimfire Mesquite Smoked Texas Single Malt Whiskey was based off of Mesquite Smoked Porter beer distilled into a whiskey. Rimfire took its name from the type of ammunition used by the Winchester 1866 “Yellow Boy” repeating rifle used by many frontier Texans. It could be said that rimfire was a smoked whiskey that used Texas mesquite wood to smoke the malt instead of using Scottish peat. 3 Ibid. 4 Ibid.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 4 9B14D011 Rimfire constituted 40 per cent of the production mix at Ranger Creek. Another 10 per cent of the mix was made up of unaged and experimental whiskeys that were sold in limited quantities. THE BEER BREWING PROCESS AT RANGER CREEK The oldest written recipe for brewing beer appeared in a clay tablet written in cuneiform some 6,000 years ago in Mesopotamia (currently Iraq). Beer is one of the oldest produced beverages known to mankind, and it was routinely enjoyed by pharaohs in ancient Egypt. Brewing beer involves breaking down a source of starch into sugars, which are then fermented by yeast, a unicellular fungus. The two main products of yeast fermentation are alcohol and carbon dioxide (CO2). Malted barley is the predominant source of starch in beer making. Grains are malted when they are soaked in water and allowed to partially germinate, and then dried and toasted, developing enzymes that help break down the grain’s starches into a complex mixture of molecules with a characteristic flavour and aroma. In general, beers do not use a single kind of grain but a combination of several malted and sometimes unmalted grains. The bitter and aromatic flavour in beer is provided by the addition of hops — the dried flowers of the hop vine, humulus lupulus. Other ingredients are added in different parts of the process to flavour the beer, such as dried or fresh fruits, herbs or seeds. At Ranger Creek, the grains were bought malted, and the first process was to “mill” or crack the grains in a mechanical mill. There was one mechanical mill at Ranger Creek, shared with the whiskey-making process, and it had a capacity of 1,800 pounds of grain per hour. A batch of beer took, depending on the final ABV of the product, between 1,500 and 2,000 pounds of malted grain per batch. The milling of a batch took about one hour for a 1,800-pound load. Since the milling process released some amount of grain dust, to prevent contamination the milling process was isolated from the rest of the plant and enclosed in a milling shed. The milled grains were loaded into a hopper at the milling shed, from where they were transported by means of an auger to the plant’s beer mash tank. The mash tank is a stainless steel tank equipped with a false bottom to later strain the grains from the liquid, producing 930 gallons of liquid per batch. In the mash tank, the malted grains were mixed with hot water and allowed to steep for around one hour so that the starches decomposed into fermentable sugars. After steeping, the liquid part of the mash was transferred through the mash tank’s false bottom to another tank, the boiler kettle, and boiled to deactivate further enzymatic action. The transfer of the 930 gallons to the boil kettle took an hour. The spent grain or “draff” was rinsed with enough water to assure an output of 930 gallons, which was transferred to the boil kettle. The draff was donated to local farmers to be used as animal feed. Cleaning the mash tank took one and a half hours. During the boiling, hops and other flavouring agents were added. Boiling not only deactivated further enzymatic action, but also sterilized the liquid, called “wort,” which was necessary to prevent microorganism competition with the yeast. After one hour, to get the mix to boiling temperature, the boiling proceeded for an additional hour and then the liquid was cooled down to fermentation temperature by means of a heat exchanger in about one hour. Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 5 9B14D011 After the wort cooled down, it was transferred to one of the fermentation tanks, where yeast was added, and the wort would spend about two weeks fermenting until the alcohol content was correct. After the fermentation took place, the mixture less all the dead yeast and other debris collected at the bottom of the fermentation tank was transferred to the 930-gallon conditioning tank. In the conditioning tank the liquid would settle for a week, separating sediment or “dregs” at the bottom and the clear beer on top. Conditioning took about a week, and fermentation tanks could be used for conditioning, but conditioning tanks could not be used for fermentation. From the conditioning tank, the beer layer was sent directly to packaging after correcting the CO2 levels by injecting commercially bought CO2, and was sent to be poured in either bottles or kegs. The conditioning stage could be avoided if instead of letting the fermented beer settle to the bottom of a tank it was filtered with a plate filter, which would cost around $8,000. Filling and labeling one batch of beer took about eight hours, with one hour for setup tasks, and if the beer was filled in kegs, it would take two hours per 930-gallon load in addition to the eight hours. Cleaning kegs prior to filling a batch was a manpower-intensive task that took a worker about three days. Automatic keg-cleaning equipment would do it in half a shift, but would cost $20,000. Any tank fillings had to be preceded by tank cleaning and sanitization, which took about an hour to complete and was a manual task. The price of select equipment can be found in Exhibit 6. THE WHISKEY DISTILLING PROCESS It was a natural choice to distill whiskey in addition to brewing beer, since “you have to make beer first if you want to get whiskey.” Whiskeys have different names depending on the composition of the grain mash. Bourbon whiskey, the kind made by Ranger Creek, has to be made from mash that consists of at least 51 per cent corn. Other kinds include corn whiskey (at least 80 per cent corn), malt whiskey (at least 51 per cent malted barley) and rye whiskey (at least 51 per cent rye). The milling batch size for a case of whiskey was 1,600 pounds, and mashing proceeded similarly to beer brewing, but since the mash was heated at higher temperatures, it took four hours to be ready, after which it took two and a half hours to cool off to fermentation temperature. After the mash was cooled down, it was transferred to the fermentation tank, inoculated with yeast and fermented for three days, after which the mixture of grain and liquid was boiled and distilled in a tall copper still whose column was filled with perforated plates. Legally, once the yeast was added and alcohol was formed, whiskey and beer products each had to be stored in dedicated that could not be shared with the other product’s process. This explains why at Ranger Creek bourbon had its own dedicated fermentation tank. Every distillation batch took around 15 hours. A vapour with high alcohol content was produced from the top of the still and cooled by room-temperature water to turn it back into a liquid state. The first portion of the distillate was called “heads” and was discarded. The middle portion was called “hearts” and was the bulk of the volume, and the last portion was called “tails” and was eventually recycled into new batches. The entire first distillation was called “low wines.” In the first distillation, 225-gallons were obtained from the original 930-gallon fermented mash, together with 65-gallons of heads and 75-gallons of tails. At Ranger Creek, the whiskey was distilled twice and the second time the distillation process took seven hours. Double distilling seemed to be the norm for bourbon production, although there were some Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 6 9B14D011 whiskeys with single or triple distillation. About 200-gallons of 140-proof alcohol were obtained after the second distillation. After the alcohol was distilled for the second time and the “cuts” were made, the hearts were diluted with water to 120 proof and filled in charred oak barrels for the maturation process. The barrels were placed on one of two shipping containers, where each container held 200 5-gallon barrels or 24 53-gallon barrels. In the maturation process, the temperature fluctuations during the day and night made the liquid seep into the oak wood and extract various aromatic compounds that yielded distinctive flavours. Though maturation took four to six years in Kentucky, smaller barrels and the heat of Texas made bourbon ready for bottling in about one year. When the bourbon was ready to bottle, it had lost about 30 per cent of its original volume due to evaporation. The lost portion was sometimes called the “angel’s share.” The bottling was done by a simple gravity tank, and took three hours to set up and six hours to fill 100 cases, with each case being 12 375-millilitre bottles. Obtaining manpower for labeling was difficult, because it presented a localized demand of several workers, and Ranger Creek’s strategy so far had been calling for volunteers who could donate four hours of their time to help label bottles and in return obtain a plant tour and Ranger Creek merchandise. REGULATORY FRAMEWORK Distribution of alcoholic beverages in the United States was organized into a “three-tier” system, where the first tier was made of the brewers and importers, the second tier was constituted by wholesalers or distributors and the third tier was the group of retailers. Brewers and importers sold their products to wholesalers, who in turn sold to retailers. This structure was enacted by law and manufacturers could not sell directly to retailers, with some exceptions varying on the state the transactions occur. Since 1995, the number of distributors had dropped from 5,500 to about 2,000 as smaller multi-brand wholesalers had been bought out. Rising costs as well as the need for scale and for broad, well-balanced portfolios had driven consolidation among U.S. beer wholesalers. Retailers sold products to the consumer either for on-premises or off-premises consumption. Each state had its own regulatory body overseeing alcoholic beverage industry operations within its borders. In Texas, this was the Texas Alcoholic Beverage Commission (TABC). TABC was responsible for approving distilling permits, collecting excise taxes and regulating the value chain separation among manufacturers, wholesalers and retailers. The federal excise tax was $13.5 per proof gallon of spirits and in the case of beer was $7 per barrel for the first 60,000 barrels if the producer manufactured less than two million barrels per year and was $18 per barrel otherwise.5 Of great relevance for Ranger Creek was that in June 2013 the Texas Legislature passed a series of bills that would potentially revolutionize the microbrewing and brewpub business. Committee Substitute to Senate Bill (CSSB) 515 raised a brewpub’s annual production limit to 10,000 barrels and allowed for a limited self-distribution permit for up to 2,500 barrels per year per licence or

5 Texas Alcoholic Beverage Commission, 2013.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 7 9B14D011 1,000 barrels per brewpub location, though a brewpub operator would have no limit on how much beer it was allowed to sell through a beer distributor.6 CSSB 516 and CSSB 517 raised the cap for craft brewers to 125,000 barrels per year and created a new self-distribution permit for craft brewers that brewed up to a total of 40,000 barrels. They also allowed out-of-state craft brewers to self-distribute equal to in-state craft brewers.7 CSSB 518 allowed craft brewers whose annual production did not exceed 225,000 barrels to sell directly to consumers for on-premises consumption up to 5,000 barrels per year, though no sales were allowed for off-premises consumption.8 CSSB 639 restricted a brewer from selling its territorial brand distribution rights and included language to mutually tie these five bills together (CSSBs 515 to 518 and 639).9 Senate Bill 642 allowed any companies holding a licence to package and sell food to buy distilled spirits directly from the manufacturer, which saved the companies the retailer markup.10 Under Senate Bill 905, Texas distillers could now sell their products directly to consumers when earlier the distillers were only permitted to sell through distributors.11 The passing of these laws was currently creating quite a stir in the brewpub/microbrewery/craft distilling community. Positioning as a microbrewer or brewpub had implications for business models. Microbrewers were considered manufacturers, and accordingly they had to locate in light industrial zoning areas, normally away from the customer traffic vital for attracting retail business. An important change related to the new laws was that from now on, craft brewers could not sell distribution rights to third-party distributors. Normally, a startup worked very hard to develop a network of clients, and that hard work used to be rewarded by monetary compensation when a third-party distributor entered into an agreement with the manufacturer. Since the passing of the new laws, such rights transfers could not be made for money. In late June 2013, Jester King and Hops & Grain, two microbrewers in the Austin area, decided to change their brewing licences into brewpub licences. Ranger Creek’s first promotional email after the passing of the new laws is shown in Exhibit 7. MARKETING, SALES AND DISTRIBUTION As part of its promotions strategy, Ranger Creek ran promotional events at restaurants and bars, and plant tours which were conducted every Saturday, while open-house events were held quarterly, to educate the public and build brand awareness. Ranger Creek was active on social media such as Facebook, and tried to be present at local events. All Ranger Creek products were supported by merchandising such as promotional material, t-shirts with the company’s logo and other items such as beer and whiskey glasses. Sales of beer and spirits were seasonal, with quarterly fluctuations (see Exhibit 8). 6 Texus state senate website, www.senate.state.tx.us, accessed December 2014. 7 Ibid. 8 Ibid. 9 Ibid. 10 Ibid. 11 Ibid.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 8 9B14D011 Beer was self-distributed by Ranger Creek to retailers in the San Antonio, Houston and Austin areas, and a third-party distributor was used for the Dallas–Fort Worth area. For the San Antonio/Austin areas, there were two salespeople who took care of client accounts, controlled spoilage and processed returns. One of them took the San Antonio area and the other took the Austin market. Additionally, there were five “tasting representatives” who organized promotions and tastings and made sure products were well presented and that taps were properly cleaned (including often cleaning the lines themselves). In 2012, Ranger Creek sold 3,500 barrels of beer and 2,800 cases of spirits (see Exhibit 9). Historically, for beer there had been a 4 per cent spoilage rate and about a 1 per cent rate of returns. Bottled beer was marked up about 25 per cent by retailers, while draft beer was sold at around $1 per pint to retailers, who sold it for about $5 per pint. By law, the distribution of spirits had to follow strictly the “three-tiered system,” where independent producers could sell their products only to wholesale distributors who then sold to retailers, and only retailers could sell to consumers. Ranger Creek used a wholesale distributor for its whiskeys, and there was one person who acted as a liaison with the distributor. Currently, a bottle of Ranger Creek .36 retailed for $35. A cost structure for both beer and whiskey is shown in Exhibit 10. CORPORATE SOCIAL RESPONSIBILITY Ranger Creek did its best to source raw materials locally, and to support the local community whenever possible. Corn for bourbon and other ingredients were sourced locally. For instance, strawberries used for Strawberry Milk Stout were sourced from Poteet, Texas, and the spent grain from the fermentation process was donated free of charge to local grass-fed cattle and pig farmers. The main environmental concerns were the conservation of water and energy, and waste management, both solid and liquid. The use of scarce water was a constant concern in Texas, and at Ranger Creek water was used not only as a component of the mash, but also to refrigerate the still. Currently, Ranger Creek was on a “one-through” system where it discarded the spent refrigeration water after using it only once. Growth prospects and changing laws made the decision more complicated on where to add capacity than originally expected by the company founders. Miller, however, was ready to discuss several alternatives to assure that Ranger Creek would not omit all of its potential.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 9 9B14D011

EXHIBIT 1: ORGANIZATION CHART

Source: Ranger Creek Brewing and Distilling and case author.

EXHIBIT 2: PLANT LAYOUT

Source: Ranger Creek Brewing and Distilling and case author.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 10 9B14D011

EXHIBIT 3: SELECT EQUIPMENT

Fermentation tanks Conditioning tank Distillation still Source: Ranger Creek Brewing and Distilling and case author.

EXHIBIT 4: BEER MIX

Beers Volume Share (%) Oatmeal Pale Ale 30 Lucky Ol’ Sun 25 Mesquite Smoked Porter 10 La Bestia Aimable 10 Seasonals 25

Source: Ranger Creek Brewing and Distilling.

EXHIBIT 5: SPIRITS SALES IN THE UNITED STATES

9-litre cases, thousands Year Value Premium High-end Premium Super Premium Total 2002 2,972 4,281 5,576 309 13,138 2003 2,972 4,278 5,823 332 13,405 2004 2,927 4,318 6,237 385 13,867 2005 2,816 4,388 6,666 431 14,301 2006 2,633 4,499 7,116 496 14,744 2007 2,619 4,415 7,310 568 14,912 2008 2,642 4,267 7,500 662 15,071 2009 2,808 4,367 7,231 658 15,064 2010 2,778 4,529 7,372 764 15,443 2011 2,717 4,637 7,782 907 16,043 2012 2,796 4,984 8,079 1,019 16,878

Source: Ranger Creek Brewing and Distilling.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 11 9B14D011

EXHIBIT 6: PRICES OF EQUIPMENT

Equipment prices (USD) Mash kettle 25,000 Brew kettle 20,000 Conditioning vessel 10,000 Grain mill 5,000 Still, complete system 150,000

Source: Ranger Creek Brewing and Distilling.

EXHIBIT 7: FIRST PROMOTIONAL EMAIL AFTER NEW TEXAS BEER LAWS

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 12 9B14D011

EXHIBIT 7 (CONTINUED)

Source: Ranger Creek Brewing and Distilling.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 13 9B14D011

EXHIBIT 8: SEASONALITY

Seasonality index Q1 Q2 Q3 Q4

Beer 1 1.2 1.1 1.2 Spirits 1 1.05 0.95 2

Source: Ranger Creek Brewing and Distilling.

EXHIBIT 9: SALES VOLUMES

Volume (barrels for beer and cases for spirits) 2011 2012 2013

Beer 1,000 2,000 3,500 Spirits 0 1,200 2,800

Source: Ranger Creek Brewing and Distilling.

EXHIBIT 10: COST STRUCTURE

% of Cost Beer Spirits

Raw Materials 48 49 Direct Labour 10 2 Utilities 3 7 Rent and Insurance 5 7 Overhead 13 20 Distribution 16 2 Promotions 5 12

Source: Ranger Creek Brewing and Distilling.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

S w

9A97D010 INTERNATIONAL DECORATIVE GLASS Jim Barker prepared this case under the supervision of Professors Robert Klassen and Paul Beamish solely to provide material for class discussion. The authors do not intend to illustrate either effective or ineffective handling of a managerial situation. The authors may have disguised certain names and other identifying information to protect confidentiality. Ivey Management Services prohibits any form of reproduction, storage or transmittal without its written permission. Reproduction of this material is not covered under authorization by any reproduction rights organization. To order copies or request permission to reproduce materials, contact Ivey Publishing, Ivey Management Services, c/o Richard Ivey School of Business, The University of Western Ontario, London, Ontario, Canada, N6A 3K7; phone (519) 661-3208; fax (519) 661-3882; e-mail [email protected]. Copyright © 1997, Ivey Management Services Version: (A) 2010-02-03 In June 1996, Delta, British Columbia, remained overcast and rainy. Frank Lattimer, vice-president operations of International Decorative Glass (IDG), mused that it really didn’t matter, as there would be little time for golf this year. Rapidly increasing demand for decorative glass panels by steel door manufacturers in the United States, IDG’s primary market, had its two production facilities in Delta and Shuenyi, China scrambling to keep up. Lattimer had been asked to develop a recommendation for capacity expansion for consideration by the board of directors. The board had emphasized the need to move quickly as sales were increasing faster than IDG’s ability to meet them. Although either existing plant could be expanded, IDG also had recently been approached about considering further off-shore sourcing in the rapidly developing country of Vietnam. Frank knew that any decision would have significant ramifications for the company’s long-term positioning and ability to meet its ambitious goals for growth. THE INDUSTRY Decorative glass panels typically are inserted into residential steel doors, and were increasingly being used by builders and home renovators to add architectural interest and a customized appearance to doorways (Exhibit 1). Growth in the industry was being fuelled by the general trend away from wooden exterior doors to steel doors. Forestry restrictions, lumber prices, energy efficiency and increasing criminal activity all contributed to the growing demand for retrofitting wood doors with steel replacements, often with decorative glass panels. In addition, the lower price of steel doors relative to the traditional wood door, with wholesale prices starting as low as Cdn$300, further eroded market share in new home construction. Decorative glass was now being incorporated into 10 per cent of new home construction. The total North American sales for decorative glass panels was conservatively estimated at $2 billion in 1995 (all figures are reported in Canadian dollars), and the market showed signs of continued strong growth. Industry experts predicted that annual sales could reach $4.5 billion in the United States alone, within five years. Canada’s weighting of the North American market was disproportionately high, at 15 per F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 2 9A97D010

cent, reflecting the somewhat earlier development of the market there for these panels. By 1996, panels were found in approximately 85 per cent of steel doors in Western Canadian homes. Manufacturers in Canada tended to be more vertically integrated than their U.S. counterparts, with plants fabricating both the steel door and the decorative panel. Locally, British Columbia’s supply capacity grew well past the sustainable growth rate during the late 1980s and early 1990s as new market entrants scrambled to ramp up production capability to capitalize on the residential construction boom. The result was steadily eroding margins, followed quickly by industry consolidation, with high cost producers closing or being absorbed by more competitive operations. In spite of these changes, Canadian industry continued to be characterized by oversupply, underutilized capacity and commodity pricing. Lattimer had recently completed a basic competitive assessment of several key Canadian competitors as part of IDG’s business plan (Exhibit 2). By contrast, U.S. manufacturers of decorative glass panels acted as original equipment manufacturers (OEMs) for large residential steel door fabricators and retail chains. The industry was quite fragmented, with the largest three producers in the United States each having less than six per cent of the total market. Unfortunately, information on these producers was limited (Exhibit 3). Manufacturers ship panels to predetermined central warehousing and assembly points where their panels are fitted into the steel doors and distributed by the door fabricators through their retail channels. In general, the U.S. marketplace demanded high quality, fast service and increasingly, low price. At this time, the United States, unlike Canada, was rapidly growing and underserved. In addition, Canadian manufacturers generally were about three years ahead of their U.S. counterparts in product functionality and design and, thus, able to develop strategic partnerships with steel door manufacturers. An undervalued currency also provided Canadian suppliers, such as IDG, with an initial competitive advantage. Combined, these factors created a significant market opportunity for any Canadian supplier who could meet rigorous quality standards and maintain a high level of customer responsiveness to design customized panels. Early attempts by Canadian firms to develop their export sales quickly revealed that a customer would pay only so much for quality, service and product differentiation, and price was becoming an increasingly important driver in the purchase decision. In response, manufacturers on both sides of the border began to source production of the glass panels at lower cost to facilities located abroad. Because labor represented a large portion of cost of goods sold, production was increasingly being moved to countries with low labor costs, such as Mexico, Thailand and China. At this time, only a few Canadian manufacturers had been able to address all of these challenges successfully. PRODUCTION OF DECORATIVE GLASS PANELS The production process for decorative glass panels was quite standardized, with little variation among firms and plants. As might be expected with a product that until recently was considered a “craft,” the process was very labor intensive, with the equivalent of up to two person-days required for each panel. Production equipment was generally quite flexible, and could be purchased from several suppliers. Decorative glass panels consist of multiple glass panes of different sizes, colors and grades assembled between soldered brass rods to form a decorative picture. The production of the panels used a multistep process that cut and formed the glass and brass components, and assembled the parts into sealed decorative glass units that could withstand the harsh exposure needed for exterior doors. F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 3 9A97D010

The manufacturing process began with the cutting of raw glass sheets of various colors and finishes into pieces of the precise shape and size needed for the final design. Some of these pieces were then bevelled to give a more attractive final appearance. The specialized cutting and bevelling of the glass pieces were the most capital intensive steps in the production process. In a separate area, brass rods were cut and shaped into segments that ultimately serve to hold and separate the glass pieces. The correct set of glass pieces and brass rods were grouped into panel-specific “kits.” These kits were assembled and soldered into predetermined patterns that formed semifinished panels. Several cleaning and touch-up steps followed. Next, clear solid glass panes were added to each side of the inlay, creating a “sandwich” that protected the more delicate decorative inlay. Swizzle, a sealant material, was added around the edge to insulate and protect the panel from water damage. The panel was then put through an automated sealing machine, washed and inserted into a frame. Finally, the finished panel was labelled and packaged for shipment. These operations typically were performed in small batches of panels. THE COMPANY Located near Vancouver, British Columbia, IDG was founded in 1984 by Michael Jeffrey, decorative glass designer and entrepreneur. Initially, the company started as an integrated manufacturer of steel doors and decorative glass panels, and IDG enjoyed modest prosperity through the 1980s as the housing market boomed in that province. During this period, numerous firms entered the market, hoping to share in the prosperity of the industry. As real estate development slowed and even stagnated in the early 1990s, and the competitive basis shifted to cost, Jeffrey realized that the company was losing money in their manufacturing of steel doors. He felt that IDG could significantly enhance profitability by concentrating exclusively on decorative glass panels. Jeffrey also recognized the need for a senior operations and business development person to make the operations more competitive in that market. Lattimer was hired in 1991 with the mandate to grow the international market, to improve cost efficiency, to set up a fully integrated management information system, and to create a corporate structure and culture that would support continued expansion. To meet these objectives, contacts and sales were further developed with several U.S. steel door manufacturers, the largest being Midwest. Lattimer also gained concessions in wage rates and flexibility in staffing requirements during collective bargaining with the union. Finally, a management information system, including materials requirements planning (MRP), was installed and brought on line to improve access to timely information and to raise customer responsiveness. Historically, IDG’s sales had been driven by custom orders for the glass panels. However, with recent efforts to increase sales volume, an increasing number of higher volume orders were being pursued, although often at much lower margins. In spite of labor concessions, high wage rates and limited flexibility continued to make IDG’s plant in Delta increasingly less cost competitive. To reduce production costs, Lattimer was forced to explore alternative, off-shore sources of production. CENTURY GLASS In January 1995, IDG began sourcing some of its high volume, low skill production through a strategic partnership with Century Glass, located in Shuenyi, approximately an hour’s drive outside of Beijing, Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 4 9A97D010

