SEGMENTATION USING CUSTOMER LIFETIME VALUE HYBRID K-MEANS AND ANALYTIC HIERARCHY PROCESS

• Main Author: Rahmadhan, Radit
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/70051
• Institution: Institut Teknologi Bandung

Developing predictive analytics based on understanding customers' electricity consumption patterns is essential to effectively manage the increasing electricity demand. This study presents a hybrid approach to customer segmentation by combining K-Means clustering, the concept of customer lifetime value, and an analytic hierarchy process to better understand customers' electricity consumption behaviour. We use K-Means clustering to identify initial market segments. Next, we evaluate and validate the customer segmentation results using the customer lifetime value concept and the analytical hierarchy process. Segment 1 has 282 business customers with a total capacity of 938,837 kWh, peak load usage of 27,827 kWh, and non-peak load of 115,194. In segment 2, there are 508 customers with a total capacity of 938,837 kWh, a peak load usage of 27,827 kWh, and a non-peak load of 115,194. In segment 2, there are 508,615 business customers with a total power of 4,260 kWh, then a peak load of 35 kWh and a non-peak load of 544. In segment 3, there are 37 business customers with a total power of 2,226,351 kWh, then a peak load of 123,297 kWh and a non-peak load of 390,803. Strategies to be taken based on the segmentation of these three customers will be integrated with CRM. For the least profitable segment, we propose an ongoing partnership program to encourage increased electricity consumption during non-peak periods and retail account marketing. For profitable and medium profitable customers, we propose a premium business to business approach that can accommodate their increased energy consumption without excessive electricity usage during peak periods. This approach will be supported by dedicated executive accounts for these customers.