Negative Keyword Identification for Ad Relevance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current advertisement targeting systems face challenges in accurately filtering out irrelevant queries, leading to decreased relevance and increased presentation of unrelated advertisements, which affects click-through rates and user experience.
Innovation Solution
A system and method for identifying negative keywords associated with content items, such as advertisements, by determining search criteria relevance using Bayesian networks and evaluating off-topic queries to generate a list of exclusion keywords, thereby reducing the likelihood of presenting irrelevant ads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If advertisement targeting systems use keyword matching to serve ads, then ads can be presented to users based on search queries, but irrelevant ads are presented increasing and decreasing relevance
Solution Approach 1:
Instead of only identifying positive keywords to match ads with relevant queries, the system inverts the approach by automatically identifying negative keywords (words that should exclude ads). The system analyzes search queries and ad performance data to determine which words, when present in queries, should prevent ad presentation. This inversion allows the system to filter out irrelevant ads by blocking specific keyword combinations rather than relying solely on positive matching.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring ad performance data, click-through rates, and search query patterns. This feedback is used to automatically refine and update the negative keyword lists. When ads are presented and user interaction data is collected, the system analyzes this feedback to identify which queries resulted in low-value interactions, thereby automatically adding those query terms to negative keyword exclusions. This closed-loop feedback enables continuous improvement of ad relevance over time.
2Reliability
If the system filters out more irrelevant queries, then ad relevance improves, but the complexity of evaluating search criteria increases
Solution Approach 1:
The system segments the complex task of query relevance evaluation into multiple independent components: positive keyword matching, negative keyword exclusion, and query scoring. By separating these functions, the system can independently optimize each component. The negative keyword module operates independently to filter out specific query patterns, while the positive keyword module handles matching. This segmentation reduces the complexity of the overall evaluation process by breaking it down into manageable, independent evaluation stages rather than requiring a single complex analysis.
Solution Approach 2:
The system performs preliminary action by pre-identifying and storing negative keywords before ad presentation occurs. Through continuous analysis of historical data, the system builds and maintains a database of negative keywords that can be quickly referenced during query evaluation. This preliminary identification and storage of exclusion criteria allows the system to rapidly filter out irrelevant queries without performing complex real-time analysis, thereby reducing the computational complexity during actual ad serving operations.
Data Source
AI summary
The present disclosure includes a system and method for identifying negative keywords associated with advertisements. In some implementations, a method for advertising includes identifying one or more keywords associated with an advertisement. The keywords used to embed the advertisement in Web pages such as those including search results. By scoring keywords to identify those keywords that are not relevant to the web page content, one or more negative keywords can be identified in order to limit the embedding of non-relevant advertisements in web page content.