Database implementation best practices and techniques

Database System in a Retail Environment

We are in a world of information boom and efficient management of data holds the key for success in this new business environment. The world of Internet has opened new vitas for corporations to take their business into a global level. Data management is at the center of all business applications and database systems are the very core of this new business perspective. Database systems have helped us take a huge stride in our ability to collect, interpret, and even assist in our decision making process. In short we can say that database management systems have bought a paradigm shift to our mode of business and have evolved from being simple store houses of data to more complex and highly efficient business decision making tools. Let us take Wall-Mart, the giant retailer as our case study and how the company profited from the implementation of a highly sophisticated database management system. In the process we will also analyze some of the problems faced by the company and possible solutions for the same.

Summary

Wall-Mart (Data Warehouse)

Wal-Mart Inc. is one of the biggest retailers in the United States with more than $200 billion in annual sales. Spread throughout the nation with more than 2600 discount stores and around 480 SAM clubs Wal-Mart has an exhaustive presence within the United States and even in Canada. Managing such a huge network of stores is no easy task and in order to integrate the different stores and to and optimize its performance Wall-mart had turned to information technology for intelligent business management solutions. Wal-Mart was quick to recognize the role of I.T in the decision making process and with this view it implemented an exhaustive data warehouse. Wal-Mart had a pact with NCR (in 1997) to provide the data warehouse solutions. NCR provided Wal-Mart with the "NCR WorldMark 5100M," a very extensive parallel processing database server and also constructed a huge 22 terabytes data warehouse. Armed with this massive data warehouse and integration between all the stores national wide Wal-Mart was able to have an exhaustive analysis of the trends in business and the local stores were able to better understand the buying patterns of their customers. [Tom Chester]

Data Mining

Customer service was taken to new heights with access and intelligent interpretation of data from more than 2800 stores. With excellent interconnectivity each customer can have access to the details pertaining to products from any of the stores around the country. Wal-Mart also invested on data mining software to enable better analysis of data from its massive data warehouse. The data mining software is very useful in pruning through the huge pile of information in the data warehouse and in identifying the best prospects for the business. Wal-Mart turned to Neo-Vista Software Inc. To deliver its data mining needs. The Neo-Vista decision series ™ was tailored to match the requirements of the database environment of Wal-Mart. One important advantage of the implementing the data-mining tool was that the company was able to optimize the inventory management. (Better management of stocks). According to said Randy Mott, the senior vice president and the chief information officer of Wal-Mart,

Our investment in data mining is part of Wal-Mart's drive to deliver what it's customers want: the right item, at the right store, at the right time and at the right price, the NeoVista Decision Series is enabling Wal-Mart to look at individual items for individual stores to determine seasonal sales profiles. This additional level of detailed analysis is helping us make even better informed business decisions and helps us further utilize our data Warehouse."[Author not Available] With data from all its POS (Point of Sale terminals) available in the data warehouse the Neo-vista decision series gives a clear picture of the business trends and provides good demand projection.

Analysis (Major Problem)

In spite of the very good and extensive data warehouse Wal-Mart faced numerous problems with querying the data form any of its online transaction processing systems. With an astounding 2900 stores throughout the country it is very often necessary to query the OLTP database directly. This proved to be a major debacle for the company as querying the OLTP resulted in very slow response times and in many instances it lead to complete freezing of the systems. This is mainly due to the entirely different data models used in the OLTP systems compared to the Data warehouse. Normalizations performed on the OLTP databases have affected the seamless integration with the data warehouse. This problem is due to the different stores having different database models and differing levels of Normalizations. [Phillip Greenspun]

However, individual stores can customize their database by using simple procedural language tools (SQL commands) to reduce the response time to their database queries. It is also advisable for a retailer like Wal-Mart to create data marts and make them accessible for the concerned departments rather than allow access to the data warehouse directly. Using data marts (customized view of data) significantly reduces the access time and also improves the security as sensitive information can be screened from the departments that do not require access to them. With all the innumerable software tools and the wide range of hard ware that is available with Wal-Mart it is pretty clear that a thorough customization and integration is all that is required to optimize the resources.