Primary Article for this page.

Data Mining

Kanhu Qi, Ph.D.

The value of information may now be higher than the actual value of the goods or services being transacted. A carefully planned and accurately mined loyalty program can help a business understand who its customers are, what they are doing or not doing, and whether the company is meeting its customers' expectations.

Data mining can be used to produce high-quality business intelligence—such as classification, association, and regression analysis—to determine which customers have a higher propensity to conduct business with you. Conventional wisdom used to say that banking customers with multiple accounts were the bank's best customers; however, using data mining it was found that they are actually unprofitable. The business that formalizes the way it sorts, analyzes and applies data will have the competitive advantage. In the end, the company which has the largest database—and knows how to use it effectively—wins.

Companies, looking to turn information into new products, begin by sifting through enormous quantities of customer transaction data to find the golden nuggets. A customer credit card company can have almost 1000 attributes, or data-points, per customer. An insurer may track 150 variables with 40 to 50 parameters on 6 million customers. A mortgage application can have 300 data-points per application.

Data mining can evaluate millions of records with hundreds of variables. Wal-Mart analyzes each of its 100 million transactions per week and manipulates a year's worth of cash-register transactions, right down to each shopping cart, to see how purchases of different items are related. Wal-Mart even found that customers who buy a Barbie doll have a 60% likelihood of buying one of the three types of candy bars.1

Data mining enables target advertisements, while age and purchase pattern profiling helps identify the best business practice. Sabre Inc. uses data mining to identify its most valuable customers. When these customers call, they are automatically sent to high-level specialists instead of entry-level support agents. Likewise, e-talk Corporation provides a VIP channel for its most important customers to call directly instead of going through the calling queues. Satisfying and keeping the most profitable customers is the most important job in a competitive business environment, because the most profitable customers also are the most difficult to satisfy.

Technology has changed customer relationship management from an art into a science through data mining. Due to the availability of low cost storage, fast processors and parallel processing power, loyalty programs that use data mining are more affordable than ever before. The little guys can, increasingly, do what only the big guys, like Wal-Mart, could do several years ago. Technology levels the playing field.

Prairie Trail Software has written programs that capture the transaction data for our clients and we continue to write customizable programs to help our clients mine their data to get the full benefits of their loyalty programs. Contact us for more information. ♦

Next Article>>