Etsy Turns to Google Cloud to Improve Search, Boost Sales

Etsy, the online crafts marketplace, boasts more than 60 million unique items and is using Google’s machine learning technology to help boost sales. Because more than 80 percent of its search-based purchases come from the first page of results, it is crucial to provide relevant items on that page. With that in mind, Etsy started a move to Google Cloud in 2017, motivated by the platform’s artificial intelligence capabilities. About three-fourths done with the migration, it’s already seen $260 million in incremental gross sales.

“Since Etsy’s products are mostly handmade by independent artists and typically customized for each order, the company doesn’t keep catalogs of items. It relies on machine-learning algorithms to serve up relevant products that match a customer’s search terms. The algorithms take the results from all product searches for the past several weeks and analyze buyer behaviors,” reports The Wall Street Journal.

The process is called ‘context-specific ranking,’ and it re-ranks search results in real time, making the first page as relevant as possible. This sort of work requires a significant amount of computing power, which Etsy can now access via Google Cloud “instead of purchasing additional computing capability for its own data centers,” according to WSJ.

Before moving to the cloud, Etsy utilized what’s called ‘index-based searching,’ which tries to match consumer search terms with seller descriptions. But that wasn’t always successful.

“For example, searches for “blue desk” would turn up a long list of blue desk calendars, blue desk lamps and blue desk clocks instead of a blue desk, which would often be on page five or 10,” explains WSJ.

Etsy’s engineering team is working on developing other complex machine learning models to do things like look at a product category (i.e. home decor) and assign styles automatically (i.e. modern or country manor).

“Engineers also are looking at ways to use more contextual data, such as location and time of year, to suggest relevant products. A person searching for a sweater in Michigan should come up with a list of search results that’s different from a person looking for a sweater in Florida,” notes WSJ.