• Irina Mikhaylina

Visibly better recommendations for online stores



Since the first prototype, we strongly believe that our Visual Product Recommendations (VPR) will make the digital shopping experience more intuitive, significantly improving the user experience. Now, we have conducted a representative A/B test to investigate the impact of a classic recommendation engine and our Visual Product Recommendations. The results, measured by common eCommerce metrics such as click-through rate (+153%) and conversion rate (+258%), are clearly and amazingly in favor of Visual Product Recommendations.


A total of 170,000 sessions were tested on the schuhcenter.de site within one week. The only difference between the test groups was the product recommendation on the product detail page.





For the "DPR" (Default Product Recommendation) test group, recommendations from a professional provider of onsite search and marketing were displayed based on text attributes and the customer journeys.


For the "VPR" (Visual Product Recommendation) test group, intelligent recommendations that suggest style-appropriate alternatives based on visual product features were displayed.


Test results strongly favor the use of Visual Product Recommendations



Explanation of the key figures


Clickrate x 2,5



The click-through rate is the ratio of the number of times the recommendation was seen and subsequently clicked.


Conversion rate x 2,5



The conversion rate is the ratio of the number of times a recommendation was seen to the number of products purchased, which were clicked on in the recommendation and then added to the shopping cart.


Use rate x 3,1



The use rate is the percentage of visitors (session) who clicked the recommendation at least once.


Share of ordered products x 3,6



The purchased products were clicked in the recommendation and added to the shopping cart and then purchased in the same session.


Conclusion and recommendation for action


The intuitive visual perception and the strong expression of personality by the visual appearance are responsible for the significantly more effective impact of Visual Product Recommendations. When calling up a product page, website visitors already have a clear style preference for the selected product. Style-matching recommendations will be accepted as more helpful and interesting and, as with all good and personal customer service, visitors will feel properly addressed and be more likely to buy.


In addition to comparing two different recommendation solutions, the test basically confirmed that product recommendations support customers in their shopping process and lead to more purchased products. In our test, the visibility of Visual Product Recommendations was low and yet over 12% of visitors clicked on the recommendations, subsequently purchasing significantly more products (+264%).


This result is convincing proof that a higher visibility of Visual Product Recommendations will lead to more sales. Therefore, Visual Product Recommendations will be integrated more prominently on the product page at Schuhcenter, supporting visitors in their selection process on the product list page.


What can Visual Intelligence do for your store?

If you have any questions about this article or how visual intelligence can improve your company's online store, feel free to contact us any time. In a joint web meeting, we can quickly find the appropriate use cases and discuss possible next steps. Write us a mail to info@mediaopt.de or just call us +49 30 92102990.