logo

NAVER Boosts Hyper-Personalized Commerce through AI Product Recommendation Technology, Elevating Customized Shopping Experience

- NAVER incorporates hyperscale AI “HyperCLOVA” into its self-developed AI product recommendation technology “AiTEMS” to strengthen customized product recommendation
- NAVER introduces “custom blocks” that show product recommendations based on shopping history as well as the reasons for the recommendations, and “recommendation blocks” that suggest additional shopping interests, providing a detailed and expanded shopping experience
- With sophisticated analysis of the correlation between search queries and shopping history, NAVER enables hyper-personalized recommendations based on a database of 1.5 billion products, raising expectations for enhancing exposure and sales of SME products

July 26, 2023

NAVER Corporation (CEO Choi Soo-yeon) is strengthening its AI-based product recommendation technology to provide a stronger hyper-personalized AI product recommendation service.
 
On July 20, NAVER combined its self-developed AI product recommendation technology, “AiTEMS,” and the hyperscale AI “HyperCLOVA” to apply the upgraded AI technology in shopping search. When users enter product-related keywords in the NAVER search, they will be provided with a “custom block” that analyzes their shopping history and shows the reasons behind the product recommendations, as well as a “recommendation block” that suggests shopping interests based on user queries.
 
Considering a higher demand and user satisfaction for AI-powered personalized recommendations, NAVER plans to further enhance AI curation through these personalized product recommendation blocks. In June, there was a 30% increase in sales for AI-recommended products, compared to the previous year.
 
From “Personalized shopping for OOO” to “#Brunch items frequently searched together” NAVER exerts efforts to enhance technology for hyper-personalized product recommendation and search
 

[Image] “Custom blocks” (left and middle) that show the reasons for recommendations based on shopping history and search queries and “recommendation blocks” (right) that recommend shopping themes and products of user interest
 
To provide a hyper-personalized experience for a product recommendation, NAVER has strengthened the personalized recommendation model and engine structure, and built a large-scale data recommendation system to ensure that a database of 1.5 billion NAVER Shopping products can be recommended with increased diversity and precision.
 
In particular, the “custom blocks” leverage advanced modeling technology to analyze users’ shopping histories in real-time, as well as the reasons behind the recommendations, immediately presenting the users with product recommendations based on their preferences and search queries.
 
For instance, if a user who has previously clicked on or had a history of adding moisturizing cream to their cart, wish list, or purchasing the item, searches for “moisturizing cream” again, the custom block for the user will show the recommended product along with the reasons for the recommendation, such as “Clicked on this product 3 days ago,” “Purchased twice at this store in the past 3 months,” or “Visited this brand 8 times in the past 3 months.” 

 

As these blocks take into consideration different queries and shopping histories depending on the user, the effectiveness of personalized recommendations is further enhanced even with the same keyword. Internal pre-tests showed that the click-through rate in the custom blocks was up to 27% higher than the previous shopping recommendation model. 
 
The “recommendation blocks” leverage NAVER’s hyperscale AI “HyperCLOVA” to recommend shopping themes and products of user interest based on the type of search query. HyperCLOVA analyzes the correlation between users’ search queries and shopping histories and suggests the keywords that might be of interest to the users. 

 

For example, when searching for “basil,” users who have a history of viewing flowerpots or other plants will be presented with a “#Balcony Décor items frequently searched together” block showcasing air-purifying plants and seed products. On the other hand, users who have searched for dishes or salad dressing will be recommended with the “#Brunch items frequently searched together” block, suggesting bagels or other salad vegetables.
 
By recommending unique products transcending the confines of popularity, AI presents expanded opportunities for SME products

As NAVER's AI-based personalized recommendation technology evolves, users can anticipate a more refined and extensive product exploration experience. This means that unique products from SMEs will also have a better chance of being discovered by users.

 

Currently, the AiTEMS technology recommends 35% of less popular SME products and 52% of new store items, effectively reducing the bias toward mainstream, popular items. According to internal preliminary tests, when the “Personalized shopping for OOO” block with stronger personalized recommendations was enabled, product click-through rates more than doubled compared to the traditional shopping section, and transaction volumes increased fourfold compared to the previous shopping recommendation model. This suggests that there will be significantly more opportunities for SMEs to boost sales through this personalized product recommendation block.
 
Lee Jung-tae, NAVER’s executive officer who leads commerce AI technology, stated, “NAVER’s AI-based recommendation technology has become increasingly sophisticated through the combination of AiTEMS product recommendation expertise, accumulated since 2017, and the hyperscale AI model, HyperCLOVA. NAVER’s technology encompassing search and shopping continues to evolve, and based on Team NAVER’s technical synergy, we aim to connect user satisfaction with SME growth while cultivating a distinct competitive edge in AI technology within the industry.”
 
In 2017, NAVER established an AiTEMS-based personalized recommendation model, expanding its coverage in shopping and search to advance the technology and services. In 2021, NAVER Shopping introduced the AI-powered personalized recommendation service “FOR YOU,” and since October 2022, the “FOR YOU Shopping Section” has been available on the NAVER App. <END>