China. This manufacturing facility was developed solely to meet the production needs of IDG, although the actual plant was owned and operated by the father of a former employee, Jianwei (Jerry) Lo. Lo had returned to China to set up the joint venture with IDG. When IDG first arrived, the Shuenyi facility was little more than a deserted warehouse, situated across the highway from the village of 2,000 people where Lo had been born. The Lo family was well respected in the area, even though they came from modest means relative to Canadian standards. There was no electricity, telephone or plumbing in the village, and fresh water was unavailable. With minimal infrastructure in place, power requirements, communication and capital equipment challenges all needed to be addressed. Co-generation power supplies and inverters were supplied by IDG; satellite and cellular phones were used until Century received a land line (faxes were sent from Beijing in the interim). Basic production equipment needed to cut glass sheets and brass rods were sourced locally; however, one large panel sealing machine was imported from Korea. Practically everything else at the facility was built by the local workforce. Approximately one-third of the workers lived in four-person dormitory rooms located on the premises, and the production plant also included space for the workers to grow their own food in the courtyard. Family ties of the Los facilitated the shipment of goods, as Chinese bureaucracy was legendary. Jerry’s uncle was the police chief of the local district and, thus, extremely well connected politically; IDG benefited from the association. The movement of raw materials into China and finished goods out of China, via Tientsin to the Gulf of Chihli, was expedited through Jerry’s uncle. Because of differences in proximity to the market and cost structure, the Chinese production facility concentrated on producing high volume, low cost glass panels for IDG. These panels were then shipped in bulk to the Delta production facility for final processing, followed by packaging and shipment to U.S. or Canadian customers. The additional processing in Canada resulted in a change in product classification under the North American Free Trade Agreement (NAFTA), which allowed the finished product to be imported duty-free into the U.S. market. (By contrast, if complete, sealed panels were imported directly from China into the United States, a 60 per cent duty would apply.) For some customers, the standardized panels produced at Shuenyi were modified and further assembled at Delta to form larger, more complex, customized panels. By necessity, these arrangements required a long lead time, currently 18 to 20 weeks (Exhibit 4), well above that of the Delta plant, where lead times averaged one month. Initial start-up problems in 1995 centered on logistics and quality. Rather than allow IDG’s reputation for excellent customer service to suffer by missing delivery dates, orders of panels were, at times, air freighted to Delta from China, at an extra cost of $250,000 in the first year. These problems were gradually overcome as typical production lead times were reduced to their current levels. Low yields and high waste/breakage also plagued the start-up. However, as the skill levels of the local workforce improved, yields increased dramatically. By mid-1996, finished panel yields consistently surpassed 99 per cent, although in-process breakage and other losses remained a problem. Current status By June 1996, Century Glass produced 80 per cent of IDG’s panels, representing 60 per cent of revenues. The remaining somewhat more specialized, lower volume panels were produced by 70 employees in the F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 5 9A97D010

Delta plant. The Century plant was operating close to capacity, with approximately 100 employees producing 8,000 panels per month. Dorms were overcrowded and people were elbow-to-elbow in the manufacturing area. The joint venture agreement specified that IDG purchase all materials, own all inventories, and specify all finished product standards. The production arrangement with Century stipulated a fixed charge per employee and a variable cost per finished panel. Specifically, IDG paid $140 per employee, per month. In addition, IDG also paid Century a product transfer price of $4 for each panel that met IDG’s rigorous quality standards for finished panels. Employment levels could be varied as needed to match sales volumes. Employees worked seven, eight-hour days per week, every week. This was high by Chinese standards, where the five- or six-day work week was more common. By comparison, in Canada, unionized employees received $9.75 per hour, per 40-hour work week. Combined, these differences in labor translated into a significant cost advantage for Shuenyi, without accounting for the operational advantages of increased labor flexibility. Relative product costs are illustrated in Exhibit 5. Labor savings were offset to some degree by a higher working capital investment necessary to finance larger inventories and longer payment cycles. For example, inventory turnover at Century Glass was only two turns per year in 1995, whereas Delta averaged six. In addition, banks refused to finance or factor raw material and work-in-process (WIP) inventories located in, or in transit to or from, China as the risk of recouping funds in the case of insolvency was considered too high. This risk varied by country. Some developing countries, such as Mexico, were viewed as less risky, while others, such as India, offered government guarantees for export-oriented manufacturers. The Lo family was anxious to keep 100 per cent of IDG’s business at their facility. However, Lattimer was very concerned about having only a single supplier in China, where political risks were perceived to be significant for such a large portion of their production. For example, the repatriation of Hong Kong in 1997, adverse trade tensions and possible trade restrictions between China, the United States and Canada, all indicated that a move to establish another production source might have strategic and operational merits. FINANCIAL RESULTS IDG’s revenue growth had been impressive since 1990, increasing from $2.6 million to $5.4 million for fiscal 1995. Financial results for the last two years are summarized in Exhibits 6 and 7. Revenues were projected to reach $10.5 million this year, with 95 per cent of sales being made in the United States. As noted earlier, margins had eroded during the early 1990s as residential construction slowed, and competition increased. Sales levels had risen significantly in 1995 as new production capacity became available at Shuenyi. However, profitability fell as a result of poor initial yields and air freight shipment costs at this new plant. Looking forward, Lattimer expected margins to increase as productivity further improved in Shuenyi. Both Jeffrey and Lattimer strongly felt that the market for strong growth by IDG was there. IDG had already been turning away business as they struggled to meet existing customer commitments from their two production facilities. Current plans called for revenue growth to $30 million by the year 2000. Critical to achieving these long-term results was an increase in production capacity to match the forecasted sales volumes. F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

Page 6 9A97D010

This aggressive growth necessitated access to additional capital to finance investment in new capacity and additional working capital. In August 1995, IDG approached a venture capital firm, Working Opportunity Fund, for $2 million of equity financing. The structure of the investment was negotiated, due diligence conducted, and the deal finalized in November of that year. In addition, IDG paid down its line-of-credit from the bank by financing its inventory in China with a guarantee from Canada’s Export Development Corporation. This effectively reduced IDG’s investment in working capital and made the sourcing of manufacturing to Asian facilities increasingly attractive. Combined, these additional sources of capital enabled IDG to increase its operating flexibility, and further develop its presence in the U.S. market. CAPACITY EXPANSION Lattimer had narrowed the options for expansion of production to three alternatives. Expansion was possible at either existing plant. In addition, another strategic partnership could be developed in another low labor-cost country, similar to IDG’s earlier decision to expand into China. After exploring options in other developing nations with low labor costs, Lattimer, in consultation with senior management, had narrowed the candidate list of countries to one: Vietnam. This country offered a critical advantage in Lattimer’s mind over other developing nations: a potential local partner, Dan Kim. Kim’s firm currently supplied raw glass to IDG, and Kim had approached Lattimer about establishing a manufacturing joint venture. Expansion in Delta At this time, company-wide capacity could be doubled by investing a relatively modest amount of capital, $30,000, in the Delta plant. Labor costs would rise based on existing wage levels. Given the close proximity of this plant to the U.S. market, the existing production planning system could be further leveraged and customer responsiveness further improved. Expansion in Shuenyi Because production at the Shuenyi plant was already very tight, any expansion would involve a significant increase of middle management and support staff, and an expanded production planning system, mirroring the earlier MRP investment made in Delta. Existing arrangements for labor would be maintained, where IDG would pay a flat monthly fee per person, plus a variable rate per panel. Although some of the existing production equipment still had excess capacity, additional equipment would be needed. In total, an estimated capital investment of $30,000 would be needed in new cutting equipment to double company-wide capacity. Incremental manufacturing overhead costs would be approximately $150,000 per year. Direct labor costs would increase proportionately with production volumes. These costs did not include either a desperately needed new building or additional inventory carrying charges. Timing for ramp-up to this volume level would be approximately six to eight months. The most significant concerns with expansion at Shuenyi were related to further dependence on a single supplier and issues related to political risks associated with production in China. Trade uncertainties between China and the United States also aggravated long-term planning efforts. Management was apprehensive that existing tensions could escalate over any, or all, of repatriation of Hong Kong in 1997,

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 7 9A97D010

intellectual property rights (software piracy and patents), dissident protests, strained relations with Taiwan and a general trade imbalance. Smaller manufacturers that supply the U.S. market, like IDG, inadvertently have been punished by short- term high tariffs, customs delays and other non-tariff barriers. Although quite unlikely now, the worst case scenario would be a ban on importation from China. Unfortunately, because of the general income levels in China and construction norms, there was little local market for IDG’s products at this time, although it did look promising in the longer term. Foreign Operations in Vietnam Vietnam had only recently begun to exhibit the economic growth characteristic of other countries in Asia- Pacific. Like many developing countries, infrastructure at this time was terribly inadequate. Lattimer estimated that development was at least five years behind China, and conditions were even more challenging than those first faced by IDG when they established their joint venture in China. In recent years, Vietnam had been plagued by internal political problems, and foreign investors were apprehensive to invest. This situation now was beginning to change, as the United States had moved to reestablish diplomatic relations with the Socialist Republic of Vietnam in 1995. In turn, this thawing of the political climate had encouraged foreign investment which had grown rapidly as a result. Vietnam also had a strategic location for re-export to other markets in Asia. Although a Communist state, the central government had instituted the beginnings of “Doi Moi” or “open door” policy as early as 1986. The objectives of Doi Moi were to develop export-oriented production capabilities that create jobs and generate foreign currency, to develop import substitutes, to stimulate production using natural resources, to acquire foreign technology and to strengthen Vietnam’s infrastructure. Incentives offered included: the option to establish wholly-owned foreign subsidiaries; favorable corporate income tax and tax holidays; waivers on import/export duties; and full repatriation of profits and capital. With 75 million citizens and a labor force of 32 million, Vietnam had the second lowest wage rate in the Pacific Rim. Only about 11 per cent of the working population was employed in manufacturing, another 19 per cent in the service sector, and the remainder in agriculture. Inflation was high, at 14 per cent in 1995, partially because of the devaluation of the “new dong” as the government had allowed the currency to float in world markets for the first time. The primary industries of Vietnam included food processing, textiles, machinery, mining, cement, chemical fertilizers, tires, oil and glass. Vietnamese companies already supplied some of the standard glass and bevelled glass components used by IDG. Generally, the labor force was energetic, disciplined and hard working, although unemployment remained high, at 20 per cent. English and French were widely spoken but literacy was relatively low, at 88 per cent. Unfortunately, basic human rights and freedoms had received little attention. There was widespread conflict between local and central governing bodies, extensive corruption and exhaustive bureaucracy at both levels.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 8 9A97D010

Production of Decorative Glass Panels in Vietnam The State Committee for Cooperation and Investment (SCCI) identified seven areas of the Vietnam economy where foreign investment would receive preferential tax treatment. Of particular relevance to IDG, labor-intensive manufacturing was one such area. The SCCI would assist the new venture in whatever way they could, typically through the development of contacts with customers and suppliers, as well as guiding the investor through the government bureaucracy that approved any business venture. The Vietnamese government also had legislated five approaches for establishing a business venture in the country. Of the five, the international business community and the government widely favored the joint venture approach. Under this approach, a foreign firm such as IDG would sign a contract with one or more Vietnamese parties to create a new legal entity with limited liability. Foreign capital had to constitute at least 30 per cent of the new entity’s total capital. A foreign investor could then leverage the local partner’s contacts, knowledge of the local market, and access to land and resources. The Vietnamese had a saying: “Nhap gia tuy tuc,” which means “When you come into a new country, you have to follow the culture.” Clearly, identification of a strong local partner would be critical for meeting the cultural norms in Vietnam and ensuring the success of any investment by IDG; this had been a major obstacle for many other foreign firms. Lattimer saw many parallels with the earlier joint venture into China. That investment had succeeded largely as a result of IDG’s strategic partnership with Century Glass and the Lo family. IDG had been able to limit their investment risk to supplying capital equipment for the facility and inventories. By contrast, other decorative glass suppliers operating in China were paying higher costs, and making larger investments in plant and infrastructure. The partnership with Century also had provided IDG with additional political clout and allowed them to bypass much of the Chinese bureaucracy. One obvious choice for a local partner was IDG’s beveled glass supplier, managed by Dan Kim. Kim operated a glass plant in Da Nang, which was well under capacity, and had an oversupply of qualified labor. Kim had approached both IDG and government authorities and essentially paved the way for IDG to begin joint venture operations within a six- to 12-month time frame. Labor and product transfer prices were likely to be significantly lower than either the Delta or Shuenyi plants, with these costs being approximately half those of Shuenyi. Additional overhead costs were estimated at $50,000 annually. Finally, a significant investment would be needed in new equipment to reach the same, company-wide production volume possible with the other options (Exhibit 8). Lattimer wondered whether he might be able to extract more favorable terms for any joint venture relationship, such as shifting responsibilities for financing inventories to Kim. THE DECISION As Lattimer was putting together his proposal for the board, he reflected on a conversation he had with Jerry Lo last month. Lo had indicated that Century would soon expect their piecework compensation to increase from $4 to $7 per finished panel. While seemingly a small fraction of total production costs, Lattimer worried that further requests for increases would follow unless other alternatives were developed. He also was only too aware that with up to $1 million invested in inventory at Century at any given time, IDG was in a very precarious position. Single sourcing had given Century a level of bargaining power that might limit IDG’s future options and cost competitiveness. Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 9 9A97D010

Lo had become agitated as Lattimer described IDG’s exploration of additional manufacture sourcing arrangements, but had to agree it made sense from IDG’s perspective. Lattimer reassured Lo that IDG wanted to add capacity, not replace it. This discussion had reinforced the need to delicately handle IDG’s existing relationships. Any recommendation for locating new production capacity would have to take into account the skilled Canadian workforce, Century Glass and the Lo family, and Dan Kim’s offer for an expanded relationship in Vietnam.

The Richard Ivey School of Business gratefully acknowledges the generous support of The Richard and Jean Ivey Fund in the development of this case as part of the RICHARD AND JEAN IVEY FUND ASIAN CASEE SERIES.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 10 9A97D010

Exhibit 1

SAMPLE OF DECORATIVE GLASS PANEL APPLICATIONS

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 11 9A97D010

Exhibit 2

SUMMARY OF MAJOR CANADIAN COMPETITORS

COMPANY ACCENT JCX GLASS ROSEVIEW Target Market Small regional distributors. Anyone who calls. Small regional distributors. Supply Custom — None

Volume — Langley, B.C. and Tacoma, WA

Custom — None Volume — New Westminster, B.C.

Georgia Buy from China

Custom — None Volume — Surrey, B.C.

Positioning Good Quality. Copy designs of others. Design Leader. Lower Quality.

Cost Base - Two locations, 38,000 sq. ft. - Heavy overheads. - Non-union. - Small orders, but purchase

materials in volume. Thus very high raw material & finished goods inventory.

- Efficient production system.

- Two locations, 105,000 sq. ft. - Heavy overheads. - One year left on collective

agreement. - Volume purchase. - Finished goods inventory of

$3.2 million. - Efficient production system.

- One location, 38,000 sq. ft. - Heavy overheads. - Non-union. - High raw material costs. - Finished goods inventory of

$1.6 million.

Sales (est. 1995) $11 million Down, some of their lowest months.

$14 million Up 39 per cent.

$3 million Down.

Warranty One year. 10 years. One year. CAD Yes. No. No. MRP Some implementation. Some implementation. No. Reputation/Customer Relations

Very good in Pacific Northwest with the “old boys” network.

Generally poor, can let the customer down.

Generally poor, always lets customer down

Management Good, but have lost their spark and desire.

Aggressive, but weak in the middle management.

Generally weak.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 12 9A97D010

Exhibit 3

SUMMARY OF MAJOR U.S. COMPETITORS

COMPANY SPANNER DOOR WESTERN DESIGN BILLINGS NEW ENGLAND GLASS Target Market National (U.S.) National (U.S.) National (U.S.) Eastern (U.S.) Positioning Good quality.

Simple, high volume panels.

Broad product line. Broad product line; focus on high volume commodities, although some lower volume panels

Fast delivery, high quality.

Supply Good operations in Mexico, with long-term commitment.

Plants in Mexico and Thailand.

No offshore production. High cost producer. Focus on automation.

Est. 1995 Sales $120 million $85 million $60 million $25 million Reputation/Customer Relations

Extensive distribution system.

Product line is narrower than IDG.

Strong, dependable supplier.

Management

Three top managers have left recently.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 13 9A97D010

Exhibit 4

ORDER CYCLE TIME FOR PRODUCTION AT SHUENYI PLANT

Exhibit 5

TYPICAL PRODUCTION COSTS

Raw materials ordered and received for shipment 2-4 weeks

Components in transit to China facility 5 weeks

Raw materials conversion to WIP and semi-finished goods 4 weeks

Sub-assemblies shipped to Canada 5 weeks

Final assembly completed at Delta, B.C. facility 2 weeks

Finished goods shipped to customer 1/2 week

Total Time 18-20 weeks

Product Shuenyi Delta

#677, Oval-San Marino Materials 95.19 92.97 Labor 6.61 69.44 Freight 7.82 1.25 Total direct costs $109.62 $163.66

#936, 22" x 36" panel Materials 44.27 44.27 Labor 3.18 40.27 Freight 7.08 1.25 Total direct costs $54.53 $85.79

#445, 7-1/2" x 18-1/2" panel Materials 15.51 15.51 Labor 1.10 10.13 Freight 1.08 0.50 Total direct costs $17.69 $26.14

Production location

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 14 9A97D010

Exhibit 6

INCOME STATEMENT FOR INTERNATIONAL DECORATIVE GLASS as of September 30

(all figures reported as $000s)

1995 1994

Sales 5,404$ 3,634$

Cost of Sales 4,365 2,610

Gross profit 1,039 1,024

Expenses administration and marketing 388 413 travel and promotion 97 44 rent and assessment 120 138 amortization of debt 48 55 bank charges and interest 141 48 interest on long-term debt 18 17 other expenses 182 258 subtotal 994 973

Income (loss) from operations 45 51

Other income 28 -

Income (loss) before taxes 73 51

Income taxes current 24 - deferred (6) 11

Net income (loss) for the year 55$ 40$

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 15 9A97D010

Exhibit 7

BALANCE SHEET FOR INTERNATIONAL DECORATIVE GLASS as of September 30

(all figures reported as $000s)

1995 1994

Current cash 1 2 accounts receivable 1,513 474 income taxes recoverable 15 22 inventories 1,422 988 prepaid expenses 54 28

3,005 1,514

Capital assets 233 296

3,238 1,810

Current bank loans 1,435 593 accounts payable 886 482 income taxes payable 17 - current portion of long-term debt 32 39

2,370 1,114

Long-term debt 152 177

Deferred income taxes 13 20

Due to (from) affiliated company 522 372

3,057 1,683

Share capital 0.1 0.1 Contributed surplus 45 45 Retained earnings 136 82

3,238$ 1,810$

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

Page 16 9A97D010

Exhibit 8

PRODUCTION EQUIPMENT REQUIRED FOR START-UP IN VIETNAM (all figures reported as $000s)

Production equipment Cost

Electrical back-up generator $13

Air compressor 2

Glass equipment two-shape cutter (pneumatic, from Korea) 7 shape cutter (CNC, from Canada) 110 glass washer 60

Brass equipment roll-former 55 roll-forming dies 22 circle rollers (large and small) 5 saws (4)/blades/sharpeners 4

Bevelling equipment straight-line beveller 125 curved bevelling machines (12) 30

Miscellaneous equipment small forklift 7 pallet jack 2 computer, fax, etc. 3 hand tools, tables, etc. 5

Total capital equipment $449 F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

9-608-093 R E V : M A R C H 3 1 , 2 0 0 9

________________________________________________________________________________________________________________ Professor Zeynep Ton and Research Associate Catherine Ross, Global Research Group, prepared this case. This case was developed from published sources. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 2008, 2009 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545- 7685, write Harvard Business School Publishing, Boston, MA 02163, or go to www.hbsp.harvard.edu/educators. This publication may not be digitized, photocopied, or otherwise reproduced, posted, or transmitted, without the permission of Harvard Business School.

Z E Y N E P T O N

C A T H E R I N E R O S S

The Home Depot, Inc.

The year 2007 started with a bang at the Atlanta, Georgia-based do-it-yourself home improvement chain the Home Depot, Inc. (“Home Depot”). With 364,400 employees, Home Depot was the largest specialty retailer in the U.S., and the second largest American retailer after Wal-Mart.1 On January 3, 2007, Frank Blake took over as chairman and CEO, one day after former chairman and CEO Robert Nardelli’s surprise resignation. A January 2 Board of Directors meeting had addressed the board’s simmering discontent with the company’s depressed share price despite improved revenues and profits under Nardelli. During the company’s May 2006 annual meeting, the board had suggested that Nardelli’s generous compensation package be tied to the company’s share price, a proposal Nardelli had flatly rejected.2 On January 2, 2007, Home Depot and Nardelli announced they had “mutually agreed” on Nardelli’s immediate resignation.

Formerly a top executive of General Electric (GE) in charge of its Power Systems Division, Nardelli had arrived at Home Depot to considerable fanfare in December 2000 to take over for retiring founders Bernard Marcus and Arthur Blank. After years of rapid growth, Home Depot in 2000 was a $45.7 billion, 1,134 store home improvement retail chain.3 (See Exhibits 1 and 2 for Home Depot’s financial data.) Nardelli set out to bring greater discipline to the sprawling retailer’s operations, including merchandising and store management. He introduced cost-cutting measures at various levels, including centralizing purchasing and shifting from mostly full-time store staffing to depend on more flexible part-time contracts. By 2006 sales had shot up to $90.8 billion. Profits had more than doubled from 2000 to 2005 to $5.8 billion.4 (See Exhibit 2.) Yet Home Depot’s stagnant share price throughout Nardelli’s tenure compared poorly with close competitor Lowe’s, which saw soaring gains in its share price over the same period.5 Furthermore, Nardelli’s centralization had been criticized by employees, managers and customers as negatively affecting the quality of service and company morale and identity.6

As successor to Nardelli, Blake faced significant challenges. Share price remained low, although it rose 2.3% the day after the announced change in leadership.7 (See Exhibit 3.) In 2007, Home Depot was facing a downturn in the housing market, a principal driver of the home improvement retail industry.8 Home Depot’s chief competitor, Lowe’s, was adding stores and directly challenging Home Depot in its previously uncontested turf of urban markets.

Blake also faced skepticism on Wall Street. Like Nardelli, Blake lacked experience in retail. He was also a former GE man, having served as general counsel and head of business development in Nardelli’s GE Power Systems division.9 A lawyer by training, he had worked for the U.S. government after leaving GE and before following Nardelli to Home Depot in 2002 to become his deputy.10 However, in contrast to the abrasive Nardelli, Blake had a reputation for being a listener and for

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

seeking consensus.11 Some analysts expressed surprise that he had not been named merely an interim leader while Home Depot looked for a top retail executive for the post.12 An analyst at Goldman Sachs commented to Business Week: “This is the second-largest retailer in the nation, and [Blake] has never run a company, or one of Home Depot’s major operating businesses.”13 To reassure Wall Street and all of the firm’s stakeholders, Blake would have to decide rapidly which of Nardelli’s strategies to maintain and strengthen, which ones to modify, and which ones to dismantle.

The Home Depot Story

Home Depot was founded in 1978 in Atlanta, Georgia by Bernie Marcus and Arthur Blank. The founders had both been fired from their executive posts at the Handy Dan Home Improvement Centers in California. Rather than seeking employment elsewhere, they sought start-up capital to fund their vision of “one-stop shopping for the do-it-yourselfer”14—vast home improvement stores akin to warehouses that would offer a broad selection of tools and products and would be staffed by knowledgeable experts in home improvement and customer service.15 The new company’s first two 60,000 square-foot stores opened in 1979 in Atlanta with 25,000 stock keeping units (SKUs), far more than offered at that time under any one roof.16

The company popularized the concept of “do-it-yourself” (DIY), in which homeowners and other individuals purchased products and tools and then built, repaired and improved homes on their own. Home Depot facilitated the DIY concept not only by prioritizing customer service, but also by providing customers with training workshops and clinics to teach them how to go about “laying tile, changing a fill valve, or handling a power tool.”17 Sales associates underwent rigorous training in product use before attending to customers. From the outset, Home Depot was characterized by a close focus on the customer: Bernie and Arthur, as Marcus and Blank were known to employees, championed the customer service philosophy of “whatever it takes,” encouraging sales associates to develop relationships with customers rather than seeing sales as a transaction.18

The company grew quickly, achieving $22 million in sales in four stores by 1980, and $3.8 billion with 145 stores by 1990, as it became the top U.S. retailer in the home improvement industry.19 It was the youngest retailer ever to reach $30 billion in revenues, a feat that would be repeated as it attained $40 billion, $50 billion, $60 billion and $70 billion in sales faster than any other retail company worldwide.20 By January 2007, Home Depot had 1,800 stores in the U.S.,21 had expanded into Canada, Mexico, Puerto Rico and Chile, and had just entered China.22 (See Exhibit 1 for Home Depot performance data.) An average Home Depot store was 105,000 square feet in size and carried 40,000 SKUs23.

