Wal-Mart Case Example
Wal-Mart Partners with NCR
Gains competitive edge with
new customer preferences data
Wal-Mart partnered with NCR in 1997 to
dramatically increase the size and information analysis capabilities of its
data warehouse by adding new customer preferences data.As
part of the contract, NCR supplied a WorldMark (TM)
5100M massively parallel processing (MPP) server and upgraded a second
installed NCR 5100M from 32 to 96 nodes. The contract also covered NCR
professional services, including database design, data transformation and data
management, as well as maintenance.
The two WorldMark
5100M servers -- running NCR's Teredata(TM)
relational database management system (RDBMS) -- tripled the size of Wal-Mart's
existing data warehouse from 7.5 terabytes to over 24 terabytes. With the
increased capacity, Wal-Mart retained its position of having the `world's
largest commercial data warehouse,' with more than 30 applications running on
the system and handling as many as 50,000 queries in one week. "This
expansion is part of Wal-Mart's drive to deliver what its customers want: the
right item, at the right store, at the right time and at the right price,"
said Randy Mott, senior vice president and chief information officer for
Wal-Mart. "It's the ultimate form of customer service." "Our
business strategy depends on detailed data at every level," Mott
explained. "Every cost, every line item is carefully analyzed, enabling
better merchandising decisions to be made on a daily basis. It is the
foundation for maintaining Wal-Mart's competitive edge and its continuing
success in providing everyday low prices and superior customer
satisfaction."
That competitive edge and commitment to
customer satisfaction is underscored by the unique aspects in Wal-Mart's
implementation. For example, although Wal-Mart's data warehouse incorporates
information on a nationwide basis, that information can be tailored by store to
allow merchandise buyers to gain insight into local purchasing patterns.
"Our extended partnership with Wal-Mart is another significant milestone
for NCR," said Bill Eisenman, senior vice
president of NCR's Computer Systems Group. "It underscores NCR's
commitment to customer service -- as it does Wal-Mart's -- and NCR's expertise
in building very large data warehouses to support complex customer information
and decision-support systems."
Information stored in the data warehouse
-- which is collected from the retailer's 2,900 stores -- will be used to
enhance Wal-Mart's Decision Support System, particularly for the retailer's
rapidly growing chain of Supercenters and Wal-Mart
International units. Using the 65 weeks of data kept by item/by store/by day,
Wal-Mart buyers and vendors can query and analyze information to make informed
decisions on replenishment, look at customer buying trends around the world,
analyze seasonal buying trends, make mark-down decisions, and react to
merchandise volume and movement at any time.
The NCR WorldMark
servers were installed and upgraded by mid-97. One system is configured with 32
nodes, 256 Pentium Pro 200 processors and 8 terabytes of storage; the second
server will be upgraded to 96 nodes, 768 Pentium Pro 200 processors and 16
terabytes of storage. For more historical information, read the press release from January 9, 1995 announcing "Wal-Mart buys
world's largest decision-support system from AT&T".
Wal-Mart (www.walmart.com) also implemented a Strategic Decision Support System
called Retail Link. It is an example of an Inter organizational DSS. Managers
at Wal-Mart felt that by integrating their processes and information flows with
their suppliers they would bind the suppliers to their organization, resulting
in improved revenues for both parties. In 1991, Wal-Mart launched Retail Link
which allows Wal-Mart to collect sales data from various stores, consolidate it
into useful reports, and distribute it to suppliers with weekly forecasting
information. In addition to forecasting information, suppliers get electronic
order forms that help ensure there is an adequate supply of the items that
Wal-Mart needs.
This system used existing electronic data
interchange (EDI) and satellite technologies to create a competitive advantage
that other retailer’s have tried to mimic but none have replicated. The
result of Retail Link is less inventory in stores,
more inventory of the right products at the right time and place, improved
revenues for both supplier and retailer, and better partner relationships with
suppliers. In 1998, Wal-Mart and Warner-Lambert began using the Internet to
communicate interactively about sales forecasts. They have reduced the time a
product is in the supply chain by 2 1/2 weeks. That translates into millions of
dollars in reduced inventory. Check on the Collaborative Forecasting and
Replenishment Initiative (CFAR) website at http://www.benchmarking.com.
About Wal-Mart Well-known as the world's largest retailer, Wal-Mart
had annual revenues of over $104.4 billion for the year ending
Jan. 31, 1997, and has more than 2,900 stores and over 700,000 employees. Based in
More information about Wal-Mart can be
found on the World Wide Web at: http://www.wal-mart.com .
About NCR NCR Corp is a
leader in delivering commercial open computer systems for transaction processing and decision-support solutions to customers in all
industries. The company, with headquarters in
More information on NCR can be found on
the World Wide Web: http://www.ncr.com .
NCR WorldMark is
a trademark and NCR Teradata RDBMS is a registered
trademark of
NCR Corp. All other product names are
trademarks of their respective holders.
Questions for Discussion:
1. How large is a terabyte? How much
processing power is needed to support 50,000 queries a week?
2. Why does Wal-Mart have a data
warehouse? Does it provide a competitive advantage? If so, what type of an
advantage?
3. What is Retail Link? Who uses it?
4. What does it mean to say the NCR WorldMark server is configured with 32 nodes, 256 Pentium
Pro 200 processors and 8 terabytes of storage?
5. Does the NCR website help you
understand how a data warehouse is developed?