Challenge
E-commerce with a wide range of apparel offerings failed to effectively leverage data on customers' previous purchases. Despite a rich transaction history, the recommendation system was quite basic and did not pay attention to shoppers' preferences. In such a situation, cross- and up-selling were practically impossible, and buyers were getting misguided suggestions for more products
The company decided to implement a system with personalized recommendations based on recently purchased products.
Action
- Analysis of purchase history
All customer transaction data was uploaded to iPresso - what products they bought, when, in what categories, what were their characteristics (e.g. type of material, size). - Integration of recommendation frames with Vertex AI
The online store combined artificial intelligence and machine learning with product recommendations. Customers were shown items specifically tailored to their recent purchases, plus the recommendations dynamically changed - depending on what they were doing on the Web. - Recommendations for complementary products
"Frequently Bought Together" type sections were dynamically displayed on product pages and in the shopping cart. Recommendations were precisely tailored to the products currently viewed or added to the shopping cart, based on previous purchases by similar customers. - Automated post-purchase emails
Once the transaction was finalized, customers received personalized emails with product recommendations that matched their most recent purchase (e.g., "Buy a leather jacket from our latest collection and complement your styling").
Results
- The company saw a 14% increase in average order value. Customers were significantly more likely to add additional recommended products to their shopping carts that perfectly matched their needs and interests.
- Customers who received personalized recommendations after a purchase were more likely to return to the store for subsequent transactions. Retention increased by 18%.
- Customers appreciated that the store understood their needs and offered them exactly what they were looking for or might be interested in. A campaign with NPS surveys sent by iPresso showed that customer satisfaction increased by 19%.
