This article was downloaded by: [27.6.227.

119] On: 09 November 2020, At: 23:49

Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
INFORMS is located in Maryland, USA

INFORMS Transactions on Education

Publication details, including instructions for authors and subscription information:
http://pubsonline.informs.org

Case—KEY Electronics—Sourcing and Warehouse Analysis

Tim Kraft, Yenho T. Chung, Feryal Erhun,

To cite this article:

Tim Kraft, Yenho T. Chung, Feryal Erhun, (2012) Case—KEY Electronics—Sourcing and Warehouse Analysis. INFORMS
Transactions on Education 12(2):92-99. https://doi.org/10.1287/ited.1110.0065cs

Full terms and conditions of use: https://pubsonline.informs.org/Publications/Librarians-Portal/PubsOnLine-Terms-and-

Conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact permissions@informs.org.

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.

Copyright © 2012, INFORMS

Please scroll down for article—it is on subsequent pages

With 12,500 members from nearly 90 countries, INFORMS is the largest international association of operations research (O.R.)
and analytics professionals and students. INFORMS provides unique networking and learning opportunities for individual
professionals, and organizations of all types and sizes, to better understand and use O.R. and analytics tools and methods to
transform strategic visions and achieve better outcomes.
For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org
 Vol. 12, No. 2, January 2012, pp. 92–99
ISSN 1532-0545 (online)

I N F O R M S http://dx.doi.org/10.1287/ited.1110.0065cs
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

© 2012 INFORMS
Transactions on Education

Case
KEY Electronics—Sourcing and Warehouse Analysis
Tim Kraft, Yenho T. Chung, Feryal Erhun
Department of Management Science and Engineering, College of Engineering, Stanford University, Stanford, California 94305
{tkraft@stanford.edu, yhthomas@stanford.edu, ferhun@stanford.edu}

T his case is based on a real-life scenario. Names, dates, and company information have been disguised to protect
the identity of the company.
Key words: product sourcing; inventory modeling; global operations
History: Received: February 2010; accepted: December 2010.

Charles Thomas, Director of Procurement at KEY for products. Before leasing a new warehouse or hir-
Electronics, rummaged through his desk drawer for ing a 3PL, he wanted to correct KM’s sourcing prob-
his Spanish dictionary. He read the last paragraph of lems. One month earlier, he had assigned Thomas to
the e-mail again. lead an analysis of KM’s sourcing and warehousing
requirements. Thomas’ team had collected all of the
Perdón, pero esa es toda la información de demanda
necessary data, but still needed time to complete the
que tenemos. Me habría encantado tener más informa-
ción para ustedes, pero como ya saben, la mayor parte analysis.
de nuestra información histórica está corrupta—ya sea In the end, Rickard agreed to let Thomas’ team
por problemas con la base de datos o porque nuestros finish the analysis. However, he wanted to see their
empleados hacen ajustes incorrectos en la bodega.1 recommendation by the end of next week. Thomas
knew it was going to be a rough couple of weeks.
If only he had paid more attention in high school He decided to quit looking for the dictionary 0 0 0 and
Spanish class. No matter, he got the main idea of the instead look for a bottle of aspirin.
e-mail. There was no more data; the Mexico office had
sent everything it had. He and his team would just
have to make do.
1. Corporate Background
Three days earlier, Andy Jacobs, KEY’s Chief Sup- 1.1. KEY Electronics
ply Chain Officer (CSO), had stormed into Thomas’ Based in Oklahoma City, Oklahoma, KEY Electron-
cubicle after one of his weekly flare-ups with KEY’s ics is a consumer electronics retailer with more than
CEO, Reilly Rickard. As usual, the source of Jacobs’ 1,500 stores throughout the United States. It sells an
and Rickard’s disagreement was the KEY de Méx- assortment of personal electronics, including batter-
ico (KM) store expansion project. This time, their ies, cell phones, TVs, and laptop computers. KEY’s
argument concerned the new Mexico City warehouse. corporate strategy is based on superior customer ser-
Rickard was tired of all the planning; he was ready vice and its unique close-to-customer, small-box retail
to go with the third-party logistics provider (3PL). locations. In 2008, KEY generated $1.4 billion2 in total
The 3PL’s quoted prices seemed reasonable and he revenue with a net income percentage of 5%.
knew that for KM to stay on schedule, it would need Although a well-known brand throughout the
a new warehouse ready by the end of 2009. Jacobs, United States, in recent years, KEY’s revenue growth
however, was not ready to commit to the 3PL. He stagnated. Management believes that there are two
had always suspected KM did not maximize its profit primary reasons for KEY’s lack of growth:
margins by identifying the optimal sourcing location (1) With the advent of large-box electronics retail-
ers, many consumers no longer value the proximity
1
of KEY’s locations. In addition, KEY cannot compete
“Sorry, that is all the demand data we have. I wish I had more for
on selection with large-box retailers.
you, but as you know most of the historical data is corrupt—either
from the database or employees making incorrect adjustments at
2
the warehouse.” All dollar amounts in the case are in U.S. currency.

