WHAT IS DATA MINING? 

Data Mining is the process of analyzing large data sets in order to find patterns that can help to isolate key variables to build predictive models for management decision making. In essence, data mining helps businesses to optimize their processes so that their customers receive the most relevant services and the costs of serving them are proportionate to the value of the profits earned from them, a company’s exposure to risk is proportionate to premiums earned, etc. Data Mining enables companies to segment their customer base and to tailor products and services to the needs and purchasing power of individual groups of customers.

WHO IS DATA MINING FOR?

Data Mining is for executives involved in strategic and tactical decision making as well as operating managers responsible for cost reduction. Strategic Managers use data mining for competitive intelligence, identifying market opportunities, product launch decisions and product positioning. Managers responsible for tactical decision making use similar tools for sales forecasting, direct marketing, customer acquisition, retention and extension purposes and marketing campaign analysis. Finally, operational managers can use similar data for decisions such as the choice of sub-prime borrowers or supply chain management.

THINGS TO CONSIDER WHEN IMPLEMENTING DATA MINING

Data Quality
It's worthwhile remembering the adage "Garbage in and Garbage out" when implementing data mining. Poor quality of data can jeopardize any attempt to use data analysis for decision making. This is especially true when data is purchased from external vendors.

Size of the Database
Data Mining comes in a variety shapes and forms depending on primarily the size of databases that have to be analyzed. Smaller companies cannot afford packages, such as SAS, which are expensive to license. Excel, with its advanced data functions, is adequate for small companies and Access for data storage. Eventually, databases inevitably grow in size and companies have to plan for migration to more complex data mining tools.

Nature of Application
The choice of data mining methods depends a great deal on the kind of decision making that a company wants to achieve. Sales related decision making is best undertaken with OLAP tools such as those offered by Cognos or Siebel. Companies looking to do predictive modeling are better-off with SAS or SPSS. Similarly, data processing can be done with a variety of databases ranging from Access for small databases, MS SQL for medium size operations to more complex enterprises types like Oracle and Teradata for data coming from networks like credit card operations.

Human Resources
Data mining requires a great deal of ingenuity that highly educated people can provide. Typically, the people hired for data mining have a graduate degree and bigger companies in the financial services companies hire PhDs. Data Mining requires a portfolio of skills in data management, statistical analysis and business decision making which are hard to find.