Data Platform Pollyanna

Business Objectives

  1. Develop robust platform to handle millions of merchant offers per month
  2. Classify at least 50k Delta (New) offers across FR (2864) & DE (2863) categories
  3. Cost efficient on-boarding of 1319 FR and 342 DE merchants with data feed updates on a daily basis
  4. Remove inaccuracy, misclassification and insufficient product information from merchant data feed


  1. Applied stochastic prediction model to classify a merchant feed based on self learning algorithm
  2. Created portal for customer’s QC and generated updates and status reports
  3. Hybrid approach brought annualized savings exceeding $100,000 and sales growth by 20%
  4. 10 million merchant feeds were classified in six months using Pollyanna
  5. Achieved 100% quality levels were achieved by moving files to QA team in US for readability check
  6. All product pages started to get indexed in Google, resulting in more than 70% increased traffic