92
 Kraft, Chung, and Erhun: Case: KEY Electronics—Sourcing and Warehouse Analysis
INFORMS Transactions on Education 12(2), pp. 92–99, © 2012 INFORMS 93

(2) KEY’s brand both hurts and helps the company. KM’s corporate headquarters are in Mexico City.
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

Many consumers associate KEY with either fix-it-type In typical KEY fashion, KM’s 30 stores were set up on
electronics (such as batteries and hard-to-find elec- a whim by the previous KEY executive team in 1998.
tronic parts) or novelty-type items (such as toys or No real market or operational analyses were per-
gadget gifts). KEY has struggled to position itself as a formed; instead management simply rationalized that
destination for popular electronics such as iPods and there were a lot of potential new customers in Mex-
flat-screen TVs. ico, especially in the Mexico City area, and that KEY
With growth stagnant in the United States, CEO should have a presence there. Fortunately for KEY,
Rickard’s impatience with the Mexico store expansion the Mexico market and KEY are a good fit. Unlike
project is justified. Mexico is one of the few markets the U.S. market where consumers’ preferences have
KEY serves where growth seems possible. shifted toward large-box retailers, the Mexico mar-
Organizationally, KEY is well-known for being ket still appreciates the more personal experience of
loyal to its employees. It is not uncommon to find KEY’s close-to-customer stores. In addition, KM stores
KEY store managers with 25+ years of experience do not have the same brand problem that KEY stores
within the company. This is also true at the cor- have. In Mexico, KM is seen as a high-end electron-
porate headquarters where many employees started ics store with a wide assortment of popular, although
with KEY as store associates. Rickard got his start pricey, electronics.
with KEY while in college, working part time in a Mexico’s store managers are very entrepreneurial.
store to pay his tuition. CSO Jacobs’ first job out Many within KEY consider the KM stores on par with
of college was as a floor manager in the Hager- and sometimes better run than the U.S. stores. Unfor-
stown, Maryland distribution center (DC). There are tunately for KEY, the previous executive team did
exceptions, however. KEY hired Chief Financial Offi- not provide KM with the resources or infrastructure
cer Anthony Chung from Wall Street. The board to properly grow its business. In 2004, for the first
pushed heavily for Chung’s hire in an effort to time in its history, KM achieved a profit. This profit
boost KEY’s slumping stock price. Since his arrival in was based on hard work, manual operations, and an
November 2007, Chung has helped to reduce oper- extensive use of Excel spreadsheets.
ational costs; however, KEY’s stock price continues Rickard recognized KM’s potential as a growth
to lag. area for KEY and, upon his appointment as CEO in
2007, took the initial steps to grow the KM business.
1.2. KEY de México Although the initiative has executive backing, KEY’s
KEY’s retail presence in Mexico is limited, with only Mexico growth strategy is still a work in progress.
30 stores and sales last year (2008) of $15 million With many employees with 20+ years experience, the
(net profit $1 million). KEY would like to expand words “change” and “flexibility” are not phrases typ-
operations in Mexico to 70 stores and $35 million in ically associated with KEY. Instead, a popular answer
sales (net profit $2.2 million) by year-end 2012 (see throughout the company is “that’s how we’ve always
Figure 1). done it.” For KEY, expanding operations into Mexico

