Project Summary
RevoBank, an Indonesian financial institution, sought to accelerate its credit card business growth by improving transaction activity among existing customers. The project analyzes a six-month performance baseline of customer spending, card usage, and profitability to identify growth potential. This analysis serves as a strategic foundation for data-driven decision-making in a competitive banking landscape.
Goals
The primary objective was to increase the total number of credit card transactions by 10% within three months. This was to be achieved by identifying specific user personas and card segments with the highest growth potential using historical performance data. The project aimed to drive revenue growth while maintaining a stable risk profile across segments.
Tools
Python
Methodology
Exploratory Data Analysis (EDA), RFM Segmentation
Process
The workflow involved extensive data cleaning, including fixing data types for identifiers and financial fields, handling missing values, and removing 31 duplicate rows. Exploratory Data Analysis (EDA) assessed net profit and fraud impact. Finally, RFM segmentation was performed over K-Means clustering to create actionable campaign playbooks for Champions, Loyal, and Hibernating segments.
Output
The analysis revealed that 40% of customers are “Hibernating“, representing the largest opportunity for scalable revenue through reactivation campaigns like fee waivers. Recommendations include selectively increasing credit limits for high-ROI “Champions” and implementing DTI guardrails for high-risk segments. RevoBank can drive growth by scaling spend in high-value RFM segments while tightly controlling risk.
Scope of Work / Achievements
- Identified that 40% of the customer base is dormant, representing the primary lever for 10% transaction growth.
- Executed data cleaning on 5,599 rows to rectify mismatched identifiers and financial data types for reliable analysis.
- Developed RFM segmentation to prioritize high-value Champions who generate 7.3M average net profit with the lowest DTI.
- Recommended targeted reactivation strategies for 772 Hibernating users to drive incremental revenue without increasing portfolio risk.
Click here to see Full Deck and Python Code
