Query Identification Using Relevancy Scores for Ad Targeting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing advertising systems often fail to accurately identify relevant queries for targeted advertisements, leading to less-than-desirable results due to broad keyword matching and the absence of 'missing keywords', which can result in advertisements linking to landing pages that are not highly relevant to user interests.
Innovation Solution
The system identifies candidate queries from query logs, generates relevancy scores for web documents, and selects those exceeding a threshold to associate with the queries, creating query-page tuples that are filtered and matched with advertisements to ensure relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If broad keyword matching is used for advertisement selection, then the coverage of advertisement targeting is improved, but the relevance of advertisements to user queries deteriorates
Solution Approach 1:
The patent segments the advertisement matching process into multiple stages: initial broad keyword matching to identify candidate advertisements, followed by query-log based relevance assessment to filter and rank these candidates. This segmentation allows the system to first cast a wide net for coverage, then progressively refine results for precision.
Solution Approach 2:
The patent introduces query logs as an intermediary data structure that captures actual user query patterns. These logs serve as a mediator between broad keyword matching and final advertisement selection, providing empirical evidence of query-relevance relationships that bridge the gap between coverage and precision.
2Ease of operation
If advertisers specify only broadly matched keywords, then the ease of advertisement setup is improved, but the accuracy of query matching deteriorates
Solution Approach 1:
The system performs self-service by automatically analyzing query logs to discover and generate specific query-page tuples without requiring manual advertiser input. This automated discovery process maintains ease of advertisement setup while significantly improving query matching accuracy through data-driven insights.
Solution Approach 2:
The patent performs preliminary analysis of query logs to pre-identify relevant query-page relationships before actual advertisement serving. This preliminary action creates a repository of validated query-relevance pairs that enhance matching accuracy without adding complexity to the advertisement setup process.
3Device complexity
If missing keywords are not accounted for in keyword specification, then the simplicity of keyword management is improved, but the completeness of query coverage deteriorates
Solution Approach 1:
The patent implements feedback loops where query log data continuously informs keyword strategy. By monitoring actual user queries against advertised content, the system identifies missing keywords and patterns, then uses this feedback to refine query-page tuple associations, thereby improving query coverage while maintaining simple keyword management.
Solution Approach 2:
The system performs preliminary mining of query logs to proactively identify missing keywords and emerging query patterns before they impact advertising effectiveness. This advance preparation ensures comprehensive query coverage is achieved automatically without requiring complex manual keyword management.
Data Source
AI summary
Apparatus, systems and methods for predictive query identification for advertisements are disclosed. Candidate query are identified from queries stored in a query log. Relevancy scores for a plurality of web documents are generated, each relevancy score associated with a corresponding web document and being a measure of the relevance of the candidate query to the web document. A web document having an associated relevancy score that exceeds a relevancy threshold is selected. The selected web document is associated with the candidate query.