Figure 1 2012 Projected Store Count and Sales in Mexico

North
Northwest Stores: 4
Sales: $1.945 M
Stores: 3
Sales: $1.48 M

Northeast
Stores: 12
Sales: $6.05 M
Central
Stores: 19
Sales: $9.6 M

Southeast
Stores: 6
Sales: $2.85 M
Mexico City
Stores: 22
Sales: $11.1 M
Southwest
Stores: 4
Sales: $1.975 M
 Kraft, Chung, and Erhun: Case: KEY Electronics—Sourcing and Warehouse Analysis
94 INFORMS Transactions on Education 12(2), pp. 92–99, © 2012 INFORMS

represents a significant change. So far, with consider- States is 60+ days for Asia-sourced stock keeping
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

able internal pushback, Rickard has accomplished the units (SKUs).

following with regard to KM:
(1) A strong and agile management team is in 2.2. KEY de México
place at the Mexico City corporate offices. The team KM’s single warehouse is located in Mexico City and
understands both the electronics and Mexican mar- services all Mexico stores. Of the 30 current stores,
kets. However, the KM team still has not filled its Vice 12 are located in the Mexico City area (52% of sales).
President of Operations position; this is a major con- On average, KM has around 300 SKUs with inventory
cern for Rickard and a major headache for Jacobs. on hand in their Mexico City warehouse.
(2) KM is currently implementing an off-the-shelf Historically, KM procures approximately half of its
IT system that will allow it to better manage the product from KEY’s Oklahoma City DC (47.6%) of
inventory in its Mexico City warehouse. The previous cost of goods sold (COGS) and half from local Mexico
system was unreliable, misused, and built on corrupt distributors (52.4%). This is somewhat misleading,
databases. The new system will integrate with KM’s though, because of the higher cost of goods for local
point-of-sale system, which is used in every store. distributor’s product (local margins are, on average,
Although the two systems will provide KM with ser- 30% versus 51% for product sourced from KEY). Mar-
viceable IT capabilities for managing 70 stores, KEY’s gins are typically much lower for local product as
management team is still worried that KM’s poor local purchases are from Mexican electronics distribu-
operational processes may corrupt the new system. tors who, with their middleman services, take a por-
(3) Chung’s finance team approved either the lease tion of the profits. KM prefers to source KEY U.S.
of a new warehouse or the hiring of a 3PL in Mexico products but because of high duties and U.S. stores
City to service all KM stores. Currently, KM leases having first priority for products, KM’s procurement
a 15,000 square-foot facility in Mexico City that CSO team is often forced to source local product. However,
Jacobs once described as “something out of the indus- with the establishment of a direct from Asia channel,
trial revolution.” The current warehouse has uneven KM has a third source for product that may reduce
cracked floors with product scattered throughout the supply uncertainty.
facility. KM has identified a 40,000 square-foot ware- High duties complicate KEY’s import of goods from
house to potentially lease located just north of Mexico the United States and Asia into Mexico. Currently,
City. KEY has also approached a 3PL in Mexico about there exists intense competition between Mexico and
the possibility of KM outsourcing its warehousing East Asia, especially China, to attract manufacturing
operations. and commercial interest from countries such as the
United States. To detract foreign imports from Asia,
2. Operations Background Mexico charges a significant duty on all products
2.1. KEY Electronics entering the country from Asia. This includes goods
KEY has three DCs in the United States, which ser- that flow through the United States (i.e., Oklahoma
vice its 1,500 stores. The West Coast DC is located in City replenishments). Although the duties on product
Richmond, California and acts as the port of entry sent from Asia to the United States are negligible, the
for all of KEY’s imports. All goods originating from duties for these same products in Mexico (either com-
Asia come through the Richmond DC and then are ing direct from Asia or through the United States) are
either stored there or forwarded to one of the other extremely high, with an average duty rate of 25%.
DCs. Goods procured from U.S. manufacturers and One regulation issue KM must consider when
distributors are shipped directly to each individual importing product from either the United States or
U.S. DC. The Oklahoma City DC services the Central Asia is proper packaging labels. Mexican importa-
U.S. and Mexico. The Hagerstown, Maryland DC ser- tion law requires that all products imported for sale
vices the Eastern U.S. KEY’s DCs are by no means into Mexico comply with Mexican Mandatory Stan-
state of the art, but they get the job done. Over the dards and be properly labeled with what are called
years, KEY’s operations have been customized to ser- “Normas Oficiales Mexicanas” (NOM) labels. These
vice KEY’s unique network of small-box stores, which labels ensure that the product being imported into
require frequent, smaller replenishments to stores. the country has been verified by the Mexican govern-
KEY sources approximately 89% of its product ment for safety standards and that the selling com-
from Asia. For most goods, it uses a periodic review pany has obtained and paid for the proper license to
inventory policy with orders occurring on weekly sell the product in the Mexican marketplace. A NOM
increments.3 The delivery lead time to the United label license costs approximately $1,000 per SKU per
year, and failure to comply with proper NOM label-
3
Example: SKU 101 is ordered weekly and SKU 202 is ordered ing can result in a fine of up to $10,000 per each
every three weeks. noncompliant unit. Typically corporations such as
 Kraft, Chung, and Erhun: Case: KEY Electronics—Sourcing and Warehouse Analysis
INFORMS Transactions on Education 12(2), pp. 92–99, © 2012 INFORMS 95

