GlassesUSA.com Sees Clear Benefits from Dynamic Yield

GlassesUSA.com is one of the largest online stores for eyewear, selling top brands like Ray-Ban and Oakley and its own products in more than 190 countries. The New York-based company uses advanced tools to help customers select the right frames and best fits at competitive prices.

“We promise satisfaction; we have very high-standard return policies, a money-back guarantee, insurance, and much more, said Ben Blanki, director of product labs at GlassesUSA.com, who adds that the company has experienced strong growth over the past few years.

But GlassesUSA.com still wanted to better understand its customers so that it could make better recommendations. “We put our customers at the center of everything we do. Instead of giving generic recommendations, we wanted to lean on what someone had already seen, loaded to their cart, etc. We also wanted to be able to do A/B testing, which I think is critical for e-commerce growth,” Blanki says.

GlassesUSA.com started looking for a vendor nearly three years ago and ultimately selected Dynamic Yield.

Though the initial relationship with Dynamic Yield started about 30 months ago, GlassesUSA.com became a beta tester for Dynamic Yield’s deep learning recommendations algorithm to offer improved recommendations for customers toward the end of 2019.

“When we initially tried it, the results were amazing. We saw a huge lift in conversion, so we fully implemented it,” Blanki recalls.

The implementation of the algorithm was the most difficult part, according to Blanki. “It takes time and resources to see the results.”

The initial implementation didn’t go as well as hoped, Blanki admits, blaming the initial poor results on his own company’s lack of dedicated resources and time. But once GlassesUSA.com started to see the great value that was possible, it made the commitment. When the back end was fully developed, the speed of the recommendations delivered from the algorithm improved greatly, meaning its use didn’t slow down the e-commerce process.

“The back end is the most expensive part, but it was worth it,” Blanki says. “It took a little under three weeks to completely implement the solution, complete with APIs.”

Now Blanki sees Dynamic Yield as more of a partner than just a business supplier.

GlassesUSA.com saw quick results once the Dynamic Yield deep learning algorithm was fully implemented. These included an 87 percent growth in revenue, a 68 percent increase in purchases, and a double-digit lift in add-to-cart rate (the percentage of customers who make an initial selection and then add additional items to their shopping carts before completing checkout).

Dynamic Yield’s algorithm also provides GlassesUSA.com with better A/B testing, Blanki says. “Now I’m testing different strategies every two to three weeks.”

Those strategies include tweaks to the home page and subsequent category pages, according to Blanki. “Our A/B testing culture has really developed.”

Going forward, GlassesUSA.com plans to continue to rely on Dynamic Yield as it further looks to customize and optimize its site. Modifications it is considering include where to place recommendations on the site as well as mixing things up a bit for all customers, for new customers, for repeat customers, etc.

“We are going to test a lot of different types of placements for different displays and different strategies,” Blanki says. “That is the first step to a more personalized approach for our customers. It will be more like an Amazon-type of approach. We’re going to look at changes in small features that can increase conversion rates.” 

The Payoff

As a result of using Dynamic Yield's deep learning recommendations algorithm, GlassesUSA.com has seen the following results:

• an 87 percent growth in revenue;
• a 68 percent increase in purchases; and
• a double-digit increase in add-to-cart rate.