Customer habits can influence revenue from financial technology-based companies in choosing the type of payment partner. Customers are very selective in having vendors oriented towards convenience, promotions given, and benefits offered. This study describes the application of data mining for clustering, using the K-means method by classifying income in a payment aggregator. The research aims to identify patterns and similarities in revenue data to help decision-making and business analysis. The K-means algorithm is used to partition income data into groups based on their similarities. The research results show that testing uses various quantities: k = 2; DBI = 0.023, k = 3; DBI = 0.209, k = 4; DBI = 0.116 with a max run of ten. This study obtained the best results at a value of k = 4, with a clustering pattern in four payment type categories: copper, silver, gold, and platinum. The results showed that the payment aggregator categories with the highest income values were the CASH, MINIATM TRANSFER, and VA TRANSFER methods with 10.78% from revenue. From the pattern and information provided, the company needs to maintain features that support the payment aggregator with the highest revenue, while for the payment aggregator generating lower revenue, an evaluation is required to consider adding merchants in order to boost transaction frequency and amounts.