Figure 2 KM Channel Overview take anywhere from 11 to 17 days. KM’s poor pro-
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

cesses and lack of information visibility are big rea-

sons for the order inconsistency. As shown in Table 1,
Oklahoma the order sizes generated by KM can vary widely.
City, OK Another challenge that hampers lead times between
Oklahoma City and Mexico City is the lack of trust
between KM’s procurement team and U.S. fulfillment
operations. Some within KEY believe KM inflates
KEY U.S. orders to acquire “hot” items that it knows will sell in
Product the Mexico market but to which U.S. stores have first
priority. KEY fulfillment personnel have full autho-
rization to adjust KM’s orders as they see fit. Unfor-
tunately, these adjustments are often based on “gut
Asia Local feeling” instead of on actual communication with
Direct Vendors KM’s procurement team.
Mexico A final problem that increases lead times is the
City NOM packaging requirement. As part of the KM
order fulfillment process, the Oklahoma City DC
manually labels each unit sent to Mexico with the
proper NOM warning label, so that it may cross the
Walmex (Wal-mart Mexico) embed NOM certification border into Mexico. So, for example, each package
requirements into their bilingual packaging. Unfortu- of batteries on a pallet must be labeled. The aver-
nately, KEY just started the expensive and compli- age shipment processing time is 3–5 days.5 Currently,
cated process of transitioning to bilingual packaging. shipping KM orders out of the Oklahoma City DC is a
The marketing department estimates the transition to hassle, but manageable because they can be treated as
bilingual packaging will finish, at the earliest, by some very large store orders. However, once KM grows to
time in 2013. A summary of the three sourcing chan- 70 stores, if KEY still does not have bilingual packag-
nels is provided below (see Figure 2): ing, then the current 15–20 employee manual labeling
(1) U.S. Imports from Oklahoma City DC. KEY’s operation may not be able to process items quickly
Oklahoma City DC is KM’s largest supplier of goods enough. KEY’s DCs and IT systems are designed for
in terms of physical quantity, supplying more than frequent small-box store replenishments, not large,
500,000 units at a freight-on-board cost of $3.6 mil- biweekly, DC to warehouse replenishments. Jacobs
lion in 2008. KM prefers goods from KEY’s Oklahoma does not want to see KM orders start to disrupt oper-
City DC over local products because of their higher ations at the Oklahoma City DC.
(2) Local Mexico Product. Because of a lack of histor-
margins (51%).
ical data and the wide range and variety of suppliers
While the shipment of goods from the United States
used, KEY’s management team has limited visibility
represents 47.6% of KM’s total unit costs (freight-on-
into KM’s local buy process. While shorter lead times
board origin4 ), for the United States, the demand cre-
for local product (3 days versus 14 days for U.S. ship-
ated by Mexico is minimal—in 2008, the amount of
ments) significantly reduce the amount of inventory
goods shipped to Mexico represented only 0.35% of
on hand, the high cost of local items typically negates
the total U.S. demand (COGS). To put this amount the local lead time advantage. Because local orders
into perspective, compared to U.S. demand regions, are easier to process, KM uses a continuous review
Mexico ranked twentieth, right behind the Orlando, inventory policy to manage goods bought from local
Florida demand region. distributors.
KM uses a periodic review inventory policy to KEY’s management team prefers a reduction in
manage goods sent from Oklahoma City. KM and local purchases to eliminate product risk. Many
Oklahoma City have an agreement where KM orders within KEY’s corporate offices are skeptical about the
goods every two weeks. To standardize the process, ethics of some of the local Mexico distributors. It is
a two-week ideal order fulfillment process has been not unheard of in the electronics industry for local
established (see Figure 3). suppliers and distributors to sell either black market
While the process is improving, issues still remain. or counterfeit items to retailers.
There is variability in the process and so orders can (3) Direct from Asia (through Manzanillo, Mexico).
Recently, KM established a new direct from Asia
4
KM pays the shipping costs and takes responsibility of the ship-
5
ments when they leave the Oklahoma City DC. This processing time is already included in the 14-day lead time.
 Kraft, Chung, and Erhun: Case: KEY Electronics—Sourcing and Warehouse Analysis
96 INFORMS Transactions on Education 12(2), pp. 92–99, © 2012 INFORMS

