Optimization Of Pricing Package Based On Data Mining

With the rapid development of the mobile communications industry, the competition between mobile operators has been intensified. Customer base has been a focus of the competition in the communications market. Currently the loss of customers has been a serious common problem for each operator. The massive and frequent loss of customers has caused significant damage to the operators. In the context that the telecommunications market is gradually becoming saturated, it is more economical to retain a customer with a potential to unsubscribe than to contract a new customer. It is thus a focus for an operator to retain its customers against the possible unscriptions.To meet the intensified competition, the telecommunications operators have continuously introduced various pricing packages for customers to choose. The pricing packages have saved telecommunications expenses for subscribers and eventually enhance the loyalty thereof. However, it is highly difficult to control the change of the market. Although it is necessary to adopt policies so as to follow the market trends, the pricing packages hastily introduced may not cater to the real needs of subscribers. This constitutes a significant potential risk for the operators as well as the subscribers. A handful instances may explain this. Many of the pricing packages introduced by the operator have not been successfully accepted by the market. The number of subscribers for a specific pricing package may not sufficient for an efficient customer management afterwards. Sometimes, the pricomg package may entails an over reduction of the prices, which makes it hard to introduce further pricing packages.Based on the current practices of Wu Zhong Office of China Telecom with regard to the pricing packages, this thesis aims to design a pricing package with the lowest expenses and the most stable portfolios in an ever changing environment for customers who are of the potential to unsubscribe. This thesis also relates to the data mining while introducing and pre-processing the relevant data. Based on the comparison made between pricing packages for customers with the potential to unsubscribe, a model is established to forecast the unsubscribing customers. Pricing packages are analyzed and compared before the most favorable pricing package emerges for the customers tending to unsubscribe. Afterwards, the most favorable pricing package has been introduced to the customers tending to unsubscribe and effectively retains the relevant customer and lower down the ratio of the customers who actually choose to unsubscribe.