| dc.description.abstract | Customer segmentation is allocation of customers into segments according to characteristics similarity to optimize the implementation of customer relationship management. The characteristics of retail customer could be identified based on their transaction record. Principal component analysis is used to summarize the characteristics of PT United Tractors Tbk. spare part retail customer from transaction record in 2012. Variables in the transaction analysis and segmentation are transaction frequency ( ), count of transaction months ( ), count of product ( ), count of product type ( ), and total transaction amount ( ). The weight of those variables are determined by principal component analysis. The weight of each variable obtained is 0.41 for , 0.78 for , 0.61 for , 0.28 for , dan 0.72 for . Transaction score of each customer is calculated by combining the standardized transaction variables. Ideal segmetation is conducted by k-means algorithm. The ideal number of segment is 4. The mean of transaction score for segmen 1 to 4 are 9.09, 3.34, -1.39, and 0.27. Customers are devided into 4 segments by percentage 2.79%, 12.09%, 54.88%, and 30.23% for segment 1 to 4. | id |