Figure 3 KM-Oklahoma City Order Process

Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

2 days 1.5 days 1.5 days

2 days

Sales information KEY Merch/

KM buyers OKC completes
received by KM Replen validates OKC processes
determine order validation and
buyers for prior and forwards orders
quantities inputs order
weeks sales order

Broker
Order is
completes Order arrives at Order labeled
delivered
approval border and shipped
to KM DC
process

1 day 1 day 1 day 4 days

to Mexico sourcing channel. So far, only two test In total, the lead time for Asia shipments is 65 days.
orders have been processed. Because of the lengthy While a large amount of coordination and process
order process and requirements set by KEY’s Asia development is still required to refine this process,
consolidators, KM will use a periodic review inven- KEY hopes that many of the fast-moving KEY prod-
tory system to manage Asia orders with orders being ucts that require fairly large stocks of inventory can
submitted every four weeks. The fulfillment process be sourced directly from Asia to Mexico, thereby:
works as follows: (1) Reducing KM’s dependency on U.S. operations,
(a) Orders are sent directly from KM’s procure- (2) Decreasing potential inventory availability con-
ment team to KEY sourcing offices in Asia. flicts between U.S. and KM operations, and
(b) A shipment is prepared in Asia for Mex- (3) Reducing processing costs, as Asia labeling
ico, including the attachment of NOM labels. KM costs are less than U.S. labeling costs.
provides the labels and processing requirements. On Managing Asia to Mexico shipments might prove
average, this process takes two days. difficult for KM’s procurement team, given the lack
(c) A KM only ocean shipment is sent from Asia of accurate information. Because of long lead times
to Manzanillo, Mexico, located on the West Coast of for overseas products, KM must submit forecasts
Mexico. On average, shipments take 60 days. three months in advance for direct from Asia product.
(d) A broker verifies the shipment and processes In the past, KM struggled to provide accurate infor-
the importation documentation. On average, this pro- mation when submitting forecasts to U.S. operations.
cess takes one day. Yet, this typically was not a problem because of KM’s
(e) Product is shipped by train from Manzanillo
small demands in comparison to U.S. demands. KM
to Mexico City where it is transferred by truck to
cannot afford these same mistakes if it begins to order
the KM warehouse. On average, this process takes
directly from Asia.
two days.
Because of long lead times from Asia, there is
Table 1 KM-Oklahoma City Order Summary also an end-of-life (EOL) inventory risk. Some popu-
lar electronics, such as cell phones and laptops, can
2008 Weight (lbs) Units Value ($)
go EOL within a matter of months if a new ver-
Average 51650 191831 1121380 sion is released. Jacobs already got in trouble with
Max 231542 881579 2551217 Chung over a major U.S. inventory write-off because
Min 342 11529 151182 of obsolescence at the start of 2008. He does not want
Standard deviation 41312 151198 261840
to go through the same process with Mexico.
 Kraft, Chung, and Erhun: Case: KEY Electronics—Sourcing and Warehouse Analysis
INFORMS Transactions on Education 12(2), pp. 92–99, © 2012 INFORMS 97