The Home Improvement Industry and Competitive Landscape

Throughout the 1980s, the home improvement industry was fragmented, with sales divided among niche players such as hardware, lawn and garden, and paint and wallpaper stores as well as big box home center stores such as Home Depot and Lowe’s. Beginning in the early 1990s, independent stores began losing market share to the growing big box giants, a trend that continued throughout the 1990s.

When the ABC television program Home Improvement debuted in 1991, Home Depot’s annual sales were about $3.8 billion and the company employed 21,500 people. Throughout the 1990s and into the new millennium, weekend “DIY-ing” became so popular that two cable networks—Home & Garden Television and the Do It Yourself Network—were devoted to the pastime. Fueled by a housing

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

boom—a bubble that since burst—home improvement outperformed most U.S. retail sectors, reaching an estimated $291.3 billion in 2005.24

In 2006, the deteriorating housing market negatively impacted the home improvement retail market; the industry reported $200 billion in sales in the U.S. that year.25 Analysts reported indications of market saturation as well. However, analyst reports were not entirely pessimistic about the strength of the home improvement industry, citing continued population growth and the need for new homes as well as for repair and renovation of existing ones.26 Providing services such as tool rental, and increasingly, installation was a new trend in the home improvement retail industry—one that became known as “do-it-for-me” or DIFM.27

Retailers in the home improvement business carried products such as building supplies and lumber, plumbing and electrical items, tools, hardware, paint and wallpaper, floor tile, upholstery, glass and window fixtures, blinds, lawn and garden supplies and even home appliances.28 Home Depot was the biggest player in the home improvement market, followed by Lowe’s. These two companies were each other’s most significant competitors, and held approximately 60% of the U.S. market share between them in 2006.29 Wal-Mart, America’s largest retailer also competed with Lowe’s and Home Depot. Hardware stores such as True Value Hardware and Ace Hardware, which maintained cooperatives among independent retailers, followed at some distance, along with hardware store Menard’s.30 (See Exhibits 4 and 5.) In 2006, the average size of a transaction for hardware stores was reported to be $15, compared to an average transaction of $41 for home centers.31

Big box stores purchased from a wide array of vendors. In 2006, Lowe’s received supplies from approximately 7,000 distinct vendors. Efficient management of information and of the supply chain was crucial. The home improvement industry was cyclical. The period from 2001-2006 demonstrated variable growth. See Exhibits 6, 7, and 8 for trends in expenditures for residential improvements and repairs, housing built, and housing sales.

Lowe’s

Founded in 1921, Lowe’s business had focused primarily on selling materials to contractors, but the housing decline in 1980 resulted in a decrease in profits, prompting the company to rethink its target market. Over a two-year period Lowe’s redesigned half of its 229 stores to serve the needs of DIY consumers, creating a friendlier and more accessible setting with softer lighting and full room displays. The design overhaul was a success, and in 1982 over half of Lowe’s customers were non- homebuilding professionals.32

Lowe’s maintained its small store format through the 1980s, a move that potentially cost it the #1 position among U.S. home retail chains, which Home Depot took over in 1989.33 It was not until the 1990s that Lowe’s pursued a warehouse-style layout for its stores. In 1991 Lowe’s began to phase out smaller stores and by 1993 it opened 57 warehouse stores, nearly doubling its overall floor space.34 Lowe’s opened more than 140 stores during the 1990s, and in 1998 announced a $1.5 billion expansion plan that would grow the company by 100 stores across the U.S. over several years.35 Three years later in 2001, Lowe’s allocated $2.4 billion for further store expansion and the building of new distribution centers.36 Between 2001-2002 Lowe’s opened over 200 new stores, a trend that continued through 2007, bringing Lowe’s total stores close to 1,500.37 (See Exhibit 9 and 10 for Lowe’s performance data.)

Lowe’s rapid growth between 2000 and 2007 reflected the company’s growing customer base. The introduction of upscale product lines, home appliances, and installation services, all offered against a welcoming backdrop, helped Lowe’s attract new consumers, specifically wealthy baby boomers and F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

women shoppers.38 The decision to sell home appliances in Lowe’s stores resulted in Lowe’s ranking second to Sears in U.S. home appliance sales.39 These successes translated into monetary gain: Lowe’s shoppers spent an average of $10 more per trip than Home Depot customers.40

Home Depot’s First Decades

For its first 20 years, Home Depot was widely known for its entrepreneurial spirit and focus on customer service and sales growth. Stores were run largely informally. Store managers had almost complete autonomy in their own store operations.41 They were not only allowed but even encouraged by senior management to ignore or send back directives from headquarters that they felt to be intrusive to their own stores.42 Because it was commonly believed that managers should spend their time on the sales floor with customers, company paperwork often ended up buried under piles on someone’s desk, tossed in a wastebasket, or even marked with a company-supplied “B.S.” stamp and sent back to the head office. “The idea was to challenge senior managers to think about whether what they were sending out to the stores was worth store managers’ time,” said Tom Taylor, who started at Home Depot in 1983 as a parking lot attendant and in 2006 was executive vice president for merchandising and marketing.43

Merchandising, Store Operations, and Vendor Management

Local autonomy was intended to encourage innovation and responsiveness to the needs and preferences of the local market. “Whether it was an aisle, department or store, you were truly in charge of it,” explained a former Home Depot store manager.44 Store managers decided on merchandizing, displays and promotions and set employee wages. Merchandising was based on each store manager’s local market intuition and not on particular metrics, tools or data for quantitative analysis. Nine regional purchasing offices negotiated separately with suppliers. Agreements would vary from region to region. Blank believed that decentralization helped boost sales 15% to 20% because Home Depot buyers understood local needs.45

Decentralized purchasing meant that “the retailer was acting as if it were nine $5 billion companies rather than a single $45 billion company, thus squandering the chance to drive down costs and boost gross margins,” wrote one industry observer.46 Promoting autonomy and responsiveness also implied difficulty in delivering on agreements with vendors. In one case, a purchasing center made a deal for garden furniture that included a 10% discount for Home Depot in return for prominent displays in the stores in that region. But many store managers simply ignored the deal, deciding the display was not what their store needed.47 Home Depot’s decentralization also meant there was little communication among managers, and limited ability to negotiate national deals.48 While the company could generate nationwide store displays from its headquarters, it could not easily coordinate them with nationwide purchasing.49

Most Home Depot stores depended principally on full-time personnel who were knowledgeable about home improvement. One analyst in early 2000 praised this strategy as one that kept productivity high and turnover low. “Home Depot traditionally shies away from minimum wage labor and believes that paying the best people will drive higher store sales and productivity.”50

In 2000, a company executive cited the high level of customer service—a central tenet of Home Depot’s management philosophy under Marcus and Blank51—as one of two major reasons for Home Depot’s success.52 The company’s 2000 annual report referenced a “recent study” in which “Home Depot associates ranked 40% higher than the competition in customer service and product knowledge.”53

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

Need for Change

Despite Home Depot’s astounding success, Marcus and Blank stated that further growth would require new ways of thinking and doing. The founders pointed out the need to apply new technologies, build efficiencies and reallocate resources, while still listening and responding to customers.54 To bring more operational discipline to the company, the founders recruited Robert Nardelli from GE.

The Nardelli Era

In his first week as CEO Nardelli asked how to email all 1,134 stores and was told it was not possible.55 The Atlanta headquarters generally communicated with store managers by fax.56 In 2000, the company employed 227,000 associates, yet had no head Human Resources officer.57 Despite its rapid growth and innovative store format, Nardelli said the company had been “in start-up mode for 20 years,”58 and that the company’s decisions were generally “based on emotion rather than data.”59 The retailer was years behind other retailers like Wal-Mart in technology. For example, bills and invoices were largely processed by hand60 and for those items that did not have barcodes, employees had to manually go to a book to look up the code. There was no general counsel, no chief marketing function, and no CFO. Nardelli quipped, “Call me old-fashioned, but I kind of like the CFO reporting to the chairman.”61

To respond to falling sales, rising costs and stiffening competition, Nardelli devised a three-part strategy: extending the business into new lines; expanding the market both geographically and with new types of customers; and making existing operations more sound and profitable.62

In 2001, Home Depot began providing “specialized products and services to smaller professional customers” and announced the intention of growing the professional business.63 By 2003, Nardelli had expanded its wholesale business, Home Depot Supply, which provided products and services to a wide range of professionals including real estate developers, plumbers, electricians, construction crews and industrial contractors rather than to the ordinary do-it-yourselfer.64 HD Supply grew largely through acquisitions,65 and was a $12 billion business by 2006.66 Nardelli also expanded Home Depot’s market by announcing steps to move into other countries. Previous leadership had already taken the company into Canada, and Nardelli brought it to Mexico. In 2001, 2002, and 2004, Home Depot bought three separate home improvement chains based there. “We went from zero to number one in Mexico,” said Nardelli.67 To him, international expansion and diversification into the wholesale business were strategic responses to the declining U.S. housing market.68

Nardelli referred to the central strategy of improving existing operations as “enhancing the core.”69 Several initiatives fit under this banner, including major changes to fundamental retail functions such as merchandising, vendor management, and store operations. In keeping with this focus, Home Depot prominently displayed the new slogan “Improve Everything We Touch.”70

Changes Introduced in Merchandising

Nardelli centralized Home Depot’s merchandising and purchasing. By July 2001, he had reassigned these functions to 12 merchandising vice presidents who would work out of the Atlanta headquarters. Stripped of purchasing power, the former regional purchasing offices’ efforts were redirected toward sales, service and supporting stores for the presentation of merchandise procured through the new centralized system.71 The centralization was aimed at realizing gains in purchasing power and addressing inefficiencies in operations.72

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

In September 2001, Nardelli used his keynote address at the National Hardware Show in Chicago to promote Home Depot’s new policy of consolidated purchasing, saying the reorganization would give the regional managers of the chain “more time to convert merchandise into sales.”73 An executive for merchandising and marketing in a Canadian home improvement store commented, “I think what they are doing is absolutely right. When you buy regionally, you find you’re losing too much control of your inventory. And when stores don’t pay attention to inventory management, you find your customer service goes down because there’s too much product on the [sales] floor.”74

Detractors feared the centralization would lead to stores offering exactly the same product mix, disregarding local variation.75 But Nardelli maintained that centralized purchasing was necessary for Home Depot to have greater control of over the mix of products available in stores, to ensure greater consistency in merchandising, and to better manage vendors and inventory.76 By 2003, the company had eliminated almost 20,000 SKUs from its overall inventory and added others—mostly higher- priced, higher-end items aimed at increasing revenue—as part of centralization of merchandising.77

Greater purchasing power allowed Nardelli to broker exclusive deals with certain vendors including John Deere, Mill’s Pride Cabinets and furniture-maker Thomasville. He placed emphasis on driving relationships with loyal vendors and developing more business with fewer entities.78 “Two years ago we never could have done [a deal with] John Deere,” Nardelli said in 2003. “Every business, every store was their own sourcing center. Now we can negotiate an exclusive arrangement with the strongest national brand in the country.”79 Home Depot used its beefed-up purchasing power to negotiate better deals, extending payment terms to 45 or 50 days up from the typical 30 days.80

To support centralized merchandising Home Depot spent over $1 billion to modernize the retailer’s technological infrastructure and IT systems, including a new system for inventory management.81 By 2004, Home Depot implemented assortment planning, markdown management, and store space planning software.82 By 2005, 11% of sales were replenished automatically.83

Six Sigma Approach to Store Operations

Nardelli was a firm believer in the power of process discipline and operational execution. Drawing on his GE training, he was steeped in the culture of Six Sigma, a management methodology that used rigorous analysis to improve quality, and was eager to implement this methodology at Home Depot. With approximately 1.3 billion customer transactions a year, a slight improvement in these key metrics could boost financial performance, he wrote to his shareholders in the 2004 Annual Report.84

The Six Sigma method had been pioneered (and trademarked) by Motorola in the 1980s and was later adopted by GE and others.85 Motorola estimated that it saved at least $15 billion by using Six Sigma from the time of its development through 1999.86 Former GE CEO Jack Welch, an enthusiastic proponent of the methodology, believed that Six Sigma techniques led to $600 million in savings in 1998 alone.87

In its initial use at Motorola and elsewhere, Six Sigma aimed at eliminating defects in manufacturing processes through reducing variation. A defect was defined as nonconformance of a product or service to its specifications and sigma referred to a measure of variation. A process with six sigma quality produced less than 3.4 defects per (one) million opportunities whereas a process with three sigma quality produced less than 2.7 defects per thousand opportunities.

The Six Sigma methodology filtered out of the manufacturing world into other business segments, including the management of sales and human resources. A former statistical methods department F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

The Home Depot, Inc. 608-093

manager at Motorola said projects that can be “broken down into discrete, manageable activities”88 are “the right projects for Six Sigma.”89 In the context of retail store operations, processes associated with the flow and storage of products (i.e. in-store logistics activities) were good candidates for Six Sigma process improvement.

The Six Sigma approach was embodied in a four-step process to achieve improved quality: measure, analyze, improve, and control. Measurement was the crucial first step which allowed a company to quantify its quality and accuracy in business processes. Anything and everything could be measured, from the number of high quality parts produced to the time it took to process orders from customers or the number of daily sales made per sales representative. Measurements and variables were then subjected to a process of analysis to determine optimal outcomes and objectives for the performance of certain business processes. Improvement came from changing processes and methods to better achieve the performance goals that had been set. Finally, once new processes were implemented, they were monitored with an eye to controlling and maintaining quality levels.90

Changes Made in Store Operations

Consistent with Six Sigma methodology, Nardelli focused on precision, measurement and quantitative analysis to bring to the fore a focus on efficiencies in operational processes. “When I came here, I was told, ‘If you get 50 percent compliance, that's a good day,’” Nardelli recalled. “And I said, ‘Well, that's not a good day because customers’ expectation is to have consistency of delivery not only in the store they shop, but as they go from Home Depot to Home Depot. “We can't live with 50 percent compliance. You have to have 100 percent compliance.”91

Changes to improve productivity Nardelli’s method for improving store productivity included simplification of each store associate’s job function with an eye toward more specialization.92 For example, with the Service Performance Improvement initiative, Home Depot began to stock new inventory at night, so that sales associates could spend more time during the day with customers.93 Sales associates were still responsible for replenishing products from storage areas during the day. Nardelli also worked on standardizing in-store logistics activities, from the moment freight arrived at the backdoor of the store to its eventual sale or return to the vendor.Productivity metrics were introduced around each step of the freight flow process (e.g. pallets per hour).94 Stores also had to track “inventory velocity,” the length of time it took for products to flow through stores.95

To improve labor productivity, Nardelli also invested heavily in technology. Through the Front End Accuracy and Service Transformation (FAST) initiative,96 all stores were equipped with new point-of-sales terminals with touch screens. These terminals made looking up unbarcoded items simpler and faster for cashiers. Self checkout registers were installed in 800 of the company’s highest volume stores.97 Cordless scan guns were introduced at a number of cash registers after a Six Sigma study found that cashiers were faster and more accurate when they used them.98 By 2004, cordless scan guns had been installed in all stores, and over 1,000 self-checkout aisles were in operation.99 By mid-2005, Home Depot had also installed a computerized process to scan incoming inventory.

Changes to increase data accuracy The FAST and back-end scanning and receiving process initiatives were intended to improve accuracy of point-of-sales data and inventory data, respectively. In addition, ordering carts were introduced to facilitate restocking of shelves from storage locations. Home Depot stores received more stock of products to the store than they could present on the display shelves. The extra stock was typically stored on the selling floor in the overheads—the upper shelves that were 8, 12, or 16 feet above. Large products were sometimes stored in the backrooms.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

Many products were placed in overheads in areas away from their display locationsi. Previously, an associate would walk down the aisle and manually write down out-of-stocks from the shelves, take that list to the computer room, generate a report with a “pick list,” and then pick those products from the storage locations. With the new ordering cart, associates could walk down the aisles and scan any shelf with an empty space in order to electronically create a pick list. If additional stock of the missing product was present in the store, it was included in the pick list so that associates would know the product was in a storage location and would look for it. In the case of a discrepancy between what the system said the store had in stock and what actually was in stock, associates could make inventory adjustments.100 That is, associates could change the inventory levels in the computer system to reflect the physical quantities at the stores.

Changes to improve labor scheduling Nardelli invested in tools for Home Depot’s store managers to better forecast staffing needs and to schedule labor. The new computerized workload management system allowed store managers to see what activities were scheduled to take place at the store and historical and future trends to inform decision-making about hiring, staffing and training.101

Seeking flexibility for scheduling labor at stores, Home Depot announced it would extend benefits to part-time staff. The company tried to increase the number of part-time employees, from 15% of personnel to 50%.103 Although a negative reaction to this caused top management to backpedal on the number of part-timers hired, the net increase of part-timers was still significant.104

Personnel Changes

Nardelli’s arrival resulted in significant changes for employees at the corporate and store level. Within two years of Nardelli’s arrival, 22 of the company’s top 29 managers left.106 Between 2001 and 2006, 98% of Home Depot’s 170 top executive positions changed hands, and over half of the new hires were from outside the company.107

In an effort to recruit new managers and train them in the new management approach, Nardelli and Dennis Donovan—a former GE executive whom Nardelli recruited to join Home Depot as his chief human resources officer—developed a two-year intensive program to fast-track new store managers.108 The program carefully selected participants and aimed to train newcomers in company practices, analytics methodologies and culture.

In 2002, 528 of 1,142 future managers hired from outside the company had military experience.109 Donovan explained that the company was seeking “people who deliver results, act strategically, and drive excellence," in contrast to traditional store managers who were frequently experts in hardware. “Leaders excel in customer service,” he said. “They inspire achievement, they live [with] integrity, they build strong relationships, and they create an environment of inclusion. Junior military officers have these essentials.”110 A former Navy Lieutenant who in 2005 was Home Depot’s director of implementation--a position created to standardize processes at Home Depot—agreed. “Military structure is very similar to how our stores are structured,” he said. “A store manager is basically the equivalent of a ship captain. You’ve got a leadership team under you, and I think the relationships you need to build between your hourly and management associates are a natural fit for military leaders.”111 “Bob Nardelli and I have hired military through a big part of our careers, so we know this model works.”112

i Information on storage locations is based on case writer’s store observations.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

Analyzing Metrics and Making Plan

Nardelli also introduced discipline through data analysis and performance evaluation. Store managers were required to “make plan” by achieving weekly and monthly sales or other performance targets. All personnel were ranked according to four “performance metrics:” financial, operational, customer, and people skills.113 Nardelli and Donovan placed Human Resource managers in every store to track these metrics.114 One market analyst gave the new discipline a positive review, saying Home Depot was becoming “much more disciplined, more numbers-oriented and more accountable for everything.”115 To reinforce the new discipline and to better communicate with employees and store managers, Nardelli established an internal TV Channel (Home Depot TV, or “HDTV”) that piped in messages to stores.116 Every Monday night, EVPs for marketing and merchandising delivered the store staff’s “marching orders for the week” through a 25-minute live program named “The Same Page.”117 At other times, HDTV ran segments reminding staff of key messages and policies.

Under Nardelli, Home Depot developed personnel incentive plans with the aim of better tying incentives to employee performance. In 2002, the “Success Sharing” program debuted, specifically targeting hourly workers.118 The program linked employee performance on the new metrics used for evaluation with bonus payments. In 2004, payouts from the program totaled nearly $90 million.119

The whole cloth culture change that Nardelli introduced, symbolized by the focus on analytics, did not always lead to the desired results. Managers had to monitor their store numbers, reducing the time they spent on the floor. Former Chief Information Officer Ron Griffin expressed his frustration at having to demonstrate a “return on every nickel.”120 Several stores responded to the “inventory velocity” metric requirement by reducing the amount of inventory coming in, leading to stockouts.121 A former Home Depot merchandising executive explained, “On paper, all these changes make sense. Unfortunately, they don’t work on the floor of the stores.”122 And employees resented the sensation that Big Brother was watching, generated by Home Depot TV. They surreptitiously dubbed it “Bobaganda.”123

Reactions from Wall Street

Wall Street’s initial reaction to the changes introduced by Nardelli was mostly positive. There was widespread agreement with Nardelli’s assessment that what had made Home Depot successful in its first 20 years was unlikely to continue to work as the company aimed for $100 billion in revenues.124 Although some analysts noted that Home Depot under Nardelli opted for a more centralized structure at a time when retailers were devolving more autonomy to stores, many agreed that in Home Depot’s case, centralization was a necessary move.125 Analysts approved of Nardelli’s investment in information systems infrastructure, improved supply chain logistics, the strengthened opportunities for economies of scale through the centralization of merchandising, and Nardelli’s introduction of greater analytics as a basis for decision-making.126 Home Depot’s strategic alliances with John Deere and others were also cited as strengths.127 While analysts noted the disgruntlement of present and former employees with the changing culture at Home Depot, it seemed par for the course. Many felt there was bound to be a reaction against cultural change in a company that had previously been famous for and proud of having little corporate discipline.128

In 2005, some analysts reflected some reservations. Many doubted Nardelli’s strategy to move away from consumer retail and place emphasis on selling to professional contractors through HD Supply. Some analysts asked whether Home Depot was pursuing too many strategic initiatives at once and gave the company a less enthusiastic rating.129 Other reports fretted about the languishing stock, the housing market and the “disconnect between Nardelli’s pay and Home Depot’s Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

performance.”130 Nardelli had received $200 million over the previous five years while shareholders saw “cumulative total shareholder returns of approximately -13%.” Still, other analysts were bullish enough about Home Depot in late 2005 to place it above Lowe’s as their top home center equity because of its positioning for future growth and its relatively low valuation.131

Performance Under Nardelli

In addition to the doubling of revenue under Nardelli’s leadership, his relentless focus on cost- cutting had been successful in pushing gross margins from 30% in 2000 to 33.8% in 2005. Between 2000 and 2005 Home Depot had a 12% average annual growth rate in sales. Riding the same housing boom, Lowe’s had managed to increase its sales by an annual average of 19% over the same time period.132

But while Nardelli successfully slashed costs and improved Home Depot’s financial management, he did little for the share price. In December 2000, when Nardelli took the reins as CEO, Home Depot’s stock price was trading slightly under $37 per share. Six years later when Nardelli left Home Depot the stock was only slightly better off (refer to Exhibit 3). By mid-2006 total return to shareholders was down 6%. During the same six-year period, Home Depot’s rival Lowe’s saw its split-adjusted stock price rise more than 200%.133 (See Exhibit 10 for a comparison of share price performance between Home Depot and Lowe’s.) Analysts explained the poor share performance as a result of Nardelli’s approach to management; one analyst noted that the “numbers were quite good [but] the fact is that this retail organization never really embraced his leadership style.”134

Customer Service

Research suggested that the slide in customer service was likely related to Home Depot’s declining stock price.135 “He's made Home Depot much more profitable and more streamlined, but messed up everything that has to do with serving the customer,” said one analyst. “They've got people in there working for less money and are less knowledgeable and less experienced. It's all about profitability, at the cost of serving the customer.”136 Increased use of part-timers was a source of concern for many. While part-time employees were generally cast in the role of cashiers or shelf- stockers,137 part-timers undercut a fundamental customer service premise—that of giving the do-it- yourselfer the confidence that prepared staff in orange aprons were available to help. “We built the entire company around the idea that a customer could come in and ask us how to do anything—fix a toilet, build a deck, whatever—and we’d tell them how to do it,” said a supervisor at a Florida store. “Now we’ve got kids who don’t even know what the products look like.”138

Nardelli rejected the notion that an emphasis on part-time staff was inconsistent with excellent customer service. He defended the shift as a positive move for customers who would have more salespeople to attend to them at peak hours. Like the self check-out machines, Nardelli saw increasing the number of part-time employees as a way to ensure better coverage on the floor, and more attention to customers.139 Investments in store processes and technologies meant that two to three additional associates could be redeployed to the selling floor per store per week to assist customers.140 Investments in store modernization made stores more attractive.ii The centralized merchandising system resulted in higher in-stock levels.141 One industry expert noted “Nardelli is very acutely aware of the personnel cost of providing personalized service and I think they've done a

ii In 2004, the company reported having spent $1 billion on store remodeling and refurbishing.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

fairly good job at making the stores as self-serviceable as possible. It's easier to find product than it was before and stores are cleaner and easier to shop.”142

Yet, many employees who had grown up in the Marcus and Blank era felt the new focus and strategies prevented them from delivering excellent service. Some insiders began referring to the chain as “Home GEpot” or “Home Despot,” monikers that leaked out to the business press.

