Style Board Recommendation System for Product Complementarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing product recommendation systems primarily rely on a canvas-based approach, recommending individual products rather than 'looks' or combinations of products, limiting their ability to understand how products complement each other when worn together and failing to provide flexible recommendations based on shared styles.
Innovation Solution
A system that allows geographically distributed stylists to recommend 'looks' to users, using a recommendation management system that includes a stylist computing device, client computing device, and a recommendation management system to normalize product records, add affiliate links, and transmit style boards, enabling recommendations based on shared looks and analytical insights.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a canvas-based approach is used to recommend individual products, then the system is simple to implement, but it cannot understand how products complement each other when worn together
Solution Approach 1:
The patent merges individual product recommendations with product combination context by creating style boards that associate multiple products with a shared image. This allows the system to track and recommend products that complement each other when worn together, while maintaining the underlying simple product recommendation structure.
Solution Approach 2:
The patent adds a new dimension to product recommendations by introducing style boards that link products to shared images. This creates a two-dimensional structure where products are connected both individually and in combinations, enabling the system to understand product relationships without completely redesigning the recommendation engine.
2Measurement precision
If stylists manually create style boards with product selections, then recommendations can be personalized and accurate, but the process is time-consuming and reduces productivity
Solution Approach 1:
The patent implements self-service functionality where the system automatically generates style board recommendations by analyzing client preferences and existing product data. This reduces the manual workload for stylists while maintaining personalized and accurate recommendations through automated preference matching and product association.
Solution Approach 2:
The patent incorporates feedback mechanisms where client interactions with recommended products (views, purchases, saves) are tracked and used to refine future style board recommendations. This continuous feedback loop improves recommendation accuracy over time while reducing the need for manual stylist intervention.
3Manufacturing precision
If product recommendations require products to be depicted in an image, then the recommendations are visually accurate, but it limits flexibility in making recommendations
Solution Approach 1:
The patent segments the relationship between products and images by allowing multiple products to be associated with a single style board image, rather than requiring each product to have its own image. This segmentation enables flexible recommendations where products can be recommended in combinations without needing individual visual representations for each item.
Data Source
AI summary
A set of computing methods and systems are provided that allow geographically distributed stylists to review and recommend “looks” for geographically distributed users, and that allow for asynchronous access by both the stylist and the user. Using “looks” instead of individual products allows the system to build an understanding of what products complement each other when worn together. Analytical insights of trends, style recommendations, customer behavior, and other internal data can be made available to stylists in order to help stylists more efficiently provide recommendations to multiple customers. Further, embodiments of the present disclosure may allow recommendations to be based around a look, but do not require that recommended products be depicted in an image associated with the look. In this way, greater flexibility can be achieved than if only products depicted in the look image were recommended.