3. KM Warehousing Requirements Costs would be based on the annual number of pal-

Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

KM’s existing Mexico City warehouse is in poor shape lets received and store replenishments made, as well
and is too small to handle the expansion to 70 stores. as on weekly inventory snapshots of the number of
For the new warehouse, KEY and KM management pallets on hand. The 3PL would store KM’s product
are considering two options: (1) leasing a new facil- in a single facility located in Mexico City. Advanced
ity or (2) hiring a 3PL to run KM’s warehousing and shipment notifications would be sent from KM to the
distribution. 3PL one day prior to any receipts from the United
As a first option, KM’s real estate team has iden- States or Asia. Local shipment receipts would require
tified a modern, 40,000-square-foot warehouse for a four-hour advanced notification. Store replenish-
potential lease, located on the outskirts of Mexico ments would be based on a weekly schedule agreed
City. The facility’s storage capacity is approximately upon by KM and the 3PL.
1,850 pallets, which Jacobs estimates should be
enough to handle KM’s store growth until 2012. The
rent and management salaries required to run the new
facility are given in Appendix A. The variable cost to 4. KM SKU Sourcing and
run the new facility will need to be estimated based Warehouse Analysis
on the variable cost at the existing facility (see Fig- In the summer of 2009, Jacobs approached Thomas
ure 4). For small warehouses (< 60,000 square feet) about a special assignment for KEY’s Mexico oper-
the variable cost typically increases linearly in square ation. Jacobs wanted Thomas and his team to ana-
footage. In addition, CSO Jacobs expects at least a 15% lyze KM’s sourcing of products to determine (1) if
reduction in variable cost if KEY moves to a new and KM was maximizing its profits by sourcing goods
improved facility. from the correct location and (2) what should KM’s
As a second option, Jacobs requested a pricing bid
optimal sourcing and warehouse strategy be. In the
from a reputable 3PL in Mexico City. Because of its
U.S., although not ideal, KEY’s sourcing and opera-
unique distribution model, KEY has never worked
tions groups typically planned sequentially, with the
with a 3PL before, so Jacobs is unsure how one might
operations team reacting to the decisions made by
perform in Mexico. However, based on the quote he
the sourcing team. Jacobs felt this method would not
received, the pricing seems reasonable (Appendix B).
work for KM because KM’s sourcing options were so
Figure 4 Existing Warehouse Weekly Variable Cost (Sample)
100 3,500

90 3,000
85% capacity
Capacity utilization (%)

80 2,500
Variable cost ($)

70 2,000
65% capacity

60 1,500

50 1,000
45% capacity

40 500

30 0
1 2 3 4 5 6 7 8 9 10 11 12
Week

Capacity utilization Variable cost

April –June (2007)
Week 1 2 3 4 5 6 7 8 9 10 11 12
Variable
3,200 2,400 2,350 2,400 1,600 2,400 525 1,700 500 3,400 2,350 1,660
cost ($)
Capacity
90 75 68 80 55 68 40 58 42 95 78 58
utilization (%)

