Unveiling Hidden Customer Segments in E-Commerce Using DBSCAN Clustering on Demographic and Behavioral Insights

Abstrak

Customer segmentation is a crucial process in e-commerce that allows businesses to tailor their marketing strategies to specific customer groups. This research applies the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm to segment customers based on their demographic and behavioral data. The dataset used includes variables such as age, annual income, total spending, and campaign engagement, which are essential for identifying meaningful patterns within the customer base. The DBSCAN algorithm was chosen due to its ability to detect clusters of arbitrary shapes and handle noise, making it ideal for complex e-commerce datasets. The analysis identified one dominant customer segment, with a small portion of the data labeled as noise, indicating that the majority of customers exhibit similar behaviors. However, the results also highlight the challenge of parameter selection for DBSCAN, as the clustering outcome was sensitive to the chosen values of ε (epsilon) and MinPts. The segmentation revealed valuable insights, such as the fact that most customers share similar characteristics in terms of spending habits and engagement, yet a few outliers exist who do not align with these patterns. These findings provide a foundation for businesses to develop targeted marketing strategies based on customer segmentation. Despite the promising results, the study acknowledges limitations in the segmentation process, particularly with the influence of outliers and the need for further tuning of the algorithm's parameters. Future research could explore hybrid clustering models that combine DBSCAN with other techniques, as well as incorporating additional behavioral features for more refined segmentation. The insights gained from this research can guide businesses in crafting personalized marketing campaigns that cater to distinct customer segments.