Representing the very top of the funnel in the customer journey, GlassesUSA.com decided to revisit an area just below the fold where it had historically displayed a recommendation widget showcasing up to six different products. Hoping to extract as much value out of this front-and-center placement, the eCommerce team hypothesized that if it could provide recommendations more heavily tailored to the individual upon entry to this page, it could not only improve add-to-cart rates, but increase purchases and revenue overall. After all, a classic collaborative filtering strategy that showcases items of interest based on what other similar users have interacted with can be highly effective, but the recommendations are not truly personalized.
After hearing about how Dynamic Yield’s deep learning recommendations could not only mine the past behavior of users but also their in-session activity, GlassesUSA.com set up an experiment to test the AdaptML™ recommendation strategy against collaborative filtering in its homepage widget for all desktop traffic.
Automatically configured per site, product feed, and individual, the team made but a few minor tweaks to the strategy before quickly seeing impressive results, most notably a 45% increase in add-to-cart rate, a 68% increase in purchases, and an 88% uplift in revenue attributed to the deep learning-based recommendations. And after running a similar test on mobile, the advanced algorithm proved yet again to be the strongest performer when compared to the control, with the team at GlassesUSA.com making deep learning the sole strategy for its popular homepage widget on this channel.