Customer Churn Prediction Model Design Using Predictive Analytics for Modern Coffee Shop
Date
2021-02-08Author
Hasyati, Besty Afrah
Djatna, Taufik
Noor, Erliza
Metadata
Show full item recordAbstract
In today's competitive world, moving toward customer-oriented markets with increased access to customer's transaction data, identifying a loyal customer, predicting customer attrition, and estimating their lifetime value makes crucial. Since knowledge of customer value provides targeted data for personalized markets, implementing a customer relationship management strategy helps a company identify, segment customers, and create long-term relationships. As a result, they can maintain loyalty and minimize attrition. The goals of this research are: (1) to model a new business process, (2) to predict the customer churn using data mining tools based on CRM, (3) to recommend a fit strategy to prevent churn and maintain the loyalty, and (4) to evaluate the result from customer churn prediction. We used the customer's past transaction data, and it had 8 attributes: customer number, name, gender, handphone number, quantity, recent visit, total spending, and frequent visit. The result was that this system had three stakeholders: customer, management staff, and management staff internal. We conducted the surveys to all the registered customers to get the satisfaction data and continued by RFM analysis, CLV, and clustering based on the customer's past transaction. The result obtained from RFM, CLV, and Clustering predicts the churn using decision tree analysis. As the final result of this research, we obtained 31 rules with 86% accuracy in the model. The marketing strategies are then designed to prevent churn and maintain loyalty.
Collections
- MT - Agriculture Technology [2272]