ABSTRACT
The conversion rate of purchasing prepaid credit on e-commerce applications is still very low. The objective of this research was to gain insights into predicting customers who have a tendency to purchase prepaid credit using the clustering and logistic regression techniques. Logistic regression was used to predict customers using 17 numeric variables. After building a development model, it was then applied to all populations and identified customers with a high tendency to purchase. To increase efficiency and effectiveness in marketing programs, marketers need to know customers’ priority. Segmentation was conducted with K-means clustering in order to determine which cluster gives the greatest predictive gain. The prediction results of high-prospect customers were divided into two clusters with the two-step cluster techniques, namely low and medium high-value customers (HVC). Based on the results, behavioral targeting can be done so that the marketing campaign is more targeted.
