Product Quality measurement is important for an e-commerce company as it helps to shape demand in favor of good quality products. Selling poor quality leads to high-returns and bad-ratings in the short term; and customer-dissatisfaction and disengagement in the long term. Here, we present a ratings and returns based approach to quantify product-quality and its impact on our bottom-line.

What can the audience expect?

Understanding of the Machine learning model built for product quality measurement, challenges in working with diverse data of e-commerce companies, and architecture of building scalable machine learning models and automated modeling frameworks.

Session Plan

• Introduction to Product-Quality
• Explaining the problem statement – Listing Quality Score
• Data formats and data collection
• Feature Engineering / Variable Transformations
• Feature Selection
• Hyper-parameter tuning Automation and Scaling of modeling framework
• Model-Selection
• Model Validation and A/B experimentation
• Impact analysis

Speakers

Shilpi Bhabhra

Shilpi leads Retail and Marketplace analytics practice at Flipkart. Her expertise is in driving data-based decisions for product and business. During her 14 years of work experience, she has led several data sciences projects for implementation in the field of Retail and Marketing. Her passion lies in solving real-life problems through analytics and data sciences.

Sumit Gupta

Sumit is a member of decision science team at Flipkart – a team focused on solving machine learning intensive – high business impact problems. In his 8+ years of work experience, he has worked with American Express (RIM CoE), WalmartLabs, Swiggy and Flipkart (current) working on a range of interesting problems like Product Quality, Selection, Recommendation Engine, Revenue Forecasting, Credit Risk Modelling.

Sumit has an Integrated M.Tech in Mathematics and Computing from IIT Delhi (class of 2010).