Client Details
Client is a financial services giant headquartered in New York
Challenges Addressed
Identify users who are likely to make a purchase in the future by modeling online user behavior
- Email campaigns
- Website visits
- Form submissions & Referral traffic
Highlights of Engagement
- Identifying customer segments by implementing K-means clustering algorithm
- Feature engineering based on segments – derived temporal and spatial features
- Predicting likelihood of purchase / conversion for each segment of customers – derived the likelihood using Random forest mode with 90% recall and precision
Benefits to client
- Market Segmentation and targeted spend of Marketing $
- Improved credit card transactions due to micro targeted offers
- Targeted promotions and offers based on spending behavior
- Reports on Top merchants / Top Industry spends
