Zappos, the online retailer, has always been at the forefront of adopting new technologies to improve its business. Over the last two years, the company has been overhauling its search algorithms using machine learning. At VentureBeat’s Transform 2019 conference in San Francisco, Zappos lead data scientist Ameen Kazerouni described how his team implemented semantic search on their website.
With semantic search, websites can determine what people are really looking for and avoid returning a bunch of wrong or incorrect items that the customer has to filter through. Kazerouni and his team, however, did not just implement semantic search. They took it a step further. The team decided that the contextual meaning behind a search term changes on a customer basis. Therefore, they strove to individually unique search results for the millions of unique search terms used by millions of unique customers. In essence, it is a one-to-one understanding of the individual and the term they applied.
Initially, the company’s search team, which maintains the database of words in the Zappos.com search index, used an old lexical-based algorithm. The results were, however, poor. Consequently, the search team approached Kazerouni and asked his data science crew for help. Part of the issue had to do with the language itself. Therefore, the team set out to understand search terms, taking in search terms, and looking at customer behavior and building machine-learning models that could create what is known as word embeddings.
The first tests of Zappos’ new semantic search algorithm were positive, resulting in a significant increase in click-through rates and engagement on the website. Kazerouni and his team have now evolved beyond word embeddings. The team has built neural networks to enhance its semantic search engine.
For Zappos, the semantic search engine has been a huge success. It has led to more searches and an increase in revenue. Additionally, tech-savvy consumers are helping the team improve their search engine by describing their experience in feedback surveys — with some customers specifically mentioning the algorithm.
Click here to read the article published in Venture Beat.
Please give your feedback on this article or share a similar story for publishing by clicking here.
