Marketing

---

Customer Labeling System

Pain Points

• Manual customer labeling was based on highly-subjective decisions
• Lack of standardized customer screening processes to find best customers and lower customer churn rate
• Difficult to differentiate customers with the current analysis system, which resulted in improperly allocated enterprise resources

System Highlights

• RapidsDB is a distributed, in-memory database
• AIworkflow integrates BI tools to provide multi-dimensional analysis
• External data is connected and federated to the unified system to provide more insights

Results

• Established a closed-loop system to create, manage, apply and evaluate customer labels
• Helped find potential KOL users based on the insights generated by the system
• Implemented fine-grained customer screening layer, improving the promotional conversion rate by nearly 1.5%