Product Bundle Recommendation Server Using Complementary Relationship Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Retailers lack a convenient method to adjust or filter product recommendations on their websites to provide a more meaningful experience for consumers based on desired product attributes.
Innovation Solution
A product bundles recommendation server that receives a selection of a first product, accesses customer add to cart data, product price data, and degree of diversity data to generate a list of bundled products with complementary relationships, and presents these recommendations on a user interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If product recommendations are generated based on historical shopping activity data, then the relevance and personalization of recommendations improve, but the system complexity and data processing requirements increase
Solution Approach 1:
The patent segments the recommendation system into distinct functional modules: a data collection module that gathers historical shopping activity data, a data store that stores structured product and shopping data, and a recommendation generation module that processes the data. This segmentation allows each component to be developed, maintained, and scaled independently, reducing overall system complexity while maintaining personalization capabilities.
Solution Approach 2:
The system performs preliminary actions by pre-collecting and pre-structuring historical shopping activity data, product attributes, and pricing information in a data store before recommendations are needed. This advance preparation of data structures and relationships enables the recommendation engine to generate personalized recommendations quickly without complex real-time processing, thus reducing system complexity.
2Measurement precision
If product recommendations are filtered by price range and diversity criteria, then the quality and relevance of recommendations improve, but the computational processing time increases
Solution Approach 1:
The patent applies partial filtering by implementing optional price range filters and diversity criteria that can be selectively applied based on specific recommendation scenarios. Not all recommendations require full filtering - the system can adjust the level of filtering applied depending on the context, thus maintaining high recommendation quality when needed while reducing processing time when simpler recommendations suffice.
Solution Approach 2:
The system changes parameters dynamically by adjusting filtering criteria such as price range thresholds and diversity requirements based on the specific product category, user preferences, and business rules. This flexibility allows the system to maintain high recommendation quality for complex cases while using simpler, faster filtering for routine recommendations, thereby reducing overall processing time.
3Adaptability or versatility
If bundled products are selected based on complementary relationships from historical data, then the meaningfulness of recommendations to consumers improves, but the data storage and retrieval requirements increase
Solution Approach 1:
The patent extracts only the essential elements needed for generating meaningful recommendations from historical shopping activity data - specifically, the relationships between products that are frequently purchased together. Rather than storing complete historical transaction records, the system extracts and stores only the derived complementary product relationships and co-occurrence patterns, significantly reducing data storage requirements while maintaining recommendation meaningfulness.
Solution Approach 2:
The system creates simplified copies of historical shopping patterns in the form of pre-computed product relationship data structures. Instead of storing and processing raw historical transaction data, the system creates condensed representations that capture the essential complementary relationships between products. These copied relationship structures enable meaningful recommendations without requiring access to the full volume of original historical data.
Data Source
AI summary
Methods and systems for providing a product bundle recommendation are disclosed. One method includes receiving a selection of a first product. Data relating to historical customer shopping activity and data describing online shopping sessions in which products are selected for addition to historical shopping carts to which the first product was previously added is accessed. A list bundled products having a complementary relationship with the first product is generated. At least one product from the list of bundled products is selected and the at least one product is presented on a user interface on a customer device as a recommendation for purchase with the first product. A filter may be applied to the list of bundled products, such as a price filter. The method may be performed by a system including one or more computing devices communicating via a network with one or more data storage devices.