Localized Product Ranking via User Intent Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing e-commerce search result ranking algorithms do not account for user intent, leading to undesirable products being listed high in search results, despite their popularity.
Innovation Solution
A system that re-orders search results based on a user's predicted intent by tracking and analyzing their online interactions, such as search terms, product hierarchy paths, and previous purchases, to provide a customized product list that prioritizes relevant products over popular ones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If products are ranked according to overall popularity, then high-popularity products are listed higher in search results, but undesirable products may appear high in rankings from the user's perspective
Solution Approach 1:
The patent segments the general product popularity ranking into user-specific localized rankings. Instead of a single global ranking based on overall popularity, the system creates segmented rankings tailored to individual users by analyzing their online interactions, purchase history, and behavior patterns. This segmentation allows the system to present different ranked orderings to different users, improving relevance while maintaining the benefit of popularity-based ranking where appropriate.
Solution Approach 2:
The patent applies local quality by customizing the search result ranking for each individual user rather than applying a uniform global ranking. The system analyzes user-specific data such as online interactions, purchase history, and behavior patterns to create localized rankings that reflect each user's preferences and intent. This ensures that the ranking quality is optimized locally for each user context rather than being uniformly average across all users.
2Device complexity
If search results are ranked by popularity alone, then the ranking algorithm is simple to implement, but it fails to account for user intent and behavioral context
Solution Approach 1:
The patent applies preliminary action by pre-analyzing and storing user interaction data, purchase history, and behavior patterns before the actual search ranking occurs. The system performs preliminary user profiling and intent analysis in advance, so that when search results need to be ranked, the complex user-specific parameters are already computed and ready to be applied. This reduces the real-time computational complexity while still incorporating comprehensive user intent information.
Solution Approach 2:
The patent introduces an intermediary layer between the simple popularity ranking and the final user-specific ranking. This intermediary component analyzes user behavior data, purchase history, and interaction patterns to generate user-specific ranking parameters that modify the base popularity ranking. The intermediary translates complex user intent information into actionable ranking adjustments, bridging the gap between simple algorithmic implementation and sophisticated user understanding.
Data Source
AI summary
A system for providing product recommendations to online visitors to an e-commerce website is provided. The system may include program comprising instructions that, when executed by a processor, cause the processor to sort a list of products based on a comparison of a user's interactions with the e-commerce website and previous user interactions with the same e-commerce website.