E-Retail

Case 1

Client – A large online retailer selling books, music and movies

Opportunity: Create a recommendation engine to bundle products (cross-sell) and increase revenue

Solution: 

  • Collect a massive data set of millions of user sessions on search, click rate, product detail views, compare product and purchase
  • Analyze the data to distill the patterns leading to purchase; identify key indicators for purchase
  • Analyzed network of individual with similar actions (browsing, search, clicks, compare and purchase) to improve the algorithm
  • Create a behavioral model (based on users past behaviors and similar click streams) to suggest bundling in real time
  • Implement the algorithm on the website

Result

  • 8X increase in purchase rate
  • 140% improvement in transaction size

 

Case 2

Client – One of the largest media houses in India

Opportunity: Improve ROI on marketing to 12 MM mobile users

Solution: 

  • Improve targeting by creating finer segmentation by users interest
  • Compile the responses to past contests to understand users interest areas and define categories such as cricket, Bollywood, politics, gaming, romantic comedies
  • Solve the challenge of matching massive number of user responses to defined categories by using smoothed Levenshtein distance algorithm that is used for spell checkers, character recognition and translation
  • Segmented the users based on the model and tested the model to improve accuracy
  • Implemented the model for all mobile marketing activities

Result

  • 150% increase in response rate in mobile marketing