Article Preview
TopIntroduction
Nowadays, with the increasing number of companies, product diversity, and the advent of technology, competition between companies has increased. Therefore, concepts such as customer satisfaction, customer loyalty, and target group have gained value. At this point, appropriate strategies must be presented to the relevant customers in order to meet the needs of the customers. Consequently, good churn management becomes involved in customer retention. Companies can develop techniques to keep their profitable customers and increasing customer loyalty in advance.
Churn management is applied in different areas such as banking, internet service providers, cosmetics, and the health sector. In this research, the loyalty status of the churn situation is discussed for a digital broadcasting platform. The real data is obtained from one of the leading and largest Pay-TV operators in Turkey.
As long as companies have more loyal customers, they will have a higher profit. With only a five percent increase in loyal customers yields in between 25 to 95 percent increase in the net present value of customers across a wide range of industries” (KhakAbi, Gholamian, & Namvar, 2010). Data mining methods are applied to distinguish the churn status in order to apply churn management actions based on customer information. Not only to increase revenues or reduce risks but also to improve customer relationships data mining process is used. By using various kinds of data mining algorithms, meaningful patterns can be obtained to predict outcomes within large data sets. With this study, significant churn prediction factors are introduced via existing data mining algorithms. The following questions are answered in this research with the help the data obtained from the company:
- •
What attribute can be reviewed to determine customer churn status in advance?
- •
Which attribute is not necessary for customer churn status?
- •
According to the customer information, possible churn customers will be defined?
In this way, the relationships through the data is represented and the customers are grouped according to their similar characteristics. The obtained data is extracted and grouped to get more meaningful results for the data mining analysis. The companies are interested in their customers' churn rate in order to determine a specific campaign for a particular group of customers to avoid customer churn.
The rest of this paper is following with the background part. The existing literature on data mining and churn management in different sectors are investigated. The existing data mining algorithms are used for the Pay-TV company to increase the company’s net present value. In the next part of this chapter, the problem statement and methodology part is briefly discussed. The information gathered data is used to determine the appropriate data mining method to obtain the results. In the following part, computational results are represented. Then a future research and conclusion parts are described.