Data Mining Essay

3. Data Mining and Predictive Analytics as Marketing Strategy Tool

After understanding the possible outcomes and usages of Big Data Mining and Analytics, the study of the process is necessary to identify the real possibilities behind this techniques and how this can improve a business performance. To do this; we should comprehend the basics about data mining and the process that leads from pure data to insights.

At this point, is important to note that Big data itself does not represent more large data set of structured and unstructured data; nowadays bigger than ever and in continuous expansion that can be defined as the "problem of big data" (Cox M. & Ellsworth D., 1997). The ability to organize this "problem" given certain parameters and to be able to build a model or representation of a reality taking care of the existing patterns and relationships to find the true value that lies hidden in data is what can be defined as Data mining (DM) (Kadiyala, S. S., & Srivastava, A., 2011).

According to Edelstein H., (1998). There are three levels of classification in the DM process to consider: Discovery, Predictive and Forensic. Each one can be used in different stages and purposes to add value in a Marketing strategy.

Figure XX DM process schema.

In a nutshell; these processes can be defined according to Rygielski (et. al 2002) as:

Discovery: Analyzing the data in a exploratory way in search for patterns and affiliations where no apparent relationship was before.

Predictive Modeling: Utilization of the patterns discovered in the discovery step to forecast possible future conducts or behavior.

Forensic Analysis: Use of the identify relationships to look for outliers or unusual elements in the data.

Other classifications ...

... middle of paper ...

... (such as linear regression and logistic regression) or forecasting techniques (such as neural networks and survival analysis) can be used to determine a given result, such as demand forecasting or customer attraction through direct marketing and evaluate their responses.

Other alternatives to predictive modeling in businesses are churn prediction and customer retention. Where the ability to anticipate their decisions; with for example, loyalty programs, ono-to-one marketing (personalized solution) or complaints management. Will allow companies to increase their profitability in industries such as Telecommunications where retention costs are lower as compared to new customers acquisition (Ngai et al., 2009).

Other association techniques can be also used to determine cross selling opportunities, increasing customer lifetime value and profitability over time.