Leveraging AI/ML In Ecommerce For Lifetime Value Analysis & Autonomous Bid Management System

Leveraging AI/ML In Ecommerce For Lifetime Value Analysis & Autonomous
Bid Management System

The Challenge

A US-based ecommerce player, was facing a squeeze on profitability. Given rising cost of customer acquisition through digital marketing and having to drop prices due to competitive pressures, the firm wanted to optimize marketing spend on consumers with positive long-term financial impact. Tatras Data developed a ML algorithm that tracks the impact of changes in digital marketing performance to continuously improve bidding rules while focusing on high lifetime value consumers.

Hypothesis

  • Customer lifetime value varies for different digital channels and product SKU’s purchased.
  • Customer transaction and historical transaction data can help predict the expected lifetime value of a prospect on the aggregate and SKU level.
  • Losing money on the first transaction by aggressive bidding on certain channels can be profitable in the long run, through repeat transactions through cheaper channels.

Execution

  • A Customer Expected Lifetime Value model was implemented using Survival Analysis and Regression methods.
  • Variance across different campaigns and channels with respect to the expected lifetime value were observed.
  • An algorithm of optimizing spend across these channels was developed that tweaked bids and budget of different campaigns to maximize expected lifetime value of acquired customers.

Outcomes

  • Life Time Value predictions were implemented with autonomous digital marketing bid management system for dynamic execution.
  • Reduced marketing budget waste by 30%.

Project Highlights

16%

INCREASE IN CUSTOMER LIFE TIME VALUE

30%

REDUCTION IN MARKETING BUDGET WASTE