Open AccessSEGMENTASI PELANGGAN BISNIS DENGAN MULTI KRITERIA MENGGUNAKAN K-MEANS
Author(s) -
Yanuar Wicaksono
Publication year - 2019
Publication title -
indonesian journal of business intelligence
Language(s) - English
Resource type - Journals
eISSN - 2621-3923
pISSN - 2621-3915
DOI - 10.21927/ijubi.v1i2.872
Subject(s) - database transaction , customer intelligence , customer retention , computer science , transaction data , customer to customer , market segmentation , customer advocacy , asset (computer security) , business , customer knowledge , customer equity , loyalty business model , voice of the customer , marketing , knowledge management , database , service (business) , service quality , computer security
Customer knowledge is an important asset, in gathering, and managing from sharing customer knowledge into valuable capital for the company. This causes the company to continue to innovate in producing products and serving according to customer needs. To find out the needs of each customer, the company needs to make customer segmentation. Customer segmentation is defined as the division into different groups with similar characteristics to develop marketing strategies that are tailored to customer characteristics. The easiest, simplest, well-known and commonly used model of customer characteristics is the model of the recency, frequency, monetary (RFM) criteria. The RFM model still has weaknesses in low customer segmentation capacity and does not provide information on the continuity of customer transactions in understanding customer loyalty. The research method used is the Knowledge Discovery in Database (KDD) method. The data is transformed into another format that suits the needs of analysis and then the customer is segmented using clustering data mining techniques with the K-Means algorithm. From the experiments, the RFM model guesses loyal customers when reviews, frequency and monetary are high. In reality, the recency only provides information on the customer making the last transaction and the high number of transaction frequencies can be done without the customer's stability in making transactions each period. Implementing multi-criteria in customer segmentation can be better than just RFM criteria. So it will not be wrong to treat customers according to the groups that have been formed.