If you’re in ecommerce, you take inspiration from the leaders. One of the talks we’re really excited about at this year’s Activate conference is with Suyash Sonawane and John Castillo from online retailer Wayfair. As one of the largest online destinations for home, Wayfair knows a thing or two about what its customers shopping onsite need and what they expect. With two million users daily, and more than 10 million products in its catalog, the scale of their operations covers, “a zillion things for the home.”
Their session, “Using Opinion Mining and Sentiment Analysis to Discover Hidden Product Features for E-Commerce Search” explores how to discover insights in the things your customers say.
When you’re in ecommerce, it’s paramount that you’re able to describe products in the written word. Describing physical dimensions or the material of a product can only do so much—often times, the best words to describe an item are those used by your very own customers. Enter stage left: Product reviews.
Senior Engineers Suyash Sonawane and John Castillo ingested a myriad of customer signals and processed millions of product reviews in order to extract useful information. They processed this data using Natural Language Processing. By combining these techniques, they developed insights about what customers say about different products.
In their talk, Sonawane and Castillo detail how Wayfair’s Search Tech team has looked beyond the product catalog to improve search relevance onsite and influence customer experience at scale.
Imagine the possibilities from this. Being able to augment your product descriptions and search terms with what customers actually say, for instance “this barbecue grill is leak-free” yields an important characteristic.
Customers also tell you when things are going wrong as well as alternatives, “I’ve been ordering this for a long time but recently the quality has gone done down so I switched to SuperX brand.”
These sorts of insights go beyond analytics and allow you to optimize keywords, make alternative recommendations, detect when something has gone wrong and even how to resolve common problems. This is a type of customer signal that with the right technique you can mine to optimize your conversion rate and your relationship with your customers!
In other words, you can Activate your AI and search capabilities like Wayfair.
- Watch Wayfair’s “Opinion Mining & Sentiment Analysis for E-Commerce Search” talk
- Find out about Retail and Recommendations
- Contact us, we’d love to hear from you.