Business Week reported that Nardelli had “alienated customers just as thoroughly as he did employees. Staffing cuts led to persistent complaints that there weren’t enough workers in Home Depot’s cavernous stores to help do-it-yourself customers.”143 In the summer of 2002, the Better Business Bureau’s Atlanta chapter—in Home Depot’s hometown—suspended the retailer’s membership in response to the skyrocketing number of customer complaints that had remained unresolved.144 The University of Michigan’s annual American Customer Satisfaction Index in 2005 ranked Home Depot at the bottom of the heap of major U.S. retailers in customer service (Exhibit 11).145 The retailer received a score of 67 that year, down from 73 in 2004. The ranking put Home Depot 11 points behind Lowe’s and 3 points behind Kmart. (See Exhibit 12 for readers’ comments to BusinessWeek’s story.)

By 2006, the CEO and his top executives had realized they had a customer service problem, and elevated the issue to one of their top priorities. Various initiatives were announced to support better merchandising and customer service.146 Nardelli announced an employee bonus program, called “orange juiced”147 that would award up to $10,000 bonuses to sales associates for outstanding customer service;148 Nardelli said, “We open every board meeting with customer service data.”149

The Road Ahead

Nardelli’s departure had been prompted by the quarrel over executive compensation and share price, and not by any rejection of the changes he had introduced to Home Depot. There was widespread agreement that Nardelli had ushered in a technological revolution that had brought the company out of the IT dark ages, and established a series of useful business functions and processes that had not existed under the prior leadership. Yet questions lingered regarding the impact the command and control approach might have had on what had previously been the prized centerpiece of Home Depot’s brand: its customer service.

Blake had to decide which of his predecessor’s strategies to consolidate, which to alter, and how to go about strengthening a culture of customer service at Home Depot. These decisions would inevitably take place within the pressure of a declining housing market and growing competition in Home Depot’s urban markets from the ever more confident Lowe’s.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

60 8-

09 3

-1

E xh

ib it

1 H

om e

D ep

ot P

er fo

rm an

ce D

at a,

1 99

7– 20

06 ($

a m

ou nt

s in

m ill

io ns

, e xc

ep t a

ve ra

ge c

us to

m er

ti ck

et , n

um be

r of

c us

to m

er tr

an sa

ct io

ns a

ls o

in m

ill io

ns )

H om

e D

ep ot

19 98

20 00

20 02

20 04

20 06

G ro

ss M

ar gi

n as

a %

o f s

al es

.1 %

.5 %

.7 %

.9 %

.2 %

.1 %

.8 %

.4 %

.5 %

.8 %

m ou

nt o

f i nv

en to

$3 ,6

$4 ,2

$5 ,4

$6 ,5

$6 ,7

$8 ,3

$9 ,0

$1 0,

07 6

$1 1,

40 1

$1 2,

82 2

In ve

nt or

y pe

r st

or e

$5 .8

$5 .9

$5 .0

$5 .3

$5 .6

N um

be r

of s

to re

s 62

4 76

1 93

0 1,

13 4

1, 33

3 1,

53 2

1, 70

7 1,

89 0

2, 04

2 2,

14 7

N um

be r

of e

m pl

oy ee

s 12

4, 40

0 15

6, 70

0 20

1, 40

0 22

7, 30

0 25

6, 30

0 28

0, 90

0 29

8, 80

0 32

3, 14

9 34

4, 80

0 36

4, 40

0 N

um be

r of

c us

to m

er tr

an sa

ct io

55 0

66 5

79 7

93 6

1, 09

1 1,

16 1

1, 24

6 1,

29 5

1, 33

0 1,

33 0

A ve

ra ge

c us

to m

er ti

ck et

3. 63

5. 05

7. 87

8. 65

8. 64

9. 43

1. 15

4. 89

7. 98

8. 90

am e-

st or

e (c

om p

st or

e) s

al es

g ro

w th

4. 0%

.5 )%

5. 1%

(2 .8

So ur

ce :

A d

ap te

d fr

om H

om e

D ep

ot A

nn ua

l R ep

or t d

at a

fo r

20 00

–2 00

6, a

cc es

se d

F eb

ru ar

y 21

, 2 00

E xh

ib it

2 H

om e

D ep

ot In

co m

e St

at em

en t D

at a,

1 99

7– 20

06 (i

n $

m ill

io ns

)

H om

e D

ep ot

19 98

20 00

20 02

20 04

20 06

S al

24 ,1

30 ,2

38 ,4

45 ,7

53 ,5

58 ,2

64 ,8

73 ,0

81 ,5

90 ,8

C os

t o f G

oo ds

S ol

d 17

,3 75

,6 14

,0 23

,0 57

,4 06

,1 39

,2 36

,6 64

,1 91

,0 54

ro ss

M ar

gi n

6, 78

1 8,

60 5

11 ,4

13 ,6

16 ,1

18 ,1

20 ,5

24 ,4

27 ,3

29 ,7

O pe

ra tin

g ex

pe ns

S el

lin g

an d

S to

re O

pe ra

tin g

4, 30

3 5,

34 1

6, 83

2 8,

51 3

10 ,1

11 ,1

12 ,5

15 ,1

16 ,4

18 ,3

P re

-o pe

ni ng

3 14

2 11

7 96

en er

al a

nd A

dm in

is tr

at iv

e 41

3 51

5 67

1 83

5 93

5 1,

00 2

1, 14

6 1,

39 9

1, 47

2 1,

76 2

N on

-r ec

ur rin

g C

ha rg

e 10

4 --

ot al

88 5

5, 94

4 7,

61 6

9, 49

0 11

,2 15

,2 78

,7 34

,5 04

,9 57

,1 10

O pe

ra tin

g In

co m

e 1,

89 6

2, 66

1 3,

79 5

4, 19

1 4,

93 2

5, 83

0 6,

84 6

7, 92

6 9,

36 3

9, 67

In te

re st

in co

m e

(e xp

en se

)

te re

st a

nd in

ve st

m en

t i nc

om e

In te

re st

e xp

en se

-4

2 -3

7 -2

8 -2

1 -2

8 -3

7 -6

2 -7

0 -1

-3 92

et in

te re

2 -7

-3

-1 4

-8 1

-3 65

E B

1, 89

8 2,

65 4

3, 80

4 4,

21 7

4, 95

7 5,

87 2

6, 84

3 7,

91 2

9, 28

2 9,

30 8

In co

m e

ta xe

s 73

8 1,

04 0

1, 48

4 1,

63 6

1, 91

3 2,

20 8

2, 53

9 2,

91 1

3, 44

4 3,

54 7

N et

E ar

ni ng

s 1,

16 0

1, 61

4 2,

32 0

2, 58

1 3,

04 4

3, 66

4 4,

30 4

5, 00

1 5,

83 8

5, 76

So ur

ce :

H om

e D

ep ot

a nn

ua l r

ep or

ts .

N ot

es :

Fi sc

al y

ea rs

e nd

in g:

F eb

ru ar

y 1,

1 99

8; J

an u

ar y

31 ,

19 99

a nd

J an

ua ry

3 0,

2 00

0; J

an u

ar y

28 ,

20 01

; Fe

br ua

ry 3

, 20

02 ;

Fe br

ua ry

2 ,

20 03

; Fe

br ua

ry 1

, 20

04 ,

Ja nu

ar y

30 ,

20 05

; Ja

nu ar

y 29

, 20

06 ;

an d

nu ar

y 28

, 2 00

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

Exhibit 3 Home Depot Share Price, 1984–2006 in $USD

Source: Adapted from Google Finance data for August 20, 1984 through August 20, 2006, accessed February 21, 2008.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

Exhibit 4 Home Channel Sales, by Retailer, 2004 and 2006

2004 2006 2004–06 2004–06

Store Name Sales

($ million) Share

(%) Sales

($ million) Share

(%) % Sales Change

% Point Share Change

Home Depot 73,094 19.9 90,837 21.5 24.3 1.6

Lowe’s 36,460 9.9 46,900 11.1 28.6 1.2

Wal-Mart 19,200 5.2 23,300 5.5 21.4 0.3

Sears 11,400 3.1 11,500 2.7 0.9 -0.4

CCA Global Partners 8,700 2.4 9,600 2.3 10.3 -0.1

Menards 7,000 1.9 8,000 1.9 14.3 0.0

Pro-Build 4,050 1.1 5,800 1.4 43.2 0.3

Stock Bldg. Supply 3,580 1.0 5,305 1.3 48.2 0.3

Sherwin-Williams 3,977 1.1 5,260 1.3 32.3 0.2

84 Lumber 3,490 1.0 3,920 1.0 12.3 0.0 Total of Top 10 170,951 46.6 210,422 50.0 23.1 3.4 All other 195,548 53.4 211,578 50.0 8.2 -3.4 Total 366,499 100.0 422,000 100.0 15.1 N/A

Source: Home Channel News/Mintel.

Exhibit 5 Top Home Improvement Centers’ Operating Statistics, Latest Fiscal Year-end, 2006

Company (end fiscal year 2006)

Annual Sales ($ billion)

Outlets (#)

Total Sales Area (‘000 sq. ft.)

Employees (#)

Home Depot (1/2006) 90.8 2,147 224,000 364,400

Lowe’s (1/2006) 46.9 1,385 157,000 210,142

Menards (1/2006) 8 210 27,000 38,000

Source: Home Channel News/Mintel.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

Exhibit 6 New Privately Owned Housing Units Started in the United States (Thousands of Units)

Source: U.S. Census.

Exhibit 7 New Houses Sold in the United States

Source: U.S. Census.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

Exhibit 8 Expenditures for Residential Improvements and Repairs

Seasonally Adjusted Annual Rate in Millions of Dollars

Improvements

Year Total

Expenditures Maintenance and Repairs Improvements

Additions and Alterations

Major Replacements

2001 627,300 191,100 436,200 308,300 127,900

2002 689,700 188,300 501,200 353,600 147,500

2003 707,000 179,600 527,400 372,300 155,100

2004 794,100 202,200 591,900 (X) (X)

2005 860,800 212,900 647,900 (X) (X)

2006 914,400 214,400 699,900 (X) (X)

2007 908,400 219,800 688,600 (X) (X)

Source: U.S. Census.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

60 8-

09 3

-1

E xh

ib it

9 L

ow e’

s Pe

rf or

m an

ce D

at a,

2 00

0– 20

06 ($

a m

ou nt

s in

m ill

io ns

, e xc

ep t a

ve ra

ge c

us to

m er

ti ck

et , n

um be

r of

c u

st om

er tr

an sa

ct io

ns a

ls o

in m

ill io

ns )

L ow

e’ s

19 97

19 99

20 01

20 03

20 05

G ro

ss M

ar gi

n as

a %

o f s

al es

27 %

28 %

30 %

34 %

35 %

A m

ou nt

o f i

nv en

to ry

,9 85

,3 85

,8 12

,2 85

,6 11

,9 68

,5 84

,9 82

,7 06

,1 44

In ve

nt or

y pe

r st

or e

$ 4

.1 6

$ 4

.5 9

$ 4

.8 8

$ 5

.0 5

$ 4

.8 5

$ 4

.6 5

$ 4

.8 2

$ 5

.5 0

$ 5

.4 3

$ 5

.1 6

N um

be r

of S

to re

s 47

7 52

0 57

6 65

0 74

4 85

4 95

2 1,

08 7

1, 23

4 1,

38 5

N um

be r

of E

m pl

oy ee

s 64

,0 70

,7 15

,1 60

,6 01

8, 31

7 12

1, 60

0 14

7, 05

2 16

1, 96

4 18

5, 31

4 21

0, 14

N um

be r

of c

us to

m er

tr an

sa ct

io ns

1 26

8 29

8 34

2 39

4 46

0 52

1 57

5 63

9 68

A ve

ra ge

c us

to m

er ti

ck et

8. 09

9. 70

3. 42

4. 88

5. 96

7. 55

9. 21

3. 43

7. 67

8. 98

S am

e- st

or e

(C om

p st

or e)

s al

es g

ro w

4. 00

6. 00

1. 00

2. 50

5. 80

6. 70

6. 60

6. 10

So ur

ce :

A d

ap te

d fr

om L

ow e’

s A

nn u

al R

ep or

t d at

a fo

r 20

00 –2

00 6,

a cc

es se

d F

eb ru

ar y

21 , 2

00 8.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

60 8-

09 3

-1

E xh

ib it

1 0

L ow

e' s

In co

m e

St at

em en

t D at

a, 1

99 7–

20 06

(i n

$ m

ill io

ns )

L ow

e’ s

19 97

19 99

20 01

20 03

20 05

al es

1, 10

8 $1

3, 33

1 $1

5, 90

6 $1

8, 77

9 $2

2, 11

1 $2

6, 49

1 $3

0, 83

8 $3

6, 46

4 $4

3, 24

3 $4

6, 92

7 C

os t o

f G oo

ds S

ol d

$8 ,1

$9 ,7

$1 1,

52 5

$1 3,

48 8

$1 5,

74 3

$1 8,

46 5

$2 1,

23 1

$2 4,

16 5

$2 8,

44 3

$3 0,

72 9

G ro

ss M

ar gi

n $2

,9 53

,5 74

,3 81

,2 91

,3 68

,0 26

,6 07

2, 29

9 $1

4, 80

0 $1

6, 19

E xp

en se

el lin

g, G

en er

al a

nd A

dm in

is tr

at iv

e 1,

95 4

2, 34

1 2,

77 2

3, 34

8 3,

91 3

4, 73

0 5,

54 3

7, 56

2 9,

01 4

9, 73

8 S

to re

O pe

ni ng

C os

13 2

14 0

12 9

12 8

12 3

14 2

14 6

N on

re cu

rr in

g M

er ge

r C

os ts

ot al

02 7

2, 41

7 2,

89 4

3, 48

0 4,

05 3

4, 85

9 5,

67 1

7, 68

5 9,

15 6

9, 88

O pe

ra tin

g In

co m

e $

9 26

1 ,1

$ 1

,4 86

1 ,8

$ 2

,3 14

3 ,1

$ 3

,9 36

4 ,6

$ 5

,6 44

6 ,3

D ep

re ci

at io

n, In

te re

st In

co m

e (e

xp en

se )

D ep

re ci

at io

n 25

6 28

9 33

7 40

9 51

7 62

6 75

8 90

2 98

0 1,

16 2

In te

re st

1 17

4 18

2 18

0 17

6 15

8 15

E B

$ 5

$ 7

$ 1

,0 64

1 ,2

$ 1

,6 23

,3 59

2 ,9

$ 3

,5 36

4 ,5

$ 4

,9 98

In co

m e

T ax

P ro

vi si

21 6

28 7

39 0

47 2

60 1

88 8

1, 13

6 1,

36 0

1, 73

1 1,

89 3

N et

E ar

ni ng

s 38

3 50

0 67

3 81

0 1,

02 3

1, 47

1 1,

87 7

2, 17

6 2,

76 5

3, 10

So ur

ce :

A d

ap te

d fr

om L

ow e’

s A

nn u

al R

ep or

t d at

a fo

r 20

00 –2

00 6,

a cc

es se

d F

eb ru

ar y

21 , 2

00 8.

N ot

Fi sc

al y

ea rs

e nd

in g:

J an

u ar

y 30

, 1 99

8; Ja

nu ar

y 29

, 1 99

9; J

an ua

ry 2

8, 2

00 0;

F eb

ru ar

y 2,

2 00

1; F

eb ru

ar y

1, 2

00 2;

Ja nu

ar y

31 , 2

00 3;

Ja nu

ar y

30 , 2

00 4;

Ja nu

ar y

28 , 2

00 5;

F eb

ru ar

y 3,

2 00

6; a

nd F

eb ru

ar y

2, 2

00 7.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

Exhibit 11 Comparison of Share Price of Home Depot and Lowe’s, 2000–2006 in $USD

Source: Adapted from Google Finance data for December 31, 1999 through December 31, 2006, accessed February 21, 2008.

Exhibit 12 University of Michigan American Customer Satisfaction Index Scores, 2000–2006

Category: Specialty Retail Stores

Company 2000 2001 2002 2003 2004 2005 2006 Costco Wholesale Corporation 77 76 79 80 79 79 81

SAM’S CLUB (Wal- Mart Stores, Inc.) 74 78 77 77 75 76 78

All Others NM 72 73 73 75 73 75

Lowe’s Companies, Inc. NM 75 76 77 76 78 74

Specialty Retail Stores Average 76 73 74 74 75 74 75

Best Buy Co., Inc. NM NM NM 72 72 71 76

Circuit City Stores, Inc. NM NM NM 73 72 70 69

Home Depot, Inc., The NM 75 71 73 73 67 70

Source: http://www.theacsi.org/, accessed February 21, 2008.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

Exhibit 13 Comments to BusinessWeek Cover Story: “Renovating Home Depot,” March 27, 2006

Our Mar. 6 Cover Story, “Renovating Home Depot,” took a close look at the corporate culture that Chief Executive Robert L. Nardelli is trying to build—one that’s based in part on military concepts and personnel. We argued that the command-and-control discipline Nardelli has imposed is getting financial results. But a deluge of letters, online posts, and message board responses—nearly 300 in all—indicates the overhaul has come at a cost. Many readers connected the military-style ethic promoted by Nardelli to a decline in customer service at Home Depot stores. In a startlingly similar refrain, they complained of indifferent workers, long lines, and unpleasant stores. Only two correspondents praised America’s largest home supply store. What follows is a sampling of reader reaction, with a reply from Home Depot

CUSTOMER VIEWS

As a submarine Navy veteran, as well as a General Electric (GE) veteran (Heavy Military Electronics), I can only applaud Home Depot’s military-style management and Robert Nardelli. What’s missing is focus on the customers. I’m also a veteran home improver. I’ve watched Home Depot under Nardelli follow in the footsteps of Northern Virginia’s now-defunct Hechinger stores, while Lowe’s (LOW) eats Home Depot’s lunch.

Anonymous

Gross sales at Home Depot may be soaring, but the customer service is, well, just gross. For example, Home Depot stores have lots of checkout lanes, but often most of them are closed, and they attempt to push customers to the awful self-checkout area. I wrote a letter to Home Depot corporate and told them I did not wish to become their unpaid employee, even for five minutes.

Jeffrey E. Schmidt Kissimmee, Fla.

I searched “Renovating Home Depot (HD)” in vain for some reference to military veterans’ value as employees—besides their maturity, discipline, and a willingness to relocate to unsavory locales.

The veteran employee’s ability to “think on his feet” was praised, but the critical plans are now being made by Home Depot central management. I read no reference to how the plans are developed—for example, how those target sales numbers (nervously tracked by store managers on their BlackBerrys) are defined. What thinking is the veteran employee empowered to do?

Similarly, Home Depot has responded to its “strategic” needs by replacing its seasoned, well-compensated professional staff with a vast crew of business personnel who will be willing to accept more demands and question authority less. It seems that decisions are being made strictly on the basis of numbers, often without a sense of their root cause or interconnected nature.

Jennifer Kirley (Formerly HT1 Hull Maintenance Technician First Class, U.S. Navy), Greene, Me.

The first problem at Home Depot after Nardelli’s arrival was extensive violation of Carl Liebert’s “Customers cannot buy what we do not have” slogan. When Home Depot first appeared on the scene, before Nardelli, it corrected two flagrant problems that were Hechinger’s undoing: out-of-stock items, poor display of existing stock, and floor employees’ lack of responsiveness to customers. Home Depot, before Nardelli, was fully stocked and hired people in the trade who were available in force on the floor. They understood the stock and could help customers use it.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

The Home Depot, Inc. 608-093

Many of us now find broader and more complete stock, along with better staff availability and knowledge, at Lowe’s. I go more than an extra mile to shop accordingly. Home Depot would do well to change its focus from command-and-control to customers.

George F. Steeg Potomac Falls, Va.

I applaud CEO Bob Nardelli’s efforts to make the shopping experience consistent throughout each store. He has brought order to what was at times an experience in futility trying to find a particular item. Unfortunately, it provides me the speed and efficiency I crave for all the wrong reasons—to get in and out of the place as quickly as I can so that I don’t have to hear the workers gripe about yet another change being implemented by senior management.

Scott Haines Foster City, Calif.

I recently visited Home Depot to buy a simple polyvinyl chloride (PVC) elbow for my sprinkler system, something that probably would cost less than a dollar. After spending an inordinate amount of time just looking for the item, I went to the front of the store only to be greeted by one clerk and two checkout lines, 10 to 12 people deep. It wasn’t worth the wait. Shop again at Home Depot? Never.

Larry Paquette Fresno, Calif.

After reading the Cover Story, I have to say I am a Nardelli supporter. Having experience in the Marine Corps during Vietnam and also as a manager of 1,000 people in business, I see great parallels for success. Most successful businesses require balanced performance—financial, operational (or process improvement), customer focus, and developing people. To deem this militaristic when it is just good business is a bit naive. In today’s all- volunteer military, there is much more focus on balanced performance and people skills, as opposed to the blood-and-guts perception created by Hollywood. To keep the volunteers, the military has actually improved more quickly than some U.S. businesses. To hire people who are well-trained is an excellent strategic advantage for which I commend Mr. Nardelli.

Richard Jozwiakowski Round Rock, Tex.

* * * * *

HOME DEPOT RESPONDS

With 1.3 billion transactions a year, we’re bound to make mistakes, but nothing disappoints us more than letting down a customer. We’re making continuous improvements in our customer service levels and are working harder than ever to make sure that our service delivery meets the high expectations our customers have of Home Depot.

Jose Lopez Senior Vice President Chief Customer Officer, Home Depot Atlanta

Source: BusinessWeek.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

608-093 The Home Depot, Inc.

Endnotes

1 “Our History,” Home Depot Company website, http://www.corporate.homedepot.com, accessed January 23, 2008.

2 Brian Grow et al. “Out at Home Depot: Behind the flameout of controversial CEO Bob Nardelli,” BusinessWeek, January 15, 2007, http://www.businessweek.com/magazine/content/07_03/b4017001.htm accessed November 29, 2007.

3 The Home Depot, 2001 Annual Report (Atlanta: The Home Depot, 2001), p. 2, http://corporate.homedepot. com/en_US/Corporate/Investor_Relations/Annual_Reports/2001/complete_annualrpt.pdf, accessed March 6, 2008.

4 Grow et al. (January 15, 2007).

5 Ibid.

6 Ibid.

7 “How Nardelli Finally Helped the Stock,” BusinessWeek, Stocks in the News, January 3, 2007, www.businessweek.com, accessed January 15, 2008.

8 “DIY Retailing—US—March 2007,” Mintel Oxygen, Mintel International Group Limited, via Investext [accessed January 18, 2008.

9 The Home Depot, “Welcome to The Home Depot, Inc. Corporate Web Site,” The Home Depot Company Web site, http://corporate.homedepot.com/wps/portal, accessed January 22, 2008.

10 Ibid.

11 Pallavi Gogoi, “Home Depot’s Surprising Choice for CEO,” BusinessWeek, January 4, 2007, http://www.businessweek.com/bwdaily/dnflash/content/jan2007/db20070103_536329.htm, accessed January 15, 2008.

12 Pallavi (January 4, 2007).

13 Ibid.

14 The Home Depot, “Our History,” The Home Depot Company Web site, http://corporate.homedepot. com/wps/portal/!ut/p/.cmd/cs/.ce/7_0_A/.s/7_0_10D/_s.7_0_A/7_0_10D, accessed January 23, 2008.

15 “The Home Depot,” The New Georgia Encyclopedia, http://www.georgiaencyclopedia.org/nge/ ArticlePrintable/jsp?id=h-1886, accessed January 23, 2008.

16 The Home Depot, “Our History,” The Home Depot Company Web site, http://corporate.homedepot. com/wps/portal/!ut/p/.cmd/cs/.ce/7_0_A/.s/7_0_10D/_s.7_0_A/7_0_10D, accessed January 23, 2008.

