Analysis Telkomsel Customer To Increase Adoption And Purchasing Via Virtual Assistant Application
Ismawaty, Gadang Ramantoko
Telkom University
Jl. Gegerkalong Hilir No. 47, Bandung, 40152
Abstract
Telkomsel as a digital telecommunication company has released digital-based applications Virtual Assistant (VA) which is used to facilitate customers to make transactions and information channel. However, customers who have purchased through VA are very low compared to the total active VA users, so it needs to analyze how to increase purchasing via VA. The objective of this research is to provide an insight of customer characteristics by using prediction models based on historical data using behavioral segmentation (occasion, benefits sought, user status, usage rate, and loyalty status). Data is analyzed using random forest algorithm by using two models, per segment data and all segment data behavioral segmentation. Theoretical approaches that used are digital marketing, behavioral segmentation and customer behavior. Random forest analysis with a total of 22 input variables and 130,388 records resulting all segment data model produced the highest accuracy compare to per segment data model. All segment data produced accuracy 96% with 49,948 predicted customers and predictor importance variables such as loyalty point, recharge, data payload, SMS and Voice transaction. Those variables are used as recommendation for Telkomsel in marketing strategy to determine product/bonus/gimmick that will be offered to increase purchasing via VA.
Keywords: Virtual Assistant, behavioral segmentation, analytical data
Topic: Digital Marketing