Query Clustering for Sponsored Search Auction Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional online content sponsorship allocation methods, such as auctions, face challenges in achieving optimal slot assignment on search engines, leading to costly and unpredictable advertiser spending due to the short time frame for displaying search results and the interchangeability of queries as commodities.
Innovation Solution
The approach clusters queries based on advertiser behavior, forming mini-markets that share similar targeting and intent, allowing for incremental adjustment of auction parameters to maximize click yield and revenue, using techniques like modularity maximization and co-clicked campaigns to group related queries and advertisers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If queries are treated as interchangeable commodities in conventional auctions, then slot allocation can be simplified, but advertiser spending becomes unpredictable and revenue optimization is difficult
Solution Approach 1:
The patent segments queries into distinct clusters based on advertiser behavior patterns, co-clicked campaigns, and semantic intent. This segmentation transforms the homogeneous treatment of all queries into differentiated clusters, allowing for customized auction parameters for each cluster while maintaining overall system manageability.
Solution Approach 2:
The patent changes auction parameters dynamically based on query cluster identification. By determining cluster-specific auction parameters such as click-through rate weights and pricing multipliers, the system adapts the auction mechanism to different query types, improving both revenue optimization and advertiser predictability.
2Productivity
If auction parameters are optimized for each individual query, then revenue maximization improves, but the complexity of the auction system increases significantly
Solution Approach 1:
The patent reduces system complexity by segmenting queries into a finite number of clusters rather than treating each query individually. This segmentation allows the system to manage a limited set of cluster-specific parameters while still capturing the diversity of query characteristics, balancing optimization granularity with computational feasibility.
Solution Approach 2:
The patent creates query clusters that serve as universal groups with shared characteristics and optimal parameters. Each cluster acts as a multi-functional unit representing multiple queries with similar advertiser behavior patterns, allowing the system to apply a single set of auction parameters to multiple queries simultaneously.
3Speed
If the auction system processes each query independently in real-time, then response speed is fast, but optimal slot assignment becomes computationally expensive and difficult
Solution Approach 1:
The patent performs preliminary clustering of queries based on historical advertiser behavior, co-clicked campaigns, and semantic analysis before the actual auction occurs. This pre-processing creates ready-to-use query clusters with predetermined optimal parameters, enabling fast real-time auction execution without repeated complex computations for each query.
Solution Approach 2:
The patent segments the computationally intensive task of query analysis into cluster formation and parameter determination steps performed in advance. By organizing queries into clusters beforehand, the system avoids repeating full analysis for each query during real-time auctions, significantly reducing computational costs while maintaining response speed.
Data Source
AI summary
Aspects of the subject disclosure are directed towards managing sponsored online content based upon advertiser behavior. Defining mini-markets to represent such advertiser behavior may be accomplished by clustering queries that generate revenue from one or more campaigns. Query revenue data between queries and a set of campaigns may be used to determine such mini-markets. To illustrate, a query whose highest revenue is attributed to a campaign may be selected for that campaign's mini-market. When that query is entered as a search term, the campaign's mini-market helps allocate space for advertisements.