A shopper who enters a search query on an e-commerce or m-commerce site and receives 9.5 million results will likely be a tad overwhelmed. And that’s what was happening too often at Etsy.

At Etsy.com, a 10-year-old online marketplace selling handmade crafts and vintage goods, consumers can shop for more than 30 million items from sellers across the globe.  About one-third of those searches are what Etsy classifies as low intent, or consumers who it deems are shopping around leisurely using general search terms such as “jewelry” as opposed to more specific phrases like “gold earrings,” says Jaime DeLanghe, senior product manager at Etsy.

Etsy today changed its algorithm so that instead of sifting through the 9.5 million items that match the search query “jewelry,” the search results display the items in categories within jewelry, such as rings, necklaces and bracelets. Consumers can then tap or click on each category to browse the items in each category.

“The goal is to make sure users find what they are look for quickly,” DeLanghe says.

Etsy is ranked No. 51 in Internet Retailer’s 2015 Mobile 500 Guide, which ranks the top 500 m-commerce retailers according to mobile sales, and No. 24 in the 2015 Top 500 Guide, which ranks the top 500 retailers according to online sales.

In an A/B testing period, consumers clicked on categorized search results 10% more often than on non-categorized results, DeLanghe says. The percentage was higher on smartphones, DeLanghe said, but declined to give specifics. The higher click rate is important, since Etsy is hoping to boost visitor time on its site and decrease bounce rate with the new site search release, DeLanghe says. Bounce rate measures the percentage of site visitors who leave after viewing just one page.

Providing whittled-down product categories that make navigating search results easier is especially helpful for shoppers on smartphones, since it’s easier for consumers to tap on product categories on a smartphone than to type in words to get to what they are looking for, DeLanghe says. Etsy’s product listings are from individual sellers, which means they aren’t classified by SKUs. Since SKUs contain defining characteristics of products, such as size, color and manufacturer, that means often times typical keyword searching isn’t effective on the site, so having the categories helps consumers filter through products, she says. 

Catering to mobile shoppers is important to Etsy’s bottom line as 57.7% of Etsy’s website visits stemmed from mobile devices and 41.4% of gross merchandise sales came from mobile devices during the first quarter of 2015, Etsy says.

Search results for more specific terms, or what Etsy calls high-intent queries, such as “silver link bracelet,” will show up as a list, uncategorized.  Results are the same on smartphones and desktops.

Etsy began overhauling the search process a year and half ago when the retailer hired a full-time taxonomist with a background in library science to comb through the site and determine the best way to improve search results.

Etsy originally thought it could use machine learning to improve search results without human involvement. But DeLanghe says the e-retailer found that the automated process did not work well. Machine learning is when computers learn without being explicitly programmed. In February of 2015, Etsy asked all of its sellers to categorize its own items, and more than 90% of Etsy’s active merchants have categorized at least one item, she says. Active sellers are Etsy sellers who have listed, advertised, or sold an item in the last 12 months.

“Re-categorization took people rather than a lot of computation power to make this happen,” DeLanghe says.  

 

Follow mobile business journalist April Dahlquist, associate editor, mobile, at Mobile Strategies 360, @Mobile360April

Sign up for a free subscription to Mobile Strategies 360, a new weekly newsletter reporting on how businesses in all industries use mobile technologies to communicate with and market and sell to their consumers.Mobile Strategies 360 is published by Vertical Web Media LLC, which also publishes Internet Retailer, a business publication on e-retailing.

Favorite