<![CDATA[This study develops an AI-driven application for analyzing sales of local fashion products and mapping customer heterogeneity to support marketing decision-making for micro, small, and medium enterprises (MSMEs). A mixed-methods approach is employed, combining a structured literature review, surveys/interviews, and Focus Group Discussions (FGDs) to validate findings and system usability. Quantitatively, first-quarter 2025 transaction data (100 respondents) are analyzed using K-Means on three standardized features age, aggregated number of items purchased, and aggregated spending. Cluster evaluation with the silhouette score for k=2-5 indicates the best separation at k=5, yielding a stable and interpretable segmentation. The resulting profiles reveal at least one high-value segment (larger baskets and higher spending) suitable for tiered loyalty programs and premium bundling; a mid-value segment responsive to targeted cross-sell/upsell offers; and a low-intensity segment that benefits from staged onboarding interventions to improve retention. These insights are integrated into a prototype analytics application that presents a segmentation dashboard and key performance indicators, providing actionable support for MSMEs’ marketing, catalog curation, and inventory allocation.]]>
Copyrights © 2026