17 Ibid.

18 Ibid.

19 Ibid.

20 Ibid.

21 Charan (April 2006), p. 5.

22 The Home Depot, “Our History,” The Home Depot Company Web site, http://corporate.homedepot. com/wps/portal/!ut/p/.cmd/cs/.ce/7_0_A/.s/7_0_10D/_s.7_0_A/7_0_10D, accessed January 23, 2008.

23 The Home Depot, “Stores, Products, and Services,” The Home Depot Company Web site, http://corporate.homedepot.com/wps/portal/!ut/p/c1/04_SB8K8xLLM9MSSzPy8xBz9CP0os3gDdwNHH0tD F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

100

The Home Depot, Inc. 608-093

U1M3g1APR0N31xBjAwgAykfC5H1MzN0MzDycDANMYdIGBHT7eeTnpuoX5EaUAwDOvP5h/dl2/d1/L2dJ QSEvUUt3QS9ZQnB3LzZfMEcwQUw5MTU1RjBVSEExR0NUMzAwMDAwMDA!, accessed March 20, 2009.

24 Alex Biesada, “Home Improvement and Hardware Retail—Industry Overview,” Hoover’s, Inc., www.hoovers.com, accessed March 6, 2008.

25 “Industry Profile: Home Centers and Hardware Stores,” Hoover’s, Inc., 2008, via Hoover’s accessed January 24, 2008.

26 “DIY Retailing—US—March 2007,” Mintel Oxygen, Mintel International Group Limited, via Investext, accessed January 18, 2008.

27 “Industry Profile: Home Centers and Hardware Stores,” Hoover’s, Inc., 2008, via Hoover’s, accessed January 24, 2008.

28 Ibid.

29 Ibid.

30 Ibid.

31 Hardware Retailing, cited in “DIY Retailing—US—March 2007,” Mintel Oxygen, Mintel International Group Limited, via Investext, accessed January 18, 2008.

32 “Lowe’s Companies—Profile,” Hoover’s Online, via Hoover’s, accessed January 24, 2008.

33 The Home Depot, “Our History,” The Home Depot Company Web site, http://corporate.homedepot. com/wps/portal/!ut/p/.cmd/cs/.ce/7_0_A/.s/7_0_10D/_s.7_0_A/7_0_10D, accessed January 23, 2008.

34 “Lowe’s Companies—Profile,” Hoover Online, via Hoover’s, accessed January 24, 2008.

35 Ibid.

36 Ibid.

37 Brian Hindo, “A Sharper Edge at Lowe’s,” BusinessWeek, January 15, 2007, http://www.businessweek. com/magazine/content/07_03/b4017006.htm, accessed December 4, 2007.

38 Hindo (January 15, 2007).

39 Andrew Ward, “Quiet Achiever Led Lowe’s to Prosperity,” FT.com, January 10, 2007, http://search.ft.com/ftArticle?queryText=Quiet+achiever+led+Lowe%E2%80%99s+to+prosperity&y=6&aje=fal se&x=13&id=070110009744&ct=0, accessed December 4, 2007.

40 Ward (January 10, 2007).

41 Charan (April 2006), p. 2.

42 Ibid.

43 Ibid.

44 Grow et al. (March 6, 2006).

45 Dan Morse, “Under Renovation: A Hardware Chain Struggles to Adjust to a New Blueprint,” Wall Street Journal, January 17, 2003, via Factiva, accessed February 15, 2008.

46 Charan (April 2006), p. 8.

47 Ibid., p. 2.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

101

608-093 The Home Depot, Inc.

48 Vanessa L. Facenda, “Cowboy Culture/GE Mentality: Bob Nardelli is trying to centralize operations and rein in Home Depot’s undisciplined past without losing the entrepreneurial spirit in the field,” Retail Merchandiser, August 1, 2002, via Factiva, accessed December 20, 2007.

49 Dan Morse, “Under Renovation: A Hardware Chain Struggles to Adjust to a New Blueprint,” Wall Street Journal, January 17, 2003, via Factiva, accessed February 15, 2008.

50 William A. Julian and Deidre Bane, “Home Depot,” Equity Research, Credit Suisse First Boston, February 25, 2000, p. 3, via Investext, accessed January 24, 2008.

51 Charan (April 2006), p. 2.

52 Dave Pennington, “Home Depot Leverages Enterprise Data to Increase Customer Satisfaction,” DM Review Magazine, January 2000, http://www.dmreview.com/issues/20000101/1784-1.html, accessed February 29, 2008.

53 The Home Depot, 2000 Annual Report (Atlanta: Home Depot, 2001), p. 1, http://corporate.homedepot.com/en_US/Corporate/Investor_Relations/Annual_Reports/2000/pdfs/hd2000. pdf, accessed December 2007.

54 The Home Depot, 2000 Annual Report (Atlanta: Home Depot, 2001), Founders’ Letter, http://corporate.homedepot.com/en_US/Corporate/Investor_Relations/Annual_Reports/2000/pdfs/hd2000. pdf, accessed December 2007.

55 William J Holstein, “Implementing Technology: IT Trials of a do-it-yourself enthusiast,” Financial Times, September 29, 2004, via Lexis Nexis, accessed December 4, 2007.

56 J. P. Donlon, “Time Out Between Nardelli Meltdowns,” January 9, 2007, http://www.chiefexecutive.net, accessed January 7, 2008.

57 Donlon (2007).

58 Neil Buckley and Betty Liu, “Fixer puts the final touches to a DIY refit,” Financial Times, July 8, 2003, p. 10, via Lexis-Nexus, accessed December 4, 2007.

59 Facenda (August 1, 2002).

60 William J Holstein, “Implementing Technology: IT Trials of a do-it-yourself enthusiast,” Financial Times, September 29, 2004, via Lexis Nexis, accessed December 4, 2007.

61 Neil Buckley and Betty Liu, “Fixer puts the final touches to a DIY refit,” Financial Times, July 8, 2003, p. 10, via Lexis-Nexus, accessed December 4, 2007.

62 Neil Buckley and Betty Liu, “Transcript of the interview with Robert Nardelli,” Financial Times, July 7, 2003, via Factiva, accessed December 4, 2007.

63 The Home Depot, 2001 Annual Report (Atlanta: Home Depot, 2002), p. 4, http://corporate.homedepot. com/en_US/Corporate/Investor_Relations/Annual_Reports/2001/complete_annualrpt.pdf, accessed December 2007.

64 “All Around the House,” HomeChannel News, December 12, 2005, via Factiva, accessed February 15, 2008.

65 Ibid..

66 Patti Bond, “A revenue-building project, Home Depot has quietly expanded its wholesaling operations into a $12 billion unit,” Atlantic Journal Constitution, August 6, 2006, via Factiva, accessed December 20, 2007.

67 From Betty Liu and Neil Buckley, “Transcript of the interview with Robert Nardelli,” Financial Times, July 7, 2003, via Factiva, accessed December 4, 2007.

68 J. P. Donlon, “Time Out Between Nardelli Meltdowns,” January 9, 2007, http://www.chiefexecutive.net, accessed January 7, 2008. F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

102

The Home Depot, Inc. 608-093

69 From Betty Liu and Neil Buckley, “Transcript of the interview with Robert Nardelli,” Financial Times, July 7, 2003, via Factiva, accessed December 4, 2007.

70 Renee Degross, “Five Years of Change: Home Depot’s Results mixed under Nardelli,” The Atlanta Journal- Constitution, January 1, 2006, via Lexis-Nexus, accessed December 4, 2007.

71 “Home Depot Shifts it Merchandise Buying,” Wall Street Journal, July 31, 2001, via Factiva, accessed December 20, 2007.

72 J. P. Donlon, “Time Out Between Nardelli Meltdowns,” January 9, 2007, http://www.chiefexecutive.net, accessed January 7, 2008.

73 John Caulfield, “Depot puts buying power in the hands of a dozen-Home Depot consolidates purchasing to regional,” September 3, 2001, http://findarticles.com/p/articles/mi_m0VCW/is_16_27/ai_78399495, accessed December 4, 2007.

74 Caulfield (September 3, 2001).

75 Ibid.

76 Ibid.

77 Tony Wilbert, “Home Depot implements another stop to cut costs,” Atlanta Journal-Constitution, May 20, 2003, p. 12, via Lexis-Nexis, accessed December 4, 2007.

78 Wilbert (May 20, 2003), p. 12.

79 From Betty Liu and Neil Buckley, “Transcript of the interview with Robert Nardelli,” Financial Times, July 7, 2003, via Factiva, accessed December 20, 2007.

80 Scott Larsen, “Heading in a Different Direction,” National Home Center News, December 17, 2001, via Factiva, accessed December 20, 2007.

81 Grow et al. (January 15, 2007).

82 “Chain Store Age,” Digital Depot, January 2004, via Factiva, accessed February 15, 2008.

83 Q3 2005 Home Depot Inc Earnings Conference Call—Final, Voxant FD WIRE, November 15, 2005.

84 The Home Depot, 2004 Annual Report (Atlanta: Home Depot, 2005), p. 2, http://ir.homedepot. com/downloads/HD_2004_AR.pdf, accessed December 2007.

85 Much of this section was informed by Hal Plotkin, “Six Sigma: What it is and how to use it,” Harvard Management Update, June 1999.

86 Plotkin (1999).

87 Ibid.

88 Ibid.

89 Ibid.

90 Ibid.

91 Harry R. Weber, “Home Depot retooling in face of challenge from Lowe's,” November 26, 2003, The San Diego Union-Tribune, via Factiva, accessed February 15, 2008.

92 John Caulfield, “Making it Happen in the Field: District Managers take vanguard role in guiding store associates in the right direction,” National Home Center News, December 17, 2001, via Factiva, accessed February 15, 2008.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

103

608-093 The Home Depot, Inc.

93 Neil Buckley and Betty Liu, “Fixer puts the final touches to a DIY refit,” Financial Times, July 8, 2003, p. 10, via Lexis-Nexus, accessed December 4, 2007.

94 Home Depot Inc. Analyst Meeting—Final, FD Wire, January 17, 2003, via Factiva, accessed February 27, 2008.

95 Dan Morse, “Under Renovation: A Hardware Chain Struggles to Adjust to a New Blueprint—Home Depot Chief Nardelli Tightens Central Control, and Employees Squawk—today he reveals more plans,” Wall Street Journal, January 17, 2003, via Factiva, accessed December 4, 2007.

96 Home Depot Inc. Analyst Meeting—Final, FD Wire, January 17, 2003, via Factiva, February 26, 2008.

97The Home Depot, 2002 Annual Report (Atlanta: Home Depot, 2003), p. 4, http://ir.homedepot. com/downloads/HD_2002_AR.pdf, accessed December 2007.

98 Home Depot Inc. Analyst Meeting—Final, FD Wire, January 17, 2003, via Factiva, February 26, 2008.

99Home Depot, 2004 Annual Report (Atlanta: Home Depot, 2005), p. 1, http://ir.homedepot.com/ downloads/HD_2004_AR.pdf, accessed December 2007.

100 Home Depot Inc. Analyst Meeting—Final, FD Wire, January 17, 2003, via Factiva, accessed February 27, 2008.

101 Home Depot Inc. Analyst Meeting—Final, FD Wire, January 17, 2003, via Factiva, accessed February 27, 2008.

103 Bond (December 8, 2002).

104 Grow et al. (March 6, 2006).

106 Bond (December 8, 2002).

107 Grow et al. (January 15, 2007).

108 Ibid.

109 Ibid.

110 Rebecca Zicarelli, “Home Depot’s Hardware Warriors: The DIY store is drafting former military officers to lead the charge against rivals Wal-Mart and Lowe’s,” Fast Company, September1, 2004, via Factiva, accessed February 27, 2008.

111 Martin Booe, “Reporting to the Depot,” Workforce Management, January 1, 2005, via Factiva, accessed February 27, 2008.

112 Martin Booe, “Agent of Change,” Workforce Management, January 1 2005, via Factiva, accessed February 27, 2008.

113 Grow et al. (March 6, 2006).

114 Ibid.

115 Terry C. Evans, “Home Depot management shift seen as healthy sign,” October 8, 2001, http://findarticles.com/p/articles/mi_m0VCW/is_18_27/ai_79353400, accessed December 4, 2007.

116 Grow et al. (March 6, 2006).

117 Ibid.

118 The Home Depot, 2002 Annual Report (Atlanta: Home Depot, 2003), p. 5, http://ir.homedepot. com/downloads/HD_2002_AR.pdf, accessed December 2007

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

104

The Home Depot, Inc. 608-093

119Home Depot, 2004 Annual Report (Atlanta: Home Depot, 2005), p. 1, http://ir.homedepot. com/downloads/HD_2004_AR.pdf, accessed December 2007.

120 Scott Larsen, “Heading in a Different Direction,” National Home Center News, December 17, 2001, via Factiva, accessed December 20, 2007.

121 Morse (January 17, 2003).

122 Ibid.

123 Grow et al. (March 6, 2006).

124 Facenda (August 1, 2002).

125 Budd Bugatch and Jessica Simmons, “Home Depot Incorporated,” Raymond James Equity Research, February 11, 2002, p. 2, via Investext, accessed January 18, 2008.

126 Donald Trott, Jennifer Malone, and Timothy Allen, “Home Depot: Conquering New Worlds,” Jeffries and Company, Inc., October 17, 2005, via Investext, accessed January 18, 2008.

127 Deborah L. Weinswig and Charmaine Tang, “Home Depot, Inc. Assuming Coverage: What Lies Beyond the Box,” Citigroup, October 28, 2005, via Factiva, accessed January 18, 2008.

128 Facenda (August 1, 2002). .

129 Deborah L. Weinswig and Charmaine Tang, “Home Depot, Inc. Assuming Coverage: What Lies Beyond the Box,” Citigroup, October 28, 2005, via Factiva, accessed January 18, 2008.

130 ISS US Proxy Advisory Services, “The Home Depot, Inc.,” March 28, 2006.

131 Donald Trott, Jennifer Malone, and Timothy Allen, “Home Depot: Conquering New Worlds,” Jeffries and Company, Inc., October 17, 2005, via Investext, accessed January 18, 2008.

132Grow et al. (January 15, 2007).

133 Ibid.

134 Ibid.

135 Grow et al. (March 6, 2006).

136 Debbie Howell, “Nardelli nears five-year mark with riveting record,” DSN Retailing Today, May 9, 2005, via Factiva, February 15, 2008.

137 Bond (December 8, 2002).

138 Patti Bond, “Nardelli Changing Culture,” Atlanta Journal-Constitution, December 8, 2002, via Factiva, accessed December 20, 2007.

139 From Betty Liu and Neil Buckley, “Transcript of the interview with Robert Nardelli,” Financial Times, July 7, 2003, via Factiva, accessed 20 December 2007.

140 Q2 2005 Home Depot Inc Earnings Conference Call-Final, FD Wire, August 16, 2005, via Factiva, accessed February 27, 2008.

141 Q1 2005 Home Depot Inc Earnings Conference Call-Final, FD Wire, May 17, 2005, via Factiva, accessed February 27, 2008.

142 Debbie Howell, “Nardelli nears five-year mark with riveting record,” DSN Retailing Today, May 9, 2005, via Factiva, February 15, 2008.

143 Grow et al. (January 15, 2007). Fo r

us e

on ly

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

105

608-093 The Home Depot, Inc.

144 Bond (December 8, 2002).

145 Grow et al. (March 6, 2006).

146 Patti Bond, “Home Depot Bumps up CFO,” Atlanta Journal-Constitution, October 13, 2006, via Lexis Nexus, accessed December 4, 2007.

147 Q2 2006 Home Depot Inc Earnings Conference Call-Final, Voxant FD Wire, August 15, 2006, via Factiva, accessed February 27, 2008.

148 Patti Bond, “Q&A: Bob Nardelli, Home Depot Chairman and chief executive,” Atlanta Journal- Constitution, July 9, 2006, via Lexis Nexus, accessed December 4, 2007.

149 Bond (July 9, 2006).

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

106

S w

9A98D005 INDUSTRIE PININFARINA: THE NEW CUSTOMER DECISION Professor Neil Jones prepared this case solely to provide material for class discussion. The author does not intend to illustrate either effective or ineffective handling of a managerial situation. The author may have disguised certain names and other identifying information to protect confidentiality. Richard Ivey School of Business Foundation prohibits any form of reproduction, storage or transmission without its written permission. Reproduction of this material is not covered under authorization by any reproduction rights organization. To order copies or request permission to reproduce materials, contact Ivey Publishing, Richard Ivey School of Business Foundation, The University of Western Ontario, London, Ontario, Canada, N6A 3K7; phone (519) 661-3208; fax (519) 661-3882; e-mail [email protected]. Copyright © 1997, Richard Ivey School of Business Foundation Version: 2011-04-27 The 25th of April is a national holiday in Italy, but it was not for Industrie Pininfarina (Pininfarina) top management in 1996. A meeting between Pininfarina and high level Mitsubishi executives lasted the entire day. The following day, a Friday, Renato Bertrandi, manager of operations at Pininfarina, sat in his office at the Pininfarina plant at Grugliasco, in the Piedmont region of Italy. In a rare quiet moment, he reflected on the challenges that lay ahead for the manufacturing operations. On Monday, he would recommend whether Pininfarina should accept European manufacturing responsibility for a new vehicle, the Mitsubishi Pajero. The vehicle presented both a major opportunity and a significant commitment, which would impact Pininfarina’s fortunes through the year 2004 and beyond and it would require major changes in manufacturing. The contract would virtually double Pininfarina’s output. Once again, Bertrandi thought through the company’s options and tried to evaluate the near-term benefits and challenges to manufacturing as well as the longer-term consequences. He thought with satisfaction about the many achievements in manufacturing since the 1980s. An active triathlete, he wondered where the next phase of the competitive race in the changing global automotive industry would leave the company. PININFARINA BACKGROUND In 1904, at the age of eleven, Battista “Pinin” Farina began work in his brother’s coach-making business — which also specialized in making seats for racecars. After long experience in the emerging and rapidly expanding Turin automobile industry, he founded his own company in 1930. Farina specialized in the design and production of custom and small series automobiles. While he expected to build relatively few “special” cars and was rooted in a tradition of highly skilled craftworkers, he was much impressed with the Ford system, which he had seen on a plant tour in the United States in 1920. The visit contributed to his conclusion that he had to draw on the strengths of Ford’s method to be successful. As he would later say,

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

107

Page 2 9A98D005

I was looking for a third state, between the craft we had to leave behind and industry. The state had to have industrial norms and structures but it must not suffocate that individual reality, which can be defined as style. There was no tradition to which we could appeal, our occupation was brand new and we paid for any mistakes we made in person.

The company soon earned a reputation for the quality and beauty of its designs. By 1939, Farina Industrie employed over 500 workers and manufactured close to 800 automobiles. For a period of time during World War II, the company product line included ambulances, airline seats, and stoves, but it returned to a focus on automobiles after the war ended. And it was in automobiles where it continued to find its greatest success — producing revered designs such as the Ferrari Berlinetta Dino and the Alfa Romeo Spider Duetto (Exhibit 1). Farina’s Cistalia automobile, designed in 1947, was celebrated in a collection of mobile sculptures at New York’s Museum of Modern Art. In 1954, after the great success of the Alfa Romeo Spider, the company added facilities to manufacture lower volume cars for major automobile manufacturers. To handle an increasing demand, in 1958 the company moved from Turin to a manufacturing plant in Grugliasco, a nearby suburb. Upon Farina’s death in 1966, management of the business was taken over by his son, Sergio, and his son-in-law, Renzo Carli. The family name and that of the business were changed from Farina to Pininfarina by presidential decree. Throughout the 1960s and 1970s Pininfarina continued to design and produce unique automobiles such as the Ferrari Berlinetta, the Lancia Flaminia, the Austin A 40, and the Morris 1100. By 1972, the company employed about 1900 people and was producing more than 23,000 cars per year. In 1979, Pininfarina split its design and manufacture divisions into Pininfarina Studi E Ricerche and Industrie Pininfarina (IPF), under the holding company Pininfarina S.p.A. In 1986, 30 per cent of the company’s shares were listed and sold on the Italian stock market, and a further three per cent of shares were sold to Mediobanca. However, the company remained closely held by the Pininfarina family who retained 67 per cent. THE NICHE MANUFACTURER Pininfarina was considered a niche car manufacturer. Niche manufacturers were chiefly distinguished by their low production volumes, which were often sub-contracted from a volume manufacturer. In Pininfarina’s case, typical production volumes ranged from only one or two cars per day (for example, the Bentley cabriolet) to perhaps 50 to 60 cars per day for “special” sedans such as the Fiat coupe (Exhibit 2). In contrast, volume manufacturers might produce a thousand cars per day or more at a factory dedicated to just a few models or even one model. However, not all volume manufacturers were the same. In the early 1990s, Japanese manufacturers on average produced around 70,000 cars per model per year, while an average European or American manufacturer produced around 200,000 per model per year. Japanese producers also had shorter model lives at around three years, while European producers had been averaging four to seven years of model life. Bertrandi felt that the best Japanese volume producers were profitable on very much lower production volume per model than even the Japanese average. One Japanese producer had told him that it would not consider outsourcing production volumes of greater than 5000 cars per year. In contrast, Pininfarina had produced over 17,000 Fiat Coupes in 1994. Volume producers could apply considerable pressure to niche producers to keep prices low. Usually, they had detailed knowledge of the product and production processes associated with a model and often had their own experience with part of the process. Further, a given volume producer was usually a vastly larger F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

108

Page 3 9A98D005

company and represented a high percentage of the niche manufacturer’s total business. Volume producer bargaining power was, therefore, high and niche manufacturer margins were narrow, especially during industry downturns. In general, margins were higher for fully assembled vehicles, and these offered more scope for production cost reductions to be achieved and captured by niche manufacturers. Profit margins for niche manufacturers were typically in the range of two to four per cent of the target unit manufacturing cost. Advantages of Niche Production Niche manufacturers provided three principal advantages to the volume producers who performed their own assembly for the vast majority of their production: niche manufacturers had lower total costs for cars made at low volumes, they could accept higher levels of volume uncertainty and their product designers brought both superior designs and famous brand names to volume producer models. First, niche manufacturer costs for small-volume products were lower than those usually achieved by volume assemblers. At low daily production rates, typical volume producer process designs were too expensive to implement. A typical volume producer might have capital and other fixed costs that were more than twice the level of a niche manufacturer. Niche manufacturers were forced to limit capital investments that were specialized to a particular model because the costs had to be amortized over fewer cars. As a result, niche manufacturers designed production processes that used general purpose equipment and required fewer dies, jigs and other specialized tools. Usually they had fewer mechanically performed operations and lower levels of automation. To achieve lower capital costs, a niche producer was also skilled in making tradeoffs between what could be accomplished by machine and what could be done by hand. Bertrandi explained:

It is mainly our engineering that provides an advantage. We get the product to 90 per cent with our process and compensate with skilled labor to provide the last 10 per cent. For example, we might decide to stamp a door in three stages instead of the four a volume producer would use. This can result in some small waves in the door metal, but we can correct this by adding five minutes of additional handwork. This can work at a production of 30 cars per day, but it would be suicide at 1,500 cars per day.

