Falling Into the Data Mine: Avoiding the Pitfalls

• Activity Number: 248
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Sam Koslowsky*+
• Companies: Harte Hanks, Inc.
• Keywords: Database marketing ; Predictive analytics ; Marketing analytics ; Data mining ; Non technical ; Practicing analytics

Abstract

Data mining plays an important role in assuring successful marketing programs. However, errors can creep into the data mining process from expected and unexpected places -- problems that all analysts can confront and be aware to prevent. More often than not, these errors are non-technical in nature. This session discusses what these situations and scenarios are, and the problems that may be encountered and avoided. From defining the wrong objective to choosing an inappropriate tool, this session, presented by a practitioner, will highlight how even the experienced data miner has erred. This paper will help practitioers avoid the most basic problems surrounding data mining efforts, and help marketers achieve better insight by considering simple, non-technical issues for mastering data analysis.