Skip to main navigation Skip to search Skip to main content

Exploring Consumer Loyalty Clusters in Indonesia's Local Cosmetic Brands Through eWOM Data Mining

  • Institut Teknologi Sepuluh Nopember

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The Indonesian beauty industry is growing, with local brands competing actively. This study clusters consumer loyalty using K-Means on SOCO reviews by analyzing ratings of long wear, packaging, pigmentation, texture, and value for money. Sentiment analysis integrates text reviews to create a consumer loyalty map. The result classifies three distinct loyalty clusters consisting of low, moderate, and high loyalty. High-loyalty products exhibit stronger repurchase and recommendation tendencies, primarily influenced by packaging and pigmentation, whereas low-loyalty products are constrained by texture and longevity. These insights help brands refine product development and marketing strategies to enhance customer loyalty in a competitive market.

Original languageEnglish
Title of host publication2025 15th International Conference on Information and Communication Technology and System
Subtitle of host publicationAI for the Now and Next: Delivering Solutions and Driving Vision, ICTS 2025
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798331566999
DOIs
Publication statusPublished - 2025
Event15th International Conference on Information and Communication Technology and System, ICTS 2025 - Bali, Indonesia
Duration: 12 Nov 202513 Nov 2025

Publication series

Name2025 15th International Conference on Information and Communication Technology and System: AI for the Now and Next: Delivering Solutions and Driving Vision, ICTS 2025

Conference

Conference15th International Conference on Information and Communication Technology and System, ICTS 2025
Country/TerritoryIndonesia
CityBali
Period12/11/2513/11/25

Keywords

  • clustering
  • consumer loyalty
  • data mining
  • local cosmetic brands
  • sentiment analysis

Cite this