A more highly skilled workforce than that typically found at a volume producer was used to assemble parts and ensure quality in fit, finish and function. Many niche producers did not use a continuously moving assembly line. The variety of work performed at each station led Pininfarina, for example, to design a stop and go process, with a time between moves that might vary from about 10 minutes to about half an hour, and even up to eight hours, depending on the volumes needed. Niche manufacturers sometimes based their product and process designs on modifications to a higher volume design that was being produced at a major automobile assembler. Often, such modifications required more skill of the workforce than would be required if a similar product had been designed from scratch. For example, Pininfarina production of the Peugeot 406 coupe was based on a sedan model produced at Peugeot. Pininfarina took bodies supplied by Peugeot and inserted a stamped part that altered the slope from the roof to the trunk lid in the rear. This alteration demanded a critical hand weld at the intersection of the mass-produced and niche-produced metal parts. Achieving proper part position and a strong, flat weld, which could be properly finished, was essential to ensure quality in this operation. The

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

109

Page 4 9A98D005

higher proportion of labor required imposed an additional cost of some four to eight labor hours per car for a niche manufacturer. The second advantage of niche manufacturers was flexibility. Consequently, they were often given contracts on models that had higher than usual volume uncertainty and larger seasonal fluctuations in sales. For example, a convertible or cabriolet might have sales in the spring that were 150 per cent of the low sales in winter. Lifetime sales and the model life of highly specialized niche vehicles were also highly uncertain, as such products were aimed at narrow consumer segments that were difficult to specify and had rapidly shifting tastes. Some of the risks associated with such products could be shifted to niche producers. Contracts typically did not fully compensate niche producers for the costs of unanticipated volume fluctuations on a seasonal or overall basis. Uncertainty over model life complicated niche manufacturer planning for new model introductions, as, for example, models with low sales might be discontinued. Niche manufacturers coped by configuring their facilities to be flexible and by developing elaborate contingency plans. Contingency planning allowed the niche producers to rapidly shift workers from one line to another as demands fluctuated. For example, work at a given station, which might be carried out by a team of five during high volume periods, could be reduced to a team of two when volumes were low. Fewer people at a station meant that each worker had to perform a greater number and variety of the operations needed, and it usually increased station time so that the line moved more slowly. The line also had to be rebalanced so that each work station’s output rate was matched to keep worker idle time to a minimum. A line that needed a higher output rate would have more workers at a station and would assign fewer and narrower tasks to each worker. When necessary, workers could be temporarily laid off or could be asked to work overtime. The third benefit provided by niche manufacturers was highly competent and often renowned design skills in product and process. Design services were an independent source of revenue for some niche manufacturers. At Pininfarina in 1994, design and engineering revenue totalled nearly £90 billion1 and was growing rapidly. Work might be performed for a production model or for prototype cars, which might never go into production. Although manufacturing contracts were not always awarded for suitable models that had been completed by niche manufacturers, participation in design significantly increased the chance of winning manufacturing business. Close links and effective joint problem-solving between design and manufacturing were considered a major advantage in the success of a new car model. Some designers, such as Pininfarina and Bertone, had widely recognized brand names. These brands were believed to command premiums and suggest luxury, fashion and high performance. Although Pininfarina’s major customers, Fiat and Peugeot, reported that they made little, if any, money on niche models, Bertrandi suspected their calculations excluded the positive impact of image and the attraction of niche cars in pulling potential buyers of other cars to showrooms. PININFARINA POSITION IN THE 1990s After relatively high profits in the late 1980s, the European auto market became less hospitable in the 1990s. Industry returns on net assets fell from their 1980s high of 10 per cent to 15 per cent to below five per cent on average in the 1990s. In the view of many, the primary problem was capacity utilization, which had averaged below 75 per cent from 1990 to 1995.

1US$1 was equal to approximately £1,600.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

110

Page 5 9A98D005

Over-capacity was partly the result of low underlying growth in the Western European consumer base and partly due to the addition of new factories by globalizing competitors. The spread of more efficient manufacturing practices, which had been pioneered by Japanese firms, also contributed to capacity growth. Although shares had been relatively stable overall, “voluntary” Japanese restraints on European market share were due to expire by the year 2000 in many large markets and Korean firms had begun to build a large European presence. In markets without restraints, Japanese producer shares were considerably higher than in markets with them. Exhibit 3 shows industry sales and share in Western Europe in the 1990s, and Exhibit 4 shows data on customer satisfaction. At the beginning of the 1990s, European producers had lagged behind other global competitors in some key areas of performance, and despite improvements were not believed to have fully closed the gap. Exhibit 5 shows comparative regional data. Manufacturing Operations at Pininfarina As it entered the 1990s, Pininfarina produced both bodies and fully assembled cars at two major production facilities, one in Grugliasco and the other at San Giorgio, about 40 kilometres away. The Grugliasco complex housed a full-scale wind tunnel test facility, which had been one of the first of its kind in the world. Production at Grugliasco was divided among three major buildings. In one, parts stamped by Pininfarina’s suppliers were welded together to form the basic “Body in White” (BIW), so named because the completed bodies were not yet painted. Suppliers made stamped parts to specifications set by the designing firm — often, but not always Pininfarina Studi E Riserche. The stamping process itself — the sequence of steps whereby the metal was formed, was typically specified by Pininfarina process design engineers. A second building contained the paint shop, which painted all production models. The paint shop performed six major steps, some separated by drying phases. In the paint shop, the bare steel was first galvanized, then phosphate-coated and given an electrostatic treatment. Next, a primer was applied and then a base coat, before a final clear coating completed the process. The paint shop had been upgraded in stages beginning in 1985 at a total cost of some £100 billion. It was initially designed for a capacity of 100 cars per shift, but its capacity had been increased to 140 cars per shift and then 160 cars per shift. Throughput time was about seven hours. The limited number of models produced by Pininfarina came in a total of 52 possible colors. The paint shop could change colors in about one minute, but required some manual setup to paint a specific model. At each arrival, the paint shop changed colors and set up for the appropriate model. Some cars, which needed special painting processes, were painted in a special area. The Rolls-Royce Bentley model, for example, required multiple steps of coating and surface preparation to achieve an adequate finish. About 100 hours of labor were required. The last step — the trim facility — installed all the rest of the parts needed to form the complete automobile. Here, engines, suspension and other mechanical parts from suppliers were installed, as were details of interior and exterior finish — from door seals, seats and instrument panels to exterior mirrors and bumpers. Trim steps were greatly complicated by the wide variety of options that were supplied to customers. For example, each Fiat model came with a choice of five different engines, and each was configured slightly differently in the engine compartment. Interior options and other options also increased the complexity of process control in assembly and resulted in inventory levels higher than comparable higher volume facilities. At San Giorgio, a more modern trim facility had been built in 1985 primarily for the Cadillac Allante business. Pininfarina’s test track was also located at San Giorgio. F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

111

Page 6 9A98D005

Improvements in the 1990s In 1992, Pininfarina faced a crisis. Production of bodies for the Cadillac Allante and Peugeot 205 and assembly for the Alfa Romeo Spider were being rapidly phased out, while volume replacement sufficient to maintain existing production levels had not yet been committed for new models (Exhibit 7). The shortfall eventually left 1993 production at less than 50 per cent of the average level for the 1990s to that point. Margins were also squeezed as European prices fell. Customers had begun to press for operational improvements in quality, cost and deliverability. Further, the company had by now concluded that despite some recent operational improvements, more fundamental and far-reaching changes to improve its manufacturing performance would be necessary to ensure future viability. Faced with deteriorating financial results, Pininfarina laboriously negotiated with its unions. The resulting accord, signed on July 28, 1992, was viewed by many as a new model for Italian labor relations. It called for the early retirement and “temporary” layoff2 of some 435 blue-collar employees — 50 per cent of the total workforce. Workforce and Quality Initiatives Two major changes were introduced with the new accord, designed to allow Pininfarina to improve its operations to near Japanese levels, while adapting to Italian conditions. First, Pininfarina introduced a work team system modelled on the Toyota NUMMI plant in California, including systems to track morale and elicit suggestions for improvement. Second, a program of training for shop-floor workers was instituted. The training program had two major components. First, skills were built in specific operations and techniques (for example, statistical process control and problem-solving techniques). Second, workers were given interpersonal skills training intended to develop the capability of the workforce to work in teams (doubts had been expressed about the potential for Italian workers to submerge a pride in individuality to the constraints of teamwork). A training program for new workers was also instituted. The training programs were a complement to an expansion of the quality initiative that had been underway since the middle 1980s. Renato Bertrandi had originally joined the company in 1986 as a manager of quality control, reporting to the general manager. After the accord of 1992, the quality control function reported to Bertrandi himself at the operations manager level. Pininfarina, while adopting some of the methods and practices of the quality movement, decided to adapt the philosophy to Italian and niche producer conditions. Renato Bertrandi explained:

As a first step in our situation, it is better not to stop the line for most types of problems. Stopping the line lowers our production and costs us more. It is better for us to have highly skilled people at the end of the line fixing problems after they have occurred. Of course, we also ask workers to identify problems they can’t fix on the line and work to remove the source of some problems. I realize this violates the philosophy of lean production, but it doesn’t pay to fix the root causes of all of our problems now. This will be our next step.

2Under Italian law, the government would, under certain circumstances, use a fund created by Italian companies to pay 80 per cent of a laid off worker’s salary for up to two years.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

112

Page 7 9A98D005

Supplier Development At about the same time, and in concert with the quality program, major programs were also initiated in supplier relations. In 1991, Pininfarina had about 650 suppliers. Typically, competitive bids were held among suppliers who were asked to meet Pininfarina’s predetermined design specifications. Volumes were then split among several suppliers. By 1993, the number of suppliers had been reduced to 350, despite a major decision to outsource the stamping operations, which had been lagging in the capital investment required to keep them competitive. This reduction was achieved by concentrating volumes in fewer, more capable suppliers, with whom Pininfarina worked more closely — even doing joint design work and parts planning. Major efforts had been made with the reduced number of suppliers to increase the frequency of deliveries, to correspondingly reduce their size, and to increase quality while decreasing the total amount of combined inspection. In the 1980s, incoming parts inspection employed 70 people to inspect all incoming supplier parts. In 1993, 30 people inspected only about 20 per cent of the incoming parts. Pininfarina also believed purchase prices and inventory levels had been improved. By 1996, the number of suppliers had increased again to 450, driven by the new production models and a shift in mix toward the assembly of complete vehicles. This trend had been offset slightly because new business was with existing customers who had substantial carryover of existing suppliers and similar needs. About 25 people were needed to manage these suppliers. Pininfarina could control the choice of supplier for about half of its purchase monetary volume and could negotiate freely on price for about two-thirds of its volume. The other one-third came mostly from major customers, who were also major parts suppliers. The progress of Pininfarina in achieving improvements in some key operating parameters is shown in Exhibit 6, and financial results and operating statistics are shown in Exhibit 7. Bertrandi was pleased with the fact that of the 20 per cent of cars produced that did not go immediately to a buyer, only 10 per cent of these were due to quality problems. The remaining 90 per cent were due to parts shortages of one type or another, typically the result of last-minute changes in option mixes in the production schedule and a consequent shortage of the correct part. Search for a New Customer Budgeted improvements called for further increases in the productivity of direct labor of three per cent annually. To utilize the extra capacity created by productivity improvements, to leverage its newly achieved production skills, and to diversify its risk of lower future volumes from current customers, Pininfarina decided to seek a third major customer in late 1994. Although some production of new models from existing customers would begin in 1996, these volumes would be insufficient to replace the production that would be phased out by the year 2000. The typical lead-time from the beginning of design until the first production vehicle was 38 months. Pininfarina actively marketed itself in the auto industry for new business in product design and manufacture, sometimes proposing joint development of prototype design projects for niche vehicles to volume manufacturers. Such projects were a source of profit to Pininfarina’s design division and could result in new manufacturing work for vehicles that were to go into production. Although at present, only the Peugeot 406 had been wholly designed at Pininfarina, the interior of the Fiat coupe had been a Pininfarina design, and Pininfarina had competed with Fiat’s internal designers for the exterior as well. Bertrandi felt that the present low level of relationship between manufacturing and design projects at

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

113

Page 8 9A98D005

Pininfarina was unusual and did not represent a trend among volume manufacturers to make the two functions more independent of one another. Pininfarina had worked with Peugeot’s styling centre to create all of its major models since the middle 1950s and had regularly performed work for Fiat for an even longer time. In addition, General Motors had been a large customer. While the identity of niche manufacturers’ design customers was a closely guarded secret, outside investment analysts’ reports stated that Mercedes, BMW, Porsche, and Honda were among Pininfarina’s current design and development customers. Analysts had anticipated the announcement of a major new manufacturing customer in 1995, but, as yet, no firm commitments from those prototyping and developing cars with Pininfarina had been received. Pininfarina’s prospects for a new niche vehicle- manufacturing customer remained good, however. THE NEW CUSTOMER DECISION Following a marketing contact with Mitsubishi proposing a niche vehicle product design project, in July 1995 Pininfarina was surprised to receive a counterproposal from Mitsubishi. Mitsubishi proposed that Pininfarina be the manufacturer of one of their sport utility vehicles, the Mini Pajero, which was to be marketed in Europe and Asia. A Pajero built in Japan was being successfully sold in Europe. A new model was already designed to the prototype stage and would be introduced first in Japan, in 1998. Vehicles for Asian sales would be manufactured by Mitsubishi; however, Mitsubishi proposed that Pininfarina adapt the design and manufacture in Italy for all of Europe. The major design work would be in adding a left-hand drive model and in adapting the process design to Pininfarina’s capabilities. Bertrandi was particularly surprised at this offer since to that point, Mitsubishi had not asked to visit or inspect Pininfarina’s factories — a common practice of volume manufacturers, who wished to verify Pininfarina’s manufacturing capabilities. Beyond the excellence of Pininfarina’s reputation and recent performance improvements, Bertrandi suspected that Mitsubishi had factored Italy’s relatively low automotive labor costs into their choice. Bertrandi believed these were one-half Germany’s levels (see Exhibit 8). The details of the Mitsubishi proposal had not been fully specified, but the basic characteristics of the proposal were clear. Based on previous experience Bertrandi believed any decision to proceed would be taken with many details not completely specified. Mitsubishi proposed that by no later than May 1999, Pininfarina should begin volume production in Europe of the new model, after a three-month trial and debugging period. Production would be at a rate of about 150 vehicles per day. Mitsubishi would pay Pininfarina a standard margin on a target cost that would be based on Mitsubishi’s own experience in producing the model in Japan and correction for differing process, parts and transportation costs. The standard margin had not been set, but it was clear it would be low — perhaps one-half of the two to four per cent margins Pininfarina earned on its current production contracts. Bertrandi believed that if Pininfarina could achieve production costs below Mitsubishi’s target, Pininfarina would be able to keep the additional profit. Mitsubishi would guarantee that total volumes would be at least sufficient for Pininfarina to recoup any model-specific capital costs. However, Pininfarina would have to bear the risk of investment in general purpose equipment such as the basic facilities themselves or robots, which could be used for other purposes. The exact guaranteed volume would be calculated on the basis of the standard margin that was allowed Pininfarina by the production contract. Total investments were expected to be £300 billion. General purpose capital equipment for a new model was usually in the range of 10 to 15 per cent of total investment. As Mitsubishi had roughly a three per cent share of the global automobile market, and over F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

114

Page 9 9A98D005

US$20 billion in worldwide sales, Pininfarina management had few doubts as to Mitsubishi’s ability to meet its commitments. The term of the production phase would be five years, expiring in 2003, with no obligations on either side to continue the arrangement with other models or services, beyond those which might be part of the Pajero contract such as warranty obligations or spare parts production. Revenues to be collected by Pininfarina each year on average over the life of the project were expected to be £900 billion. Some design changes would be needed for Europe. These were well within Pininfarina’s capabilities, although the model development time would be less than the approximately three years needed for a typical design project. As long as the Japanese schedule was kept, Bertrandi felt product and process design changes could be made in time easily, since the model was already in the prototype phase. In process design, Mitsubishi would design much of the process. Pininfarina had only to adapt the process to its facility — designing an appropriate flow and layout — and to adapt certain processes to a somewhat more labor-intensive system. Bertrandi felt such differences would mainly be in the BIW area where Japanese producers had a tendency to place more robots than American or European producers. Capital Investment As a result, new production facilities would have to be acquired and equipped for Mitsubishi production. Bertrandi felt confident such facilities could be built or acquired in time since potential expansion sites had already been identified near Grugliasco. Basic facilities were expected to cost somewhat in excess of £4 billion, including land, the trim facility and adequate parking for workers. Mitsubishi would not cover these expenses. Pininfarina would not invest in welding automation for the Pajero. The paint shop, which was currently running at capacity, would have to be run for an additional shift. To supply it, the logistics for transporting BIW from the Mitsubishi facility to the Grugliasco paint shop and back would have to be set up, but this posed no problem in principle since BIWs were already being painted at Grugliasco and transported to San Giorgio for trim. There would be additional expenses with the paint shop, however, associated with the Mitsubishi production. Currently, the necessity of cleaning the painting system with solvent to change paint color after each car placed Pininfarina near the limits of what would be acceptable under Italian pollution control regulations. Additional volumes for the Mini Pajero would force a switch to a water-based system. The Pajero would offer a two-tone painting option, and this also posed some problems for the paint shop. Two-tone painting required additional space to dry and store vehicles in between painting stages, and this space would also have to be created. Quality Although they had not yet been definitively set, Bertrandi knew that Mitsubishi considered its own quality standards to be very high and that its focus would differ substantially from those of Pininfarina’s existing customers. Some in the company believed Mitsubishi might demand defect levels of one-fourth the level of Pininfarina’s current customers, although what would be considered a defect was not clear. Experience had shown different customers considered different things in deciding what was a defect. For example, some customers closely specified the routes and positions of electrical harnesses and hoses in the engine compartment, while for other manufacturers, only the functionality was considered, aside from ensuring no basic hazards such as a plastic hose resting on a hot running part existed. F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

115

Page 10 9A98D005

Parts Supply and Logistics Major mechanical parts, including engines, would be supplied by production in Japan — either from Mitsubishi itself or one of its suppliers. Other parts would be sourced from Europe, predominantly from suppliers who Mitsubishi qualified. Parts supply and logistics from Japan would have to be established jointly with Mitsubishi. Mitsubishi agreed to own the inventory until it arrived at Pininfarina, but it would be shipped at Pininfarina’s request. Pininfarina would be responsible for having sufficient parts on hand to meet its production obligations. Pininfarina had some experience in long distance supply chains. In the 1980s and early 1990s, it had shipped BIW Cadillac Allantes to Detroit for final assembly. However, the supply chain to Japan was even longer and Mitsubishi and Pininfarina calculated that some 13 days shipping would be required, and a further three days of inventory at port in Italy would be needed, in addition to the normal supply of inventory at the plant. Pininfarina logistics staff believed these inventories would be adequate to ensure supply even in the event of strikes. During a strike or port closure, contingency plans would be established to divert production to a free port and to ship to Turin overland. Many parts from within Europe would also be shipped further than was usual for Pininfarina operations. At present, the most distant supplier was 900 kilometres away for Peugeots and 65 per cent were within 60 kilometres. However, Mitsubishi, which had production in Holland, wanted to retain many suppliers with which it had familiarity. Many of these suppliers were outside Italy, with the most distant being in the United Kingdom — a three-day shipping distance. Many of the new suppliers would be unfamiliar to Pininfarina and would present challenges. Despite the presence of Mitsubishi, Pininfarina was responsible for parts supply and negotiation of price. The volumes needed were much less than those usual for volume manufacturers, and this made negotiations of price and delivery difficult compared with what could be accomplished by larger firms. Suppliers often incurred extra costs in overhead and packaging and shipping costs, as well as additional set-up costs to supply smaller orders. Many of the suppliers would be unfamiliar to Pininfarina and might increase the number of Pininfarina’s suppliers by 150 or so. Outgoing logistics would also be more complex than usual since Pininfarina would be shipping greater volumes to dealers in major European markets, but this was not expected to present insurmountable problems. It would also be necessary to forecast sales further in advance since the interval from parts order to arrival would be about 46 days — considerably longer than for the current models. For example, for Peugeot, the current standard for orders was about 10 days in advance of production. Workforce While Mitsubishi would pay for tooling and fixtures under the volume guarantees at standard margins, the workforce needed was another matter. Bertrandi felt some 600 additional direct workers (inside the factory) would be needed to meet the Mitsubishi volume needs. At traditional ratios of direct to indirect workers, this would imply some 200 to 240 indirect workers. However, Bertrandi felt that while some classes of indirect workers could not be reduced from usual levels, other classes could. For example, purchasing might not need a fully proportional addition to staff because some suppliers would still be in common, some additional capacity could be added in information systems and present resources were not fully utilized. Bertrandi estimated that only one-half the historical number of additional indirect workers might be needed. F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

116

Page 11 9A98D005

Increasing productivity would free some direct labor capacity by 1999. Further, by the year 2000, existing contract business would be ramped down so that some of the current direct labor force could be freed to work on the Mitsubishi vehicle. The Pajero might also present Pininfarina with a learning opportunity. Bertrandi anticipated that work methods, jigs and tools would be refined in Japan prior to beginning production in Italy. Using Japanese designs and production tools and processes, Bertrandi was excited by the prospect of being able to compare his operations with world-class volume manufacturers. Pininfarina would need to learn fast. After the initial test and ramp-up phase, which would last two to three months, Pininfarina would have to meet the agreed target costs or else pay for any overages itself. Bertrandi wondered if he should accept the contract and the challenge.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

117

Page 12 9A98D005

Exhibit 1

ALFA ROMEO SPIDER DUETTO

FERRARI BERLINETTA DINO

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

118

Page 13 9A98D005

Exhibit 2

FIAT COUPÉ

BENTLEY AZURE Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

119

Page 14 9A98D005

Exhibit 3

WESTERN EUROPEAN MANUFACTURER SHARE OF TOTAL NUMBER OF VEHICLES (in per cent)

Manufacturer 1990 1991 1992 1993 1994 1995

VW Group 15.6 16.4 17.5 16.4 15.8 16.7 General Motors Group 11.3 12.1 12.4 13.0 13.1 13.1 Opel/Saab/GM 11.3 12.1 12.4 13.0 13.1 13.1 Lotus 0.01 0.01 Peugeot-Citroen 12.9 12.1 12.2 12.3 12.8 12.0 Peugeot 8.2 7.6 7.4 7.4 7.7 7.0 Ford Group 11.6 11.7 11.3 11.3 11.9 11.9 Ford 11.4 11.7 11.2 11.2 11.8 11.7 Jaguar 0.1 0.1 0.1 0.1 0.1 0.1 Fiat Group 14.2 12.8 11.9 11.1 10.8 11.1 Fiat* 10.3 9.3 8.8 8.3 8.6 8.7 Lancia 2.3 2.0 1.7 1.6 1.4 1.4 Alfa Romeo 1.5 1.4 1.2 1.1 0.8 1.1 Innocenti 0.06 0.11 0.10 0.11 Ferrari** 0.019 0.021 0.023 0.018 Maserati 0.015 0.011 Renault 9.8 10.0 10.6 10.5 11.0 10.3 Mercedes 3.3 3.3 3.0 3.1 3.5 3.4 BMW 2.8 3.1 3.3 3.2 3.3 3.3 Rover*** 2.9 2.6 2.5 3.2 3.3 3.1 Nissan 2.9 3.3 3.2 3.5 3.2 3.0 Toyota/Lexus 2.7 2.7 2.5 2.7 2.6 2.5 Mazda 2.1 2.1 2.0 1.7 1.5 1.4 Volvo 1.8 1.5 1.5 1.5 1.7 1.8 Mitsubishi/DMS 1.3 1.4 1.2 1.2 1.0 1.1 Honda 1.2 1.3 1.3 1.4 1.4 1.5 Hyundai 0.1 0.3 0.6 0.7 0.7 0.8 Suzuki/Maruti 0.7 0.7 0.9 0.9 0.7 0.8 Chrysler 0.3 0.3 0.3 0.5 0.5 0.6 Subaru 0.4 0.4 0.3 0.4 0.3 0.3 Porsche 0.1 0.1 0.1 0.1 0.1 0.1 Others 2.0 1.8 1.5 1.2 0.9 1.4

Total 100 100 100 100 100 100 As Per cent of 1990 100 102 102 86 90 91 Total Vehicles 13,258,807 13,504,345 13,497,536 11,428,352 11,910,952 12,012,415

* Includes Innocenti, Ferrari and Maserati after 1993 ** Includes Maserati after 1991 *** Part of BMW Group after 1991

Source: Company Files

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

120

Page 15 9A98D005

Exhibit 4

CUSTOMER SATISFACTION SURVEY SAMPLE RESULTS BY MAKE (UNITED KINGDOM DATA)

Overall Index Parts /Service Problem Incidence /Resolution

Subaru 142 145 150 Honda 141 151 145 Daewo 140 134 89 Mazda 137 136 145 Toyota 137 142 141 Jaguar 136 158 118 Nissan 136 127 138 BMW 132 157 121 Daihatsu 130 129 137 Mercedes 128 159 131 Mitsubishi 124 117 129 Saab 124 139 112 Audi 119 109 118 Suzuki 116 101 122 Volvo 115 116 103 Hyundai 111 102 95 Renault 106 107 109 Volkswagen 105 104 105 Citroen 100 97 96 Rover 100 108 95 Peugeot 99 96 98 Fiat 95 88 94 Alfa Romeo 94 74 80 Ford 78 72 79 Lada 62 81 28 Total Industry 100 100 97

Source: Company Files

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

121

Page 16 9A98D005

Exhibit 5

SUMMARY OF ASSEMBLY PLANT CHARACTERISTICS: VOLUME PRODUCERS (1989)