Notes. Variable costs include hourly labor wages and supplies (fuel, packaging material, etc.). Because of KM’s tendency to use a significant amount of
temporary labor, the variable cost at the existing warehouse tends to jump with capacity utilization.
 Kraft, Chung, and Erhun: Case: KEY Electronics—Sourcing and Warehouse Analysis
98 INFORMS Transactions on Education 12(2), pp. 92–99, © 2012 INFORMS

diverse. Whereas KEY primarily sourced SKUs with new direct from Asia process? When would Mexico
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

60+ day lead times from Asia, each KM SKU had orders start to truly disrupt U.S. operations? Could
three sourcing options with three different lead times. KM trust local purchases? And were KEY and KM
Because of these differences in lead time, the stor- capable of managing a 3PL? Thomas and his team
age requirements for a SKU could vary significantly had their hands full.
depending on the sourcing location chosen.
Appendix A. New Warehouse Specifications
Thomas was hesitant to accept the assignment. His
procurement team spent most of their time negoti- Warehouse parameters
ating contracts and finding the lowest unit cost for
an item, not performing complex analysis. To deter- Description Value
mine KM’s warehousing requirements, Thomas and
Capacity in pallets 11850
his team would need to incorporate inventory poli- Square feet 401000
cies into their analysis. Typically, they only modeled Rent ($/square feet per year) $3.75
inventory requirements, in the final stages of a nego- Salary per:
tiation, to determine the optimal order quantity for an Manager (one in total) $301000
SKU. Even then, they approximated values based on Supervisor (four in total) $121000
deterministic demands and weeks of supply. Benefits (% applied to salary) 30%
Eventually, Thomas agreed to take on the project.
His team spent the last month collecting the neces- Appendix B. 3PL Pricing
sary SKU, demand, and cost data from KM. Although Description Value
KM’s team was extremely responsive to requests,
they lacked reliable historical data. Most of their data Receiving cost/pallet1 $6.50
was either corrupted by their IT systems or employ- Shipping cost/replenishment 2 $11.25
ees making incorrect adjustments at the warehouse. Avg. orders per week/per store 2.50
Therefore, for their analysis, Thomas’s team identi- Storage cost/pallet per week3 Incremental
fied a noncorrupt representative SKU for each of KM’s Pallets on hand 4x5 quantity discount
10 categories (Appendices C and D). Given the lack x < 11000 $4.00
of data and limited time, they will assume the identi- 11000 ≤ x < 11200 $2.75
fied optimal sourcing location for each SKU holds for 11200 ≤ x < 11400 $1.50
11400 ≤ x $0.75
the entire category. Although KM only provided base
1
unit costs (purchase costs from the supplier) for each Includes cost to receive and put-away pallets.
2
SKU, Thomas’ team identified the additional label- Includes cost to pick, prep, and ship a replenishment
(i.e., a store order).
ing, transportation, and duty costs for each SKU and 3
Weekly inventory snapshot taken on Friday.
sourcing location (Appendix E). Type 1 service levels6
were provided by KM management. (1) Setup costs for both leasing the new warehouse and
hiring the 3PL are approximately equal, and therefore do
not impact the warehousing decision.
5. Conclusion (2) Incremental Quantity Discount implies the first
Thomas knew that his team had to work fast. He 1,000 pallets cost $4.00 per week to store, the next 200 pal-
originally estimated four weeks for the analysis, not lets cost $2.75 to store, etc.
two weeks. So far, he was struggling to see how they
Appendix C. SKU Count by Category
should incorporate KM’s inventory policies into the
analysis. Could they still base the sourcing analysis SKU category1 Description Number of SKUs
on lowest unit cost and then use this to determine
the warehouse requirements? Or should they incorpo- 13 Music 15
rate inventory policy costs such as holding and setup 16 TV, DVD 10
17 Cell phones 20
costs into their sourcing analysis? As for determining
20 Antennas, 10
the warehouse costs, he agreed with Jacobs that the Shortwave radios
attractiveness of the 3PL’s pricing would depend 23 Batteries 50
heavily on the sourcing strategy chosen. 27 Fuses, adapters 53
Thomas also realized this would not be just a cost 43 Telephones 12
minimizing exercise. There were qualitative factors 61 Surge protectors 100
that his team must consider. Could KM manage the 63 Gadgets 15
65 Calculators 15
Total 300
6
A Type 1 service level equals the probability of not stocking out
during the lead time. 1
SKU category = First two digits of SKU number.
 Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

