PackPack Case Study AI-powered desire ignites 35% cart surge.
Creating custom offers for every e-shop visitor is both time-consuming and resource-intensive. However, not when you utilize AI-based automated recommendations from ui42. The increasing number of conversions through effective cross-selling and up-selling is evident without much delay. The more data artificial intelligence gathers, the better results it achieves. The conversion rate continually grows after implementing the Recommendation System from ui42, just as it did for this client.
To validate the benefits of our solution for client PackPack, we decided to conduct an A/B test on the category page. The results of the A/B test revealed that users who were presented with AI-driven category arrangements engaged more with individual products, either by clicking through to the product details or by adding them to their cart more frequently.
Adding to the cart through categories
+ 34,81 %
when utilizing AI from ui42.
Results of AB testing in the period 05/2023 - 10/2023.
Original - We displayed the original solution to customer group A, which was the classic product sorting within the category.
In this case, it was sorting by the newest products.
32 032
impressions
6 472
click on product
1 646
add to cart
Automated AI-based recommendations - Group B displayed the updated solution.
Based on the user's previous web activity, we organized all products in the respective category.
30 327
impressions
6 481
click on product
2 101
add to cart
Assignment and Collaboration Objectives
The aim of the collaboration was to achieve higher conversions through personalized offers, increase user time spent on the website, and achieve greater interaction with products. In order to display relevant offers to every e-shop visitor, which would enhance their interest in products, it was necessary to find an automated solution. We implemented automated recommendations into the project. This allowed us to achieve the set goals for the client in a short time frame without increasing their personnel costs for administration. We verified the results through an AB test.
Key Milestones of Collaboration
We have deployed artificial intelligence in the e-shop
Each e-shop has its own specifics, which is why we configured automated recommendations with respect to the uniqueness of the project. The foundation is a minimal dataset, such as a product feed, which the system needs for machine learning. The more quality data, the better. Consequently, acquiring additional data over time allows the system to learn and improve its outputs.
We have set up AI for product recommendations in the product details, pop-up window, shopping cart, and also in category sorting through recsys.
There are several ways artificial intelligence can increase conversions in an e-commerce store. The most common ones are recommended products or upselling (you may also like..., customers who bought this also added... to their cart). In this case, we utilized cross-sell and upsell extensively in product details, categories, shopping cart, and pop-up windows.
We verified the benefits of recommendations through an AB test.
To compare the results of the previous solution - classic product sorting, and the new solution in the form of automated recommendations, we conducted AB testing. The result was a higher level of interactions with the products displayed within the automated recommendations.
What are the automated recommendations from ui42?
Automated recommendations are based on machine learning, an area of artificial intelligence capable of processing large amounts of data to create a unique personalized offering. As a result, the system does not need to know any specific user data. It can generate the most likely product or service recommendations based on a large amount of general input data for website visitors.
Large platforms like Netflix, Youtube, Amazon, and others also operate on this system. Thanks to the Recommendation System developed by ui42, these technologies are available for medium-sized and smaller e-commerce sites in Slovakia, typically not dealing with such a volume of data. The recommendation tool can fully automate the recommendation of complementary products or upselling in the product details, pop-up windows, shopping cart, or various other locations in the e-shop.
Benefits of Automated Recommendations for E-commerce Stores
- Automation without the need for administrator intervention
- Working with real user data (interests)
- Displaying relevant products/services that the user is most likely interested in
- Increasing the conversion rate in the e-shop
- Extended user time spent on the website
- Enhancing the user experience
Contact us
Would you like automated recommendations for your e-shop as well?
-
Boris Henezi
Sales Executive