Japanese in Japan

Japanese in North America

American in North

America

All Europe

Performance

Productivity (hours/vehicle)* 16.8 21.2 25.1 36.2

Quality (assembly defects/100 vehicles) 60 65 82.3 97

Layout

Space sq. ft./vehicle/year 5.7 9.1 7.8 7.8

Size of repair area (as % of assembly space) 4.1 4.9 12.9 14.4

Inventories (days) 0.2 1.6 2.9 2

Workforce

% of workforce in teams 69.3 71.3 17.3 0.6

Job rotation (0 = none, 4 = frequent) 3 2.7 0.9 1.9

Suggestions/employee 61.6 1.4 0.4 0.4

Number of job classes 11.9 8.7 67.1 14.8

Training of new production workers (hrs) 380.3 370.0 46.4 173.3

Absenteeism 5 4.8 11.7 12.1

Automation

Welding (% of steps) 86.2 85 76.2 76.6

Painting (% of direct steps) 54.6 40.7 33.6 38.2

Assembly (% of direct steps) 1.7 1.1 1.2 3.1

* Includes all labor within factory walls. Source: Womack, J. P., D. T. Jones, et al. (1990). The Machine That Changed the World. New York, Rawson Associates.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

122

Page 17 9A98D005

Exhibit 6

PININFARINA ASSEMBLY CHARACTERISTICS

1992 1996

Performance

Productivity1* (hours/vehicle) 60 42.5

Rework Cost** (% of Total) 12 – 15 9

Layout

Space sq. ft./vehicle/year 380.25

Size of repair area (as % of assembly space) N/A

Inventories (days) .5 – 3

Workforce

% of workforce in teams 0.25 95

Job rotation (0 = none, 4 = frequent) 3 2.7

Suggestions/employee 0 0.1

Number of job classes 4

Training of new production workers (hrs) N/A

Absenteeism 7.7 6

Automation

Welding (% of steps) 5 5 – 34

Painting (% of direct steps) 35 40

Assembly (% of direct steps) 5 5 * Includes all labor within factory walls. ** Includes cost of rework labor and materials only. Source: Company Files

1Includes all labor within factory walls.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

123

Page 18 9A98D005

Exhibit 7

PININFARINA DATA

1989 1990 1991 1992 1993 1994 1995

Sales (group total in billion lira) 372.0 479.5 501.9 412.4 417.2 731.4 670.0 Expenses:

Purchases & Services 270.7 258.7 175.2 201.0 556.6 410.4 Labor Cost 77.6 89.4 97.4 93.5 83.6 116.6 Depreciation 14.6 13.6 16.2 13.5 16.2 18.2 SG&A 92.7 113.4 121.8 98.0 72.5 117.1

Total Expenses 357.6 455.6 475.0 410.6 406.0 728.9 662.3 Operating Income 14.4 24.0 27.1 1.8 11.2 2.5 7.7

Production Model Mix 1989 1990 1991 1992 1993 1994 1995

Assembled Vehicles Lancia Thema SW/K SW 3,010 3,456 2,536 1,894 1,310 806 0 Alfa Romeo Spider 3,978 7,106 9,073 3,640 1,956 Fiat Coupe 276 17,332 12,500 Bentley Azure 3 170 250 Peugeot Coupe

Total Assembled Vehicles (Units) 6,988 10,562 11,609 5,534 3,545 18,308 12,750 Total Revenue - Assembled Vehicles (billion lira) 153 237 252 142 104 426 312 Revenue Per Assembled Vehicle (Million Lira) 21.9 22.4 21.7 25.7 29.3 23.3 24.5

Bodies Ferrari (Testa Rossa, 512TR, 456GT) 1,207 1,312 1,565 870 306 625 600 Cadillac 3 3,775 2,495 2,660 1,978 Peugeot 205 cabriolet 9,303 11,051 12,982 l1,718 3,450 784 Peugeot 306 cabriolet 414 11,154 11,600

Total Bodies 13,565 16,138 17,042 15,248 6,148 12,563 12,200 Total Revenue - Bodies (billion lira) 157 172 167 160 85 143 153

Workforce 1989 1990 1991 1992 1993 1994 1995

Direct Workers 803 964 889 846 824 927 864 Indirect Workers 443 444 431 408 352 351 344 Total Workers 1,246 1,498 1,320 1,254 1,176 1,278 1,208

Note: US$1 = £1,600 (approximately) Source: Company Files

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

124

Page 19 9A98D005

Exhibit 8

1996 LABOR COSTS IN THE AUTO INDUSTRY (DM per hour)

Germany 62.44 Belgium 44.6 Sweden 41.8 Japan 41.56 United States. 38.52 Netherlands 34.75 France 33.08 Spain 28.06 Italy 27.79 United Kingdom 27.08

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

125

10.

9-610-022 N O V E M B E R 5 , 2 0 0 9

________________________________________________________________________________________________________________ Professors Gary Pisano (Harvard Business School) and Pamela Adams (Franklin College) prepared this case. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 2009 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to www.hbsp.harvard.edu/educators. This publication may not be digitized, photocopied, or otherwise reproduced, posted, or transmitted, without the permission of Harvard Business School.

G A R Y P I S A N O

P A M E L A A D A M S

VF Brands: Global Supply Chain Strategy

It was August 2009. Chris Fraser, President, Supply Chain International for VF Brands, was driving to his office just outside of Milan near Lake Como. On this sunny morning, the sparkling lake was a picture of tranquility, a striking contrast to the turbulence of the global apparel industry. In the shorter term, the economic crisis of 2008-09 was taking its toll on the entire business from the largest marketing companies to the smallest sub-contractors. But beyond the crisis, Fraser also foresaw long-term structural changes in the apparel business that could call for profound changes in the way VF, the world’s largest publicly owned apparel company, managed its supply chain. Fraser noted “For the past few decades, supply chain strategy in apparel was focused on chasing low cost labor from one country to the next. Today, apparel is produced just about everywhere on Earth, and we have basically run out of new “low cost” places to source production – until, of course, penguins learn to sew. We have to start finding cost saving by how we manage our supply chain.”

For some time, Fraser had been advocating that VF shift its supply chain strategy. VF currently procured apparel both from its own plants and from a large network of suppliers. Like its competitors, VF’s outsourcing strategy emphasized flexibility. Most suppliers in the garment industry received short-term contracts (typically a few months) to produce a specific garment in specific volumes. This strategy allowed garment marketers like VF to shift production among suppliers in different locations in order to optimize costs and to respond to changes in exchange rates, tariffs, and other cost factors. Many believed that this approach also provided strong incentives for suppliers to reduce costs in order to compete for future contracts. Fraser admitted that while this approach had worked well for many years, it had its drawbacks. The lack of coordination and trust between suppliers and apparel companies led to higher inventory and long lead times. In addition, Fraser felt that a company like VF, with its strong internal manufacturing capabilities, had expertise that it could share with suppliers in order to improve processes and reduce costs. He noted, “For products coming from our own manufacturing plants, we can move things through the supply chain in days instead of weeks. That allows us to respond very quickly to the market. That’s the value of having our own plants. But from a capital point of view, it may not make sense for VF to continue to build its own plants. What I would like to see is that we create supplier relationships that work as closely with us as our internal plants do.”

Fraser called this approach the “Third Way” sourcing strategy because it represented an alternative to both in-house manufacturing and traditional sourcing. Fraser had first pitched the “Third Way” strategy five years ago, but encountered skepticism from some groups within the

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

127

610-022 VF Brands: Global Supply Chain Strategy

organization. To date, VF had experimented with a limited number of “Third Way” supplier relationships. Fraser now felt VF had the data and the experience to reflect on this experiment and to decide, once and for all, whether the “Third Way” should be implemented more extensively.

VF Brands and the Apparel Industry

In 2008, VF Corporation had total revenues of just over $7.6 billion. The company’s roots could be traced back to 1899 as the Reading Glove and Mitten Company based in Pennsylvania. In 1914, the company expanded into lingerie and in 1917 changed its name to Vanity Fair. In 1969, Vanity Fair entered the jeans business through the acquisition of the Lee Company. By 1983, jeans accounted for 75% of the company’s $1 billion in sales. In 1984, the company embarked on a series of acquisitions aimed at expanding the jeans product line and diversifying into new areas. It acquired Blue Bell (owner of the Wrangler, Rustler, and Girbaud jeans brands), Jantzen (sportswear and backpacks) and RedKap (occupational apparel and uniforms). Through much of its history, Vanity Fair pursued a vertically integrated manufacturing strategy in jeans, with many of its factories located in the United States.

In 2004, the company made a significant shift in strategy. Its new “Growth Plan” called for the transformation of Vanity Fair (now VF) from a company focused on basic apparel (like jeans) into a global lifestyle apparel company with strong brands. Fraser noted, “We used to be a company that sold what we could manufacture. With the Growth Plan, we decided to focus on marketing, and source products from the outside.” While continuing to invest in growing “heritage brands” like Wrangler and Lee, the company acquired new brands with global appeal through a series of acquisitions. These included The North Face®, Vans®, Nautica®, Reef®, Kipling®, Eastpak®, Majestic®, Napapijri®, Eagle Creek®, John Varvatos®, 7 For All Mankind® and lucy®. In 2000, heritage brands accounted for 90% of sales. By 2008, heritage brands represented only 56% of sales revenue with lifestyle brands making up the remaining 44%. The company’s goal was to have 40% of sales from heritage brands and 60% from lifestyle brands.

There were two other critical elements of the company’s strategic growth plan. One was to expand sales outside the US, particularly in rapidly developing countries like Russian, India, and China. In 2001, international sales constituted only 19% of revenues. By 2008, this figure had grown to 30%. Further growth in international sales was targeted. The final element of the company’s strategic growth plan was to expand its direct to consumer business. VF, like other apparel companies, had historically sold its products through independent stores. However, following a recent trend in the apparel business, VF was creating its own, single brand stores as well as expanding its web-based retailing. By 2009, the company had over 700 single brand stores (mostly The North Face, Napapirji, Lucy, John Varvatos, and 7 For All Mankind). These stores acted as showcases for the brand, but also drove significant revenue growth. VF planned on opening 75 to 100 single brand stores annually, with a target of 1300 stores globally by 2012. In keeping with its international expansion strategy, the company was emphasizing the Asian markets for the location of new stores. The company’s distribution strategy balanced different types of channels: specialty stores (16%), domestic and international retailers (16%), department stores (2%), chains (7%), upscale department stores (3%), mass retailers (15%), royalty income (13%), and international wholesale (28%).

VF organized its businesses into five major “coalitions.” Each coalition was responsible for the product lines, marketing, and sales of a set of related brands globally. Two of these coalitions were heritage businesses: jeanswear and imagewear. Jeanswear, consisting of the Lee, Wrangler, and Rustler brands, was the largest coalition with $2.8 billion in revenue (2008). On its own, the VF F

or u

se o

nl y

in th

e co

ur se

S up

pl y

C ha

in A

pp lic

at io

n an

d P

ol ic

y at

M ic

hi ga

n S

ta te

U ni

ve rs

ity ta

ug ht

b y

D r.

C ar

ol P

R A

H IN

S K

I f ro

m S

ep te

m be

r 07

, 2 01

7 to

N ov

em be

r 18

, 2 01

U se

o ut

si de

th es

e pa

ra m

et er

s is

a c

op yr

ig ht

v io

la tio

128

VF Brands: Global Supply Chain Strategy 610-022

Jeanswear coalition sold more pairs of jeans than any other company in the world. The Imagewear coalition provided uniforms for commercial or industrial use (e.g. Federal Express employee uniforms) as well as for sports franchises (the NBA, the NFL, and collegiate sports). The Imagewear coalition had sales of $1 billion. Three additional coalitions were associated with the lifestyle brands. The Outdoor and Action Sports coalition contained the Eastpak, Vans, Reef, The North Face, Napapirji, and Eagle Creek brands and had $2.8 billion in 2008 revenues. The Sportswear coalition housed the Nautica, Kipling, and John Varvatos brands, and had 2008 revenues of $625 million. Contemporary Brands, the newest coalition (established in 2007), included 7 For All Mankind and lucy and had 2008 sales of approximately $350 million. Exhibit 3 provides an overview of the financial performance of each coalition.

VF took great pains to preserve the organizational cultures and unique brand identities of the companies it acquired. A critical part of this strategy was to allow acquired companies to keep their design groups intact and in their original locations. As a result, design at VF was highly decentralized. For instance, Vans’ (clothing and shoes for skating, surfing, and snowboarding) design was done at the organization’s southern California home. Napapirji design continued to be based near Milan. The North Face had design studios in the US (San Francisco Bay Area) and Italy (Treviso). Chris Fraser noted, “We try not to monkey around with brand heritage. We keep the design and culture the way it was.”

The Apparel Industry

The apparel industry encompassed the design, manufacture, and marketing of clothing, accessories, and personal luxury goods. Global sales in 2008 (at retail prices) of apparel were approximately $1.3 trillion.1 The sector included an extremely broad range of products and price points, from basic garments likes socks and underwear to casual clothing and sportswear to super- premium “haute couture” suits and dresses. Most garment companies were associated with a product segment that represented their traditional “base.” VF, for instance, was known for many years as a “jeans” company. Van Heusen was a “shirts” company. But over time, the larger companies (like VF, Liz Claiborne, Phillips-Van Heusen, and Sara Lee) had branched out into a growing number of product lines. In addition, companies which had traditionally competed in shoes and footwear, like Nike and Adidas, had entered the apparel business and had become major players in certain segments (such as sportswear). Similarly, many apparel companies (including VF) had acquired shoe lines. Given the sheer size and breadth of the industry, competition was highly fragmented and even the largest players typically had only single digit market shares. Even within specific segments of the market, rivalry was generally intense, with dozens of brands competing directly. Jeanswear was a good example. While VF was the largest seller of jeans globally, its $2.8 billion in annual jeans sales constituted only about 5% of the total market (approximately $50 billion in sales). In such competitive environments, substantial and continuous investments in brand building were essential to maintaining margins. Most major apparel companies (VF, Christian Dior, Nike, Adidas, Ralph Lauren, Liz Claiborne) invested 7-12% of their revenues in advertising.

Another major trend in the apparel industry was the growing power of large mass retailing chains in the distribution of clothing. Walmart, for instance, had become the largest jeans retailer in the US. Volume gave merchants like Walmart significant bargaining leverage with respect to raw materials and logistics costs along with supplier contracts. In addition, large retailers were developing and marketing their own “private label” store brands. Walmart had introduced its own brand of jeans (Faded Glory) that it sold for $9/pair (compared to its $16 price for Wranglers). It had also recently

1 Datamonitor, Global Apparel and Textiles: Industry Profile, March 2009.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

129

610-022 VF Brands: Global Supply Chain Strategy

launched a line of jeans made exclusively for them by Jones Apparel, called L.e.i., specifically targeted at teenage girls.

Most apparel companies typically concentrated on design and marketing, and generally performed little or none of their own production, much of which had been transferred to low cost countries around the world. In fact, while 49% of retail apparel sold in the United States was made domestically in 1992, by 1999 this figure had dropped to only 12%. Extensive outsourcing had become the norm for a number of reasons. The production of garments was generally a labor intensive process offering few scale advantages. Barriers to entry into production were relatively low. As a result, there were hundreds of thousands of small contract garment manufacturers scattered around the globe. Moreover, the skills required to produce garments (cutting and stitching fabric) were relatively generic. This enabled garment companies to source production of their design on highly competitive terms. In addition, garment production was subject to complex and ever changing tariffs and quotas. Bilateral trade deals generally dictated from which countries garments were imported. Further complicating matters was the fact that there were separately negotiated duties and quotas for fabrics and textiles. These could dramatically change the economics of production in one country versus another. Historically, garment companies “chased quota”; that is, they sought out low-cost producers from countries that had not yet hit their quotas for importing into a certain country. While tariffs and quotas on textiles and garments were being reduced under the auspices of the 2005 World Trade Organization accord, it was still far from an open market. The largest apparel companies that sold products in all major global markets thus found it advantageous and less risky to have an extremely broad geographic base of suppliers so that they could shift production in response to changes in tariffs and quotas.

Fraser noted that while tariffs and quotas had come down over the past two decades, those barriers had left the industry with a highly fragmented and often illogical supply chain. For instance, for a sweater sold in the US market, the raw wool might be sourced in Australia. That wool would then be sent to China for spinning into yarn. The yarn might then be sent back to Australia for dying and knitting into fabric panels. Those panels would then be sent to China again where the sweater would be assembled. From there, it would be shipped to the US. Fraser noted, “Ideally, we are trying to line up all the vertical steps in a region or a country. Thailand, for instance, is an ideal place to assemble backpacks. We can shorten lead times significantly if we can procure fabric, components, and other raw materials there as well. If we do a similar thing with other products in China where our stores are located, and we can move the goods right out of the packing area directly into the stores, all the better!”

As supply chains globalized, the challenges of finding suppliers, managing sourcing relationships, and coordinating product flows had been steadily increasing. In the 1990s, many American and European based apparel companies found that they lacked both the skills and relationships for effective sourcing in Asia. To fill this need, some Asian manufacturers shifted their business models to provide fully-integrated supply chain services to apparel companies. Hong Kong based Li & Fung was a good example of this breed of supply chain service company. Founded in 1906 as a trading company, Li & Fung now managed the supply chains and sourcing for many of the world’s largest brands (in apparel, footwear, and other consumer products). In essence, Li & Fung acted as an intermediary between the brand companies and a network of sub-contractors scattered around the globe. In recent years, Li & Fung, like other supply chain intermediaries, had begun to integrate forward into their own brands and retail chains.

Over the past ten years, the upstream segment of the supply chain--garment production--had been undergoing dramatic changes. Since 2001 (the first year garment quotas were eliminated), fabric producers in the CAFTA region (which included Central America, the US, and Dominican Republic) Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

130

VF Brands: Global Supply Chain Strategy 610-022

had been steadily losing market share to imports from China (who had increased their share of the US market from 7% in 2001 to 45% in 2009). In the past decade alone, the US had lost 50% of its fabric production capacity. Fraser reflected “It is getting harder to maintain apparel production in places where we cannot get cheap and speedy supply of fabric.”

In 2008-09, the world economy was gripped with the worst recession since the Great Depression of the 1930s. Falling gross domestic products and plummeting consumer demand in the US, Europe, and many developing economies did not spare the global apparel industry. Total industry revenues fell by 10%. VF was weathering the storm relatively well compared to its competitors. While it had seen sales decline by 9% (and significantly less when exchange rates were taken into account) in the first half of 2009 (compared to first half of 2008) and earnings decline by 30% over the same period, the company’s financial position was strong. It had a cash position, relatively low debt, an A- bond rating, and ample untapped lines of credit. The bigger concern among some senior managers was the long-term effect of the crisis on the supply base. Many garment contractors were small shops, operating on razor thin margins, and with virtually no financial cushion. As volumes fell, many were forced to shut down. In China alone, it was reported that over 60,000 small production shops had closed their doors in 2008-2009. Sudden closures of suppliers could be very disruptive. For instance, one of VF’s suppliers of jeans (supplying over 15 million pairs of jeans per year to VF) had given VF less than 3 months notice that it was closing down its plants in Nicaragua and moving production to Vietnam, a location much less favorable to VF due to quota, tariffs, and logistics. VF had to scramble to find an alternative supplier.

VF Operations Strategy

VF used a mixture of both internal manufacturing and outsourcing, a relatively unique operations strategy in the apparel industry. Beginning in the 1980s, many major apparel companies began to sell off their internal manufacturing operations and source their products from specialized suppliers. Many of VF’s major competitors, like Liz Claiborne, Ralph Lauren, Levi Strauss, and Sarah Lee, no longer had any internal manufacturing and relied completely on outsourcing. At the other extreme, companies like Benetton and Zara were completely vertically integrated from garment production through retail, and did limited outsourcing.

As noted earlier, VF had historically been an apparel manufacturer. At one point, it owned approximately 100 factories. With the acquisition of The North Face, in the late 1990s, this began to change. The North Face, like many of the organizations VF would subsequently acquire, had no internal manufacturing. VF’s existing manufacturing infrastructure was not well suited to these lifestyle brands for two reasons. First, VF’s plants were largely focused on jeans and denim products, while many of the lifestyle brand products were not. Second, VF plants were located in Mexico and the Caribbean in order to optimize the logistic costs and tariffs to serve the US market, VF’s traditional focus. With the strategy of expanding into lifestyle brands and international markets, the company needed to expand its outsourcing in Asia. This was a significant shift in the company’s philosophy. A painful part of this new strategy included the closing of many of VF’s internal manufacturing plants. By 2009, VF produced about 30% of its products in-house (in its 40 remaining plants), and sourced the rest from independent suppliers. Of course, there was significant variance across product lines in sourcing. For instance, VF produced about 60% of its jeans in-house. Imagewear, which was largely targeted at the US market, and which required very quick response times, also sourced the vast majority of its products internally. On the other hand, VF used outsourcing for 100% of its lifestyle apparel, footwear, and backpacks.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

131

610-022 VF Brands: Global Supply Chain Strategy

Still, VF was proud of the internal manufacturing capabilities the company had accumulated over 125 years, and believed those capabilities provided it a significant competitive advantage. A recent benchmarking study by a consulting firm indicated that VF’s internal manufacturing plants were among the very best in the world in terms of quality, efficiency, and reliability. The time needed to produce a garment in VF-owned factories was much shorter than the industry average. VF’s factories also had defect rates well below industry averages. On production lead times, VF-owned factories required 10 days from “cut to ship” compared to 30 to 50 days for external suppliers. VF also believed it had built up technical and engineering capabilities for apparel manufacturing that few companies could match. For instance, its Mexican and Nicaraguan plants employed about fifty engineers focused on improving processes. The company had developed novel techniques and even proprietary equipment for manufacturing jeans. Mike Green, Managing Director of VF Asia, a member of Fraser’s international sourcing team, had spent 26 years at VF in engineering, plant management, and sourcing roles. He commented “There is no doubt that VF plants set the standard in the industry.”

On the sourcing side, building up a reliable and high-quality supplier network required an enormous investment in time. Prospective suppliers needed to be visited and their manufacturing capabilities carefully assessed. In addition, VF had a strict policy of only doing business with suppliers who followed internationally established standards for worker safety and protection. It also took time to establish good working relationships with suppliers, and only experience could really tell which ones were reliable. By 2009, VF had relationships with more than 1600 contractors and 30 distribution centers around the world. The top 20 suppliers accounted for about 45% of the outsourced volume procured by VF on an annual basis. To manage this, the company hired Chris Fraser in 2000. He came from another sourcing position in a large apparel company located in Asia. When he arrived, sourcing represented only a small portion of total sales. Between 2000 and 2009, with the acquisition of many new lifestyle brands, the company’s sourcing volume in Asia alone increased 15 fold to reach a total value of $1.8 billion. As VF gained experience with sourcing, the management team also began to understand that its supply chain network provided a significant platform for growth. Chris Fraser provided an example: “When we acquired Napapirji, they had a very strong brand, but they would have had to invest years and millions of dollars to grow to $300 million in sales. By being part of VF, they now had access to our supply chain network. We could just plug them in to our system.”

One of biggest challenges of running such a large apparel supply chain was the sheer complexity of the product line. VF, for instance, currently had over 600,000 SKUs, where an SKU was defined by only style and color (and not article size). Jeanswear alone had approximately 100,000 SKUs. Moreover, while some “classic” product lines change little from year to year, the lifestyle brand product typically had very short product life cycles, and required nearly constant replenishment of new designs. On average, about half of VF’s SKUs were essentially new product designs every year.

A second complexity was the widely differing needs and priorities of the brand coalitions. For instance, in more fashion oriented products where VF competed with companies like Liz Claiborne and Tommy Hilfiger, product design was considered “king.” Product designers in those lines focused almost solely on creating an exciting menu of products that would hit the fashion “sweet spot” in the coming season. In these product lines, cost was not such a critical issue. Products sold in VF’s brands stores typically fit into this category. In other products lines, the name of the game was low cost and rapid replenishment. For instance, large retailers in the US demanded that VF be able to replenish store inventories within 8 days in order to minimize inventory costs. For product lines competing with Zara, a chain well known for being able replenish inventories continuously throughout a season, the supply chain had to be extremely responsive. There were also significant differences in product requirements across regions, even for seemingly very similar products. Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

132

VF Brands: Global Supply Chain Strategy 610-022

Consider jeans. In the American market, jeans were by and large a non-fashion clothing item (Fraser, an American, pointed out, “Americans wear jeans as an ‘anti-fashion’ statement”). A good quality pair of Wranglers could be bought for $16-$30 depending on the retailer. But in Europe, jeans were worn as a fashion item. They had different cuts, design, and fit, and were often made of different denim than jeans sold in the US market. They also sold at smaller retailers and at much higher prices than in the US (e.g. a pair of Wranglers could sell for $60-$80/pair).

