Sparkling results with personalized recommendations

Dynamic Yield automatically selected the right recommendation strategy for each user, context, and KPI.
* Email not necessary to download
Jewelry.com case study header image
15%
Increased revenue from homepage
14%
Increased revenue from product pages
18%
Increased revenue from cart page

Introduction

Jewelry.com is a world-class curated jewelry retailer, with a growing roster of thousands of products from over twenty emerging and established designers. The company turned to Dynamic Yield to improve its product recommendation capabilities and increase purchases from recommendations. Using Dynamic Yield’s recommendations, Jewelry.com increased revenue per visitor by 39% from the homepage, 13% from product pages, and 18% from cart pages.
Jewelry.com logo
“With Dynamic Yield, we can use machine learning to make data-driven recommendations based on where visitors are in the sales funnel. The ability to assess the level of valuable information about each visitor and automatically serve the most effective strategy has empowered us to increase revenue across our site.”
Jon Azrielant, Director of Marketing
The Challenge section thumbnail

The Challenge

While typical recommendation engines leverage the main recommendation strategies, most do not allow merchandisers to fully control their recommendations. Being confined to a recommendation vendor’s “black box” often prevents brands from executing critical recommendation tactics, such as serving differing quantities of recommendations based on where visitors are in the sales funnel or embedding widgets in custom locations on a website or app screen.

To break free of black boxes and increase revenue from product recommendations, Jewelry wanted a data-driven yet flexible solution that would allow it to:

  • Implement merchandiser expertise and set custom recommendation conditions for specific pages and products

  • Use a machine-learning engine to automatically select the most effective strategy for each user, context, or KPI merchandising rules

Execution

Combining multiple strategies to increase homepage conversions
As visitors move through the sales funnel, they reveal important signals about their buying intentions and preferences for specific products. A key question for Jewelry became: how can it use its behavioral data to:

1. Personalize recommendations to visitors with a rich history of behavioral interactions

2. Promote top-selling products to new visitors, upon whom minimal information is known

To address the challenge of targeting customers in different parts of the sales funnel, Jewelry first used Dynamic Yield to implement a recommendation widget on it’s most-valuable real estate: above the fold on the homepage. Next, the company then set custom conditions for the widget to independently target new and returning customers.
Combining multiple strategies to increase homepage conversions
To increase conversions among returning customers, the widget leveraged affinity-based recommendations, recommending products according to a weighted score of what the user had added-to-cart, viewed, or purchased in the past.

To induce engagement among new visitors, the widget presented products with the highest amount of pageviews and click-through-rate on the site.
Finding the right product page strategy to increase cross-sells
Finding the right product page strategy to increase cross-sells
While many suggest the ‘most popular’ and ‘similar to the current item’ strategies to increase cross-sells on product pages, Jewelry wanted to gobeyond conventional wisdom and find the most effective recommendation strategy for its own product pages.

To do this, Jewelry implemented a Dynamic Yield widget and customized the widget’s conditions to automatically determine the most effective recommendation strategy to leverage, using purchase completions as the success metric.

45% of users were recommended products ‘similar to the current item,’ 45% were recommended ‘top-selling’ products, and 10% of users received a control variation. As a result of continuous testing, Jewelry revealed that recommending ‘top-selling’ products provided a 10% uplift compared to the other variations
Increasing upsells on cart pages
Finally, the company added an additional widget on the bottom of the cart page showcasing items frequently bought together with the current item. This encouraged shoppers to make an impulse decision on the spot rather than requiring further navigation around the site to find relevant products.
Increasing upsells on cart pages

The Key Takeaway

No single recommendation strategy will be equally effective for all site visitors. Look for a solution that assesses the level of valuable information about each visitor and deploys the most appropriate strategy based on behavioral data, current context, or general popularity signals. Targeting visitors based on where they are in the sales funnel has become an essential eCommerce recommendation tactic. Make sure your solution allows for flexible strategy setting, allowing you set unique targeting rules and determine when and to whom each recommendation strategy should be served.

GET YOUR DEMO