Coupang has filed a patent for a system and method that generates and presents relevant product search results based on a user’s past purchase history. The system retrieves a user’s search query, historical purchase data, and experimental data comprising a hierarchical list of product categories. It then determines queried product categories, generates a hierarchical list of historical product categories, and ranks the search results based on the user’s past purchases. The ranked list is then presented to the user. GlobalData’s report on Coupang gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on Coupang, AI for workflow management was a key innovation area identified from patents. Coupang's grant share as of September 2023 was 41%. Grant share is based on the ratio of number of grants to total number of patents.
Generating and presenting relevant product search results based on user history
A computer-implemented system for generating and presenting product search results based on a user query has been described in a filed patent. The system includes a memory storing instructions and at least one processor configured to execute the instructions. The operations performed by the system include retrieving a product search query by the user, at least one set of historical purchase data associated with the user, and at least one set of experimental data comprising hierarchical lists of product categories. The system then determines a plurality of attributes associated with the product query and at least one pattern associated with these attributes using machine-learning algorithms. Based on these attributes and patterns, the system generates a list of queried product categories. Additionally, the system generates a hierarchical list of historical product categories associated with the user using the user's historical purchase data and the experimental data.
Using machine-learning algorithms, the system ranks the list of product search results based on a relevancy metric to the user's product query. The relevancy metric can be based on factors such as past purchase volume or state of recency. Finally, the system presents the ranked list of product search results to the user.
The system also includes features such as determining a history of prior access to the historical purchase data and storing the data in a cache memory if there is no history of prior access. The list of historical product categories associated with the user is further ranked based on the frequency of the user's past purchases within each category. The order of displaying the product search results to the user is based on the ranking of each result.
The patent also describes a computer-implemented method for generating and presenting product search results based on a user query. The method includes similar steps as the system, such as retrieving the product search query, determining attributes and patterns, generating queried product categories, generating a hierarchical list of historical product categories, ranking the search results, and presenting them to the user.
Overall, the system and method described in the patent aim to improve the accuracy and relevance of product search results by utilizing machine-learning algorithms and historical purchase data. By considering user preferences and past purchase behavior, the system can provide more personalized and tailored search results to enhance the user's shopping experience.