Appendix D. KM SKUs
• Base unit cost is the price the supplier charges KEY per product. It does not include additional costs such as labeling, transportation, and duty costs. Note, the additional cost to
store goods at the Oklahoma City DC is included inthe U.S. base unit cost.
• Duties should be applied to base unit costs. Duties only apply to products sourced from the United States or Asia.
• Demands are normally distributed.
Projected 2012
Local base U.S. base Asia base Service Mean Standard deviation
KM Current Weight Volume Units/ Sales unit cost unit cost unit cost Duty level demand of demand
SKU description Item source (lbs) (cubic ft) Pallet price ($) ($) ($) ($) (%) (Type 1) (month) (month)

2300105 BATERIA KY357 Batteries KEY 0.10 0.016 1,000 1.33 0.80 0.41 0.40 18 0.98 1,800 250
1609313 T.V. 2500 SANYO TV Local 8.82 2.080 8 176.94 129.17 116.78 115.01 30 0.98 115 95
2780211 CONECTOR F-59 Cable/connector KEY 0.01 0.043 350 1.18 0.66 0.38 0.37 10 0.90 675 85
1709786 TEL CEL SONY Cell-phone KEY 0.60 0.130 128 142.22 128.00 108.09 106.67 35 0.98 152 69
Z200
6300612 PURIFICADOR Air purifier KEY 8.08 1.360 8 166.67 116.67 93.33 91.67 15 0.95 125 45
DE AIRE
1309034 KARAOKE Karaoke machine Local 16.50 3.500 4 138.89 108.33 75.00 73.61 50 0.95 30 16
JXOK5
6509036 CALC.CIENT.SC Calculator Local 1.23 0.200 80 16.28 10.42 8.46 8.30 25 0.95 108 65
OL153
6109015 REGULADOR Surge protector Local 0.96 0.080 200 20.51 13.33 8.41 8.21 13 0.95 540 270
INFORMS Transactions on Education 12(2), pp. 92–99, © 2012 INFORMS

KOBLENZ
BP1000
4309113 TELEFONO Telephone Local 1.50 0.140 115 9.49 6.83 4.46 4.37 23 0.98 910 334
ALAMBRICO
GONDOL
2000234 RADIO FM Radio KEY 3.00 0.160 100 174.39 148.23 120.33 118.58 20 0.95 38 17
YB400PE
Kraft, Chung, and Erhun: Case: KEY Electronics—Sourcing and Warehouse Analysis

Appendix E. Parameters
• U.S. and Asia lead times are normally distributed.
• KEY and KM use third-party shipping services. Therefore they do not need to consider truck or container capacity when making orders.
• The holding cost interest rate is based on a weighted average cost of capital of 10.50%.
• Transportation costs for locally sourced SKUs are considered negligible.
General Lead time/Replenishment (days) Order costs (per SKU) Ocean freight

Annual holding cost interest rate 25.00% Asia 65 Asia $12.60 $/Cubic ft $2.05
Asia (std. deviation) 7 U.S. $4.86
OKC 14 Local $0.28
OKC (std. deviation) 1.3
Local 3

Mexico shipping Labeling U.S. shipping

$/lb (Border to Mexico City) $0.09 Cost per unit (Asia) $0.02 $/lb (OKC to Border) $0.05
$/lb (Manzanillo to Mexico City) $0.12 Cost per unit (OKC) $0.14
99