Floyd Perkins, a 26 year VF veteran, was Corporate Vice President for Supply Chain which included all aspects of procurement, manufacturing, sourcing, and distribution. Procurement, manufacturing, and distribution were managed from the US, while international sourcing was based in Europe and Asia (Hong Kong), with local offices in China, Pakistan, India, and Bangladesh. Perkins’ organization supported all of the coalitions. While the coalitions were responsible for the designs, volume decisions, pricing, and margin goals for their product lines, the supply chain organization planned capacity (internally or externally), managed inventory, and coordinated all the processes required to go from fabric to a finished product on the store shelf.

The Apparel Supply Chain: From Design to Store Shelf

It is August 2009 and the The North Face store on Newbury Street in Boston is bustling. The summer heat grips Boston, but the shoppers are browsing the Autumn collection, contemplating the cool days ahead. Few are aware that the fleece vests, rain jackets, t-shirts, pants, shirts, and backpacks they see on the store shelves are the culmination of a process that began a slightly more than a year ago. It all began in June 2008 when VF designers began to sketch their preliminary ideas for the Fall 2009 collection, and made commitments to broad design themes and colors. Over the next few months, designers would create literally hundreds of preliminary designs encompassing the entire product line. It was a highly iterative process. Design concepts were dropped and others were added, hues and shades were tweaked, new patterns created (and discarded), pockets moved a centimeter higher (or lower), and trims adjusted. Ultimately, the design and marketing teams would have to make judgments about what would appeal to customers more than a year hence.

At some point, even the most detailed drawing was not enough to judge a design, and physical prototypes had to be sewn. These would be used internally by management to evaluate the design as well as to show to prospective customers (e.g. large retailers). For the sake of speed and secrecy, VF made prototypes in their own or partnered development centers. This process typically took 4 weeks. More iterations might follow, and in some cases, additional prototypes might be fabricated. While the designs were taking shape, the marketing groups were forecasting prices, volumes, and margins for each item. These forecasts would have a critical impact on the supply chain strategy. By September of 2008, the design and marketing management would have to make a final decision on the entire 2009 Autumn collection. Design decisions were reviewed at both the brand and coalition levels.

Once design commitments had been made and volume orders placed, Floyd Perkins’ operations took over (prior to this point, operations was often consulted informally on product manufacturing or supply chain issues that might affect design choices, volumes, or prices). The first step for operations was to develop a sourcing strategy for each product (internal vs. external supplier, location of suppliers, etc.). Sourcing had to be done at multiple levels of the supply chain, not just for final garment assembly. Sources for raw materials, fabrics, and accessories had to be identified and lined up. Several criteria went into the choice of supplier. Location choices were influenced by both economic factors (cost of production and transportation) and trade quota/tariff considerations. Suppliers were then chosen according to their managerial as well as technical skills and expertise in

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

133

610-022 VF Brands: Global Supply Chain Strategy

specific garments. Sourcing offices located throughout the world were responsible for identifying and managing the procurement process with suppliers in the local region.

From September through December, the sourcing organization was focused on identifying suppliers, obtaining price quotes, and producing samples. Sample production was critical for the supplier to really know what was involved in production of a particular garment (e.g. how much labor time would be required) and thus critical for cost estimation. Sample production was used by VF to evaluate a supplier’s capability to produce a product to its specifications. It was not unusual at this stage to discuss technical challenges in the production process that required design modifications (for instance, the angle of a seam specified by the designer might be hard to sew in high volumes). While typically minor, each potential design change required a discussion between the supplier, VF sourcing offices, and VF designers. Designers typically did not like to make changes in the design for the sake of manufacturing. Perkins explained “We don’t just make what we can make and that’s it. The product really rules. If the designer says the jean needs three legs, then we’ll put three legs on it. But if we have a better way of doing it than what’s been handed to us, and it’s OK with them, then we try to improve the design.” Samples were used by VF sales and marketing personnel to show to retailers and wholesalers in the sales process during October and November.

By January 2009, contracts would be signed for all products in the Fall 2009 collection, and suppliers (or VF) would begin placing their orders for raw materials and fabrics. The long lead times made forecasting particularly critical. For some seasonal products (like backpacks where 90% of all products were sold in the “back to school” shopping season), inaccurate forecasts could be particularly costly. Depending on the fabric, the procurement lead time could be anywhere from 4 weeks to 12 weeks. Once suppliers had all the necessary raw materials and fabrics, they would begin production of garments. Contracts generally specified a target volume. Virtually all suppliers performed work for multiple garment companies. A critical factor affecting their economic viability was running factories at full capacity, and thus they typically scheduled production in batches in order to optimize utilization. VF could expect that its Fall 2009 products were being cut, sewn, and assembled in the March to June time frame. During this time period, VF could make some adjustments to its orders as it updated its demand forecasts. The products would then be shipped to regional distribution centers where they were sorted and packed for bulk shipping to target markets. Transportation (via ship) from a distribution center in Asia to a US port could add another 2 weeks of lead time. From the port, products would clear customs and be shipped to a VF distribution center in the US, where they would then be sent (via truck) to a store. Retailers could expect the Fall collection to begin arriving during the early part of July.

Historically, apparel supply chains were very inflexible. Retailers and wholesalers placed their orders 8 to 10 months prior to a particular season. Product would usually arrive at the beginning of the selling season. After this point, the retailer had limited ability to adjust stocks in response to actual customer demand. If a particular shoe line was proving to be exceptionally “hot,” the retailer probably could not restock mid-season. They would simply run out of product and lose prospective sales. Conversely, if a highly anticipated new line of jackets was experiencing disappointing sales, the retailer would likely be stuck with excess inventory and be forced to discount this item substantially. Thus, retailers suffered the costs of both excess inventory and stock-outs.

Over the past decade, both retailers and garment manufacturers had been trying different strategies to build in more flexibility. Zara, for instance, used a combination of vertically integrated manufacturing, small lot production, and information technology to adjust inventories on a weekly basis. Zara stores stocking out of a hot item could expect a replenishment shipment in two weeks. Companies that relied solely on outsourced production had less scope to change production schedules, as they often had to commit to production volumes well in advance of the selling season. Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

134

VF Brands: Global Supply Chain Strategy 610-022

However, even in these cases, companies were trying to make the supply chain more responsive through better information exchange and through inventory management. VF had found that it could be highly responsive for jeans products manufactured in its own Mexican plants and destined for the US market. Chris Fraser explained, “For jeans sourced out of our factories in Mexico, we might cut fabric on a Monday. By Thursday, those jeans have been sewn and washed. Friday, they are packed and probably put on a truck heading to the US. By the next Monday, those jeans are in our US distribution center. By that Thursday, those jeans could be on a store shelf.”

“Third Way” Supply Chain Strategy

In 2004, Chris Fraser felt that VF had created a highly efficient, globally diversified supply chain. There were a few basic types of sourcing relationships VF could establish with a supplier. One type was known as “cut and make” (CM) contracts. Under this approach, VF would strike separate contracts for suppliers at each stage of the production process (fabric, components, cutting, sewing, washing, finishing, etc.). VF owned the inventory and suppliers were paid for the value added of their particular step. VF was also responsible for coordinating the flow of product from one supplier to the next. The advantage of this approach is that it allowed VF to maintain very tight control over costs at each stage. CM contracts were used mostly for heritage lines out of Central American and Caribbean suppliers and were managed in conjunction with internal manufacturing operations. A second approach was known as “package sourcing.” Under package sourcing, a single supplier took responsibility for the entire process from raw materials to finished goods and shipping into the market. They were paid on a piece basis and are responsible for paying sub-contractors, raw material suppliers, and logistic costs. In this case, VF did not have ownership of the materials along the process. Full package sourcing was used mostly for the lifestyle brands throughout Asia, Europe, and Northern Africa.

VF’s coalitions were very pleased with the quality and reliability of the service they received from sourcing. In addition, VF had driven down costs by continually expanding the supplier base to ever lower cost locations. An aggressive focus on low cost sourcing had enabled VF to hit its overall corporate margin targets of 10-15%. It was at that point, however, that Fraser and Green began to think about next steps. Fraser believed that VF needed to shift its focus from finding even lower cost suppliers to doing a better job managing the supplier base it had. Fraser continued, “The supply chain could be made more efficient.”

Some of the inefficiency was due to lack of coordination and, more precisely, lack of trust between apparel companies and their suppliers. Historically, apparel companies and apparel suppliers showed little loyalty to one another. Contracts were short-term (typically one season). In their aggressive pursuit of low costs, apparel companies drove hard bargains on pricing and freely shifted production from one supplier to another. There were no guarantees in either direction. Every year, suppliers had to bid to get new business from a company and never guaranteed production capacity beyond a very short time horizon (covered by the contract). They also took on products for as many companies as possible (often competitors) in order to diversify their risks. In addition, suppliers never shared production information (capacity, inventory, costs) with apparel companies, for fear such information would be used against them in the bidding process.

One manifestation of this lack of trust was excess inventory. Floyd Perkins provided an example: “When you own the factory there is no question of trust or commitment. Therefore we are able to operate on incredibly lean inventory levels. We have 3 days of work in process in our factories and maybe half a day’s work ahead of the line. By contrast, our external suppliers often require that we

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

135

610-022 VF Brands: Global Supply Chain Strategy

provide them 30 days of raw materials. Why? Because they don’t trust us to come through. The inventory gives them security.”

The process was also very time consuming. Because there was no pre-set menu of prices for different design features (e.g. a pocket or a stone wash), the price of each garment had to be negotiated from scratch with each supplier. For instance, if VF wanted to add a pocket to a pair of pants, a supplier having a good year might ask for a 27 cent price increase while another who needed to fill capacity might only ask for only 21 cents. Fraser commented, “There is a lot of transaction time and gamesmanship in the negotiations with suppliers.”

Another problem with current sourcing was the lack of process improvement. Because garment contractors operated on razor thin margins, they invested little, if anything, in process technology. Engineering skills were not only hard to come by, they were not highly valued or sought after by contractors. In fact, because they were generally born in very low wage environments, productivity improvements were never a high priority. If problems arose, the answer was generally to add more workers or overtime.

Fraser and Green believed that VF’s technical capabilities represented a “trump card” that no other competitor had. Floyd Perkins agreed:

You go to places today to source where labor in incredibly cheap. That is not going to be the differentiator any more. We need to look for other benefits like speed to market, material utilization, lower inventories, less work in process, and lower cost to quality. Competitive advantage no longer comes from reducing the amount of needle time that goes into the garment, but from managing the whole supply chain. And that is where we might be able to beat another brand that does not know how to manage the supply chain.

Fraser explained that the biggest barrier to growing internal manufacturing was VF’s desire to minimize investments in fixed plant and equipment. He noted “One could argue that we should be buying more factories and leveraging our expertise in production that way. But even if it only costs $10 to $15 million to build a good factory, such a strategy would not be consistent with VF’s corporate capital deployment strategy. That money is better invested in our brands and our retail operations.”

Was there a way for VF to leverage its internal technical expertise and gain a greater degree of control over the supply chain without actually owning suppliers? Fraser believed that if VF could put its “hard technical skills” into the supply chain, it would have an advantage over apparel companies that lacked internal manufacturing competences. Floyd Perkins commented:

We have an abundance of engineering skills dedicated to our own factories. If we could take our engineers in Mexico who are shaving pennies off the cost of garments that have been engineered for 100 years now, and transfer them to places where 60% of our products are made now, that could have a big impact. We had one recent example that just blew me away. We sent this one young engineer from Mexico to one of our Asian suppliers. He came back and said they don’t load the containers as efficiently as we do in Mexico. So we had him develop some guidelines for them for container utilization. The result was a $2 million annual savings.

Fraser thought he had a strategy for making the supply chain much more efficient. He called it “Third Way” sourcing. “Third Way” sourcing was designed to be a half-way point between full integration and traditional outsourcing. The idea was to create a true partnership between VF and the supplier. While the details might differ across specific suppliers, Fraser saw the key elements of a “Third Way” sourcing relationship as follows: Fo

r us

e on

ly in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

136

VF Brands: Global Supply Chain Strategy 610-022

• VF would strike an agreement with a supplier for a specific product line (e.g. backpacks) and commit to a volume forecast over a number of years (instead of one season). The supplier would not produce the same category of product (e.g. backpacks) for competitors, and would agree not to do so in the future.

• The supplier would set up production lines dedicated to VF’s products, investing in the building, machinery, equipment, labor supervision, logistics services, and administrative infrastructure to manage the operations.

• VF and the supplier would develop production schedules jointly to meet each partner’s needs. Information on order forecasts and production capacity were to be shared between the partners.

• VF and the supplier would work together on process improvements. VF would make available (without charge) its engineering resources to improve production processes. A portion of the savings realized by these improvements would be passed through to VF.

• The supplier would own the factory and the equipment and be responsible for managing the work force. VF would make certain investments in specialized equipment and capital when necessary.

• VF would utilize its purchasing capacity to help the suppliers procure fabric and other raw materials at discounted prices. VF would agree to buy back any unused fabric or raw material from the supplier.

• The supplier would be paid on a cost plus basis with a margin to meet its ROA requirements.

Fraser recalled his first presentation of the idea back in 2005 to a group of senior managers. Concerns about the idea came from all directions. Some within the marketing organization were worried about the loss of flexibility the new approach might entail. Fraser commented “Sourcing was viewed as a big ‘candy shop’ where you could get just about anything you wanted. The marketing folks were concerned that if we went away from our sourcing strategy, we could lose flexibility to serve them.” The manufacturing organization, on the other hand, was frustrated by the continued closing of internal plants despite their strong performance. At another level, they were also not happy about handing over their engineering resources to be managed by the sourcing team.

But the real concern of internal manufacturing revolved around the sharing of its proprietary expertise with outside suppliers. They raised serious concerns about the leakage of VF’s hard earned process expertise which could eventually be used to produce for competitors. Fraser and Green, however, did not share these concerns. They had no plans to transfer VF’s proprietary equipment to “Third Way” partners. Rather, their focus was on using the skill sets of internal manufacturing to improve supplier performance in terms of cost, quality, and speed. As Green explained, “There is a philosophy of how we train our engineers and the training programs from a junior engineer all the way through a division engineer. It’s about what we teach them and how we teach them, and copying that is not as easy as just flipping on a light switch or reading some manual.”

Despite the objections of these internal groups, Fraser and Green persisted. By 2009, they had formed five “Third Way” partnerships. One was for the production of backpacks in Thailand. A second was for jeans production in Bangladesh. A third was for jeans production in Morocco. A fourth was for jeans production in China for the Chinese market. The fifth was for outerwear production in China. One of the interesting things Fraser and Green learned was that it was much easier to convince new suppliers than existing suppliers to sign up for a “Third Way” partnership.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

137

610-022 VF Brands: Global Supply Chain Strategy

When they broached the concept to some of VF’s best existing suppliers, they got a cold reception. Green recalled, “It’s hard to convince suppliers this is a good idea for them. The experienced ones can be pretty set in their ways about how they run their plants and their operations. They were really not that interested in us coming in there and changing their processes. We also had a hard time to get partners to share information about costs and processes within their factories.”

Another challenge of implementing “Third Way” partnerships was staffing. In fact, Green noted that he found it difficult to move forward on the plans more quickly. He either had to find experienced engineers from VF’s own factories that would be willing to move across the world, or hire locals and then send them through a rigorous training program within VF’s manufacturing division. Tom Green provided an example from the Bangladesh partnership: “We hired some top graduates out of a Bangladesh university. We then put them into a mentoring program with two of our engineers with 10 and 20 years experience that we brought over to Hong Kong to work with these guys. The next phase for us will be to create a routine for this training process. Tom and Floyd are constantly saying move faster, but you can only do so much so fast.”

The experience to date, while limited, gave VF management a glimpse into the potential cost benefits of “Third Way” sourcing. Exhibit 4 provides an overview of the lead times, inventory, and costs of producing a standard “five pocket jean” in different locations and under different sourcing arrangements (internal manufacturing, traditional packaged source, and “Third Way” partnerships).

Not all partnerships ended as planned, as illustrated by VF’s experience with its Moroccan partner. Despite VF’s volume guarantees and technical help to improve operations, the Moroccan jeans plant was not able to dig itself out of debt (largely due to drastic declines in its business with other companies). By early 2009, its financial position had deteriorated to the point that the owner considered halting operations. To protect its supply base, and in the face of a highly favorable price due to exchange rates, VF decided to buy-out the partner, and the Moroccan plant is now a wholly owned VF operation. VF transferred a manager from one of its Mexican plants to Morocco to run the operations.

Floyd Perkins felt the real benefits of the “Third Way” strategy had not even been seen yet, because they lay in the design process. He commented “If you think about speed to market, which is always one of the challenges of the supply chain, about two-thirds of the time is spent in the product development process. Only one-third is the time it takes to go from the order to the delivery to the store shelf. I think we also need to focus on those first stages to see how we can shorten lead times.”

Fraser felt that the “Third Way” strategy had reached a critical cross-road. VF’s ambitious international expansion goals, particularly for Asia, meant that they would need to bring on significant new capacity over the next several years. They could do that by expanding “Third Way” sourcing, expanding internal manufacturing, or by simply doing more traditional sourcing. The decision, Fraser felt, would have a profound impact on VF’s competitive capabilities for years to come.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

138

VF Brands: Global Supply Chain Strategy 610-022

Exhibit 1 VF Consolidated Statements of Income

In thousands, except per share amounts 2008 2007 2006 Net Sales $ 7,561,621 $ 7,140,811 $ 6,138,087 Royalty Income 80,979 78,548 77,707 Total Revenue 7,642,600 7,219,359 6,215,794 Costs and Operating Expenses

Cost of goods sold 4,283,680 4,080,022 3,515,624 Marketing, administrative and general expenses 2,419,925 2,173,896 1,874,026

6,703,605 6,253,918 5,389,650 Operating Income 938,995 965,441 826,144 Other Income (Expense)

Interest Income 6,115 9,310 5,994 Interest Expense (94,050) (72,122) (57,259)

Miscellaneous, net (3,103) 2,941 2,359 (91,038) (59,871) (48,906) Income from Continuing Operations Before Income Taxes 847,957 905,570 777,238 Income Taxes 245,209 292,324 242,187 Income from Continuing Operations 602,748 613,246 535,051 Discontinued Operations ---- (21,625) (1,535) Net Income 602,748 591,621 533,516 Earnings Per Common Share- Basic

Income from continuing operations $ 5.52 $ 5.55 $ 4.83 Discontinued Operations ---- (0.20) (0.01)

Net Income 5.52 5.36 4.82 Earnings Per Common Share- Diluted

Income from continuing operations $ 5.42 $ 5.41 $ 4.73 Discontinued operations ---- (0.19) (0.01)

Net Income 5.42 5.22 4.72 Cash Dividends Per Common Share $ 2.33 $ 2.23 $ 1.94

Source: Company documents.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

139

610-022 VF Brands: Global Supply Chain Strategy

Exhibit 2 Consolidated Balance Sheets

In thousands 2008 2007 Assets Current Assets

Cash and equivalents $ 381,844 $ 321,863 Accounts receivable, net 851,282 970,951

Inventories 1,151,895 1,138,752 Deferred income taxes 96,339 104,489

Other current assets 171,650 109,074 Total current assets 2,653,010 2,645,129 Property, Plant and Equipment 1,557,634 1,529,015

Less accumulated depreciation 914,907 877,157 642,727 651,858 Intangible Assets 1,366,222 1,435,269 Goodwill 1,313,798 1,278,163 Other Assets 458,111 436,266 $ 6,433,868 $ 6,446,685 Liabilities and Stockholders’ Equity Current Liabilities

Short-term borrowings $ 53,580 $ 131,545 Current portion of long-term debt 3,322 3,803

Accounts payable 435,381 509,879 Accrued liabilities 519,899 489,160

Total current liabilities 1,012,182 1,134,387 Long Term Debt 1,141,546 1,144,810 Other Liabilities 724,248 590,659 Commitments and Contingencies ---- ---- Common Stockholders’ Equity

Common Stock 109,848 109,798 Additional paid-in capital 1,749,464 1,619,320

Accumulated other comprehensive income (loss) (276,294) 61,495 Retained earnings 1,972,874 1,786,216

Total common stockholders’ equity 3,555,892 3,576,829 $ 6,433,868 $ 6,446,685

Source: Company documents.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

140

VF Brands: Global Supply Chain Strategy 610-022

Exhibit 3 2008 Sales Revenue by Coalition

Coalition Name Brands Sales Revenue ( in Millions)

Jeanswear Wrangler, Lee, Rustler 2,751 Outdoor and Action Sports JanSport, Eastpak, The

The North Face, Vans, Reef, Napapijri, Eagle Creek

2,751

Imagewear Red Kap, Bulwark, The Force, NFL, CSA, Chase Authentics, Majestic, Harley- Davidson

994

Sportswear Nautica, John Varvatos, Kipling

611

Contemporary Brands 7 For All Mankind, lucy 383 Other 153 Total Sales Revenue 7,643

Source: Company documents.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

141

610-022 VF Brands: Global Supply Chain Strategy

Exhibit 4 Comparative Results for Alternative Sourcing Solutions: VF Jeanswear Five Pocket Jeans

Alternative Sourcing Solutions

VF - Owned & Operated

Packaged Sourced

Packaged Sourced Third Way Third Way

Fabric Source Mexico China India India India India Cut, Sew, Finish Mexico China Bangladesh Morocco Bangladesh Morocco Days Leadtime Leadtime 17 60 71 75 48 52 Finished goods inventory 24 83 99 104 68 85 Total days forward 41 143 170 179 116 137 Cost per unit ($) Cost fabric/yard (incl. freight) 2.64 2.23 2.25 2.37 2.25 2.37 Fabric/unit 3.03 2.72 2.76 2.90 2.76 2.90 Cut, Sew, Finish (incl. Transport) 2.62 1.93 1.90 2.36 1.90 2.36 Total COGS ($) 5.65 4.65 4.66 5.26 4.66 5.26 Local manufacturing and margin 0.38 0.38 0.39 0.50 0.39 0.50 FOB price ($) 6.03 5.03 5.05 5.76 5.05 5.76 VF overhead 0.14 0.32 0.32 0.34 0.22 0.24 Finished product freight 0.16 0.35 0.35 0.28 0.35 0.28 Duties 0.73 0.73 0.73 Landed cost ($) 6.33 6.43 6.45 6.38 6.35 6.28 Inventory carrying costs 0.19 0.52 0.52 0.49 0.43 0..42 Markdown provision 0.03 0.19 0.19 0.18 0..16 0..14 Net cost ($) 6.55 7.14 7.16 7.05 6.94 6.84 Charge for capital per unit ($) 0.93 0.12 0.12 Total Costs ($) 7.48 7.14 7.16 7.05 7.06 6.96

Source: Company documents.

Note: Data and locations have been disguised for purposes of confidentiality. All $ figures calculated at exchange rates prevailing on August 1, 2009. All duties and freight rates as of August 1, 2009.

F or

u se

o nl

y in

th e

co ur

se S

up pl

y C

ha in

A pp

lic at

io n

an d

P ol

ic y

at M

ic hi

ga n

S ta

te U

ni ve

rs ity

ta ug

ht b

y D

r. C

ar ol

P R

A H

IN S

K I f

ro m

S ep

te m

be r

07 , 2

01 7

to N

ov em

be r

18 , 2

01 7.

se o

ut si

de th

es e

pa ra

m et

er s

is a

c op

yr ig

ht v

io la

tio n.

142

NOTICE REGARDING COPYRIGHT This custom course package contains intellectual property that is protected by copyright law. It is illegal to copy the material within this package without the written consent of the holder(s) of the copyright. This material has been copied under licence from Access Copyright or the copyright owner. Resale or further copying of anything in this package is strictly prohibited.

www.iveycases.com

Supply Chain Application and Policy SCM470-001 and 002
Process Fundamentals
Earth Buddy
McLeod Motors Ltd
Quinte MRI
Ranger Creek Brewing and Distilling
International Decorative Glass
The Home Depot, Inc.
Industrie Pininfarina: The New Customer Decision
VF Brands: Global Supply Chain Strategy