Multi-Provider Ad Selector Optimizing Fill Rates and Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current content delivery systems face challenges in achieving high ad fill rates and revenue optimization due to limitations in selecting the most suitable ad providers and managing content requests efficiently, leading to suboptimal performance, revenue, scalability, and utility.
Innovation Solution
A multi-provider ad selector (MPAS) system that receives ad requests, selects ads from a plurality of providers based on criteria such as keywords, user demographics, and ad provider statistics, and prioritizes providers to maximize fill rates, revenue, and response times, while transforming content formats as needed for different devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single response provider is used to simplify the system, then device complexity is reduced, but ad fill rate decreases
Solution Approach 1:
The system segments the ad provider selection into multiple independent components: a priority manager that assigns priorities to ad providers, a multi-provider ad selector that selects from multiple providers based on priorities, and a fallback mechanism. This segmentation allows the system to maintain multiple providers without increasing overall complexity, as each component has a specific function.
Solution Approach 2:
The patent introduces an intermediary layer (the priority manager and multi-provider ad selector) between the content delivery system and ad providers. This intermediary manages the complexity of multiple providers by implementing a priority-based selection mechanism, thereby maintaining ad fill rate without proportionally increasing system complexity.
2Reliability
If multiple ad providers are queried sequentially to improve fill rate, then ad fill rate increases, but response time increases
Solution Approach 1:
The system performs preliminary actions by pre-assigning priorities to ad providers and maintaining a ready-to-use priority list. When an ad request arrives, the system can immediately query providers in the predetermined priority order without delay, thus improving response time while maintaining high fill rates.
Solution Approach 2:
The patent implements dynamic priority assignment where ad provider priorities can be adjusted based on performance metrics, fill rates, and revenue considerations. This dynamic adjustment allows the system to optimize response time by directing requests to the most efficient providers while maintaining high overall fill rates.
3Loss of energy
If ad provider priorities are dynamically adjusted to maximize revenue, then revenue increases, but system complexity increases
Solution Approach 1:
The system implements feedback mechanisms where ad provider performance is continuously monitored and used to adjust priorities. The priority manager receives feedback on fill rates, revenue performance, and response times, then dynamically adjusts provider priorities to maximize revenue while maintaining manageable complexity through automated decision-making.
Solution Approach 2:
The patent changes the parameter of ad provider priority from static to dynamic, allowing priorities to be adjusted based on performance metrics. This parameter change enables revenue optimization through automated priority reassignment based on fill rates, eCPM, and other performance indicators.
Data Source
AI summary
A multi-provider response selector (MPRS) receives requests and provides respective selected responses. Each request includes respective selection criteria (such as keywords and demographic/geographic information) and/or respective other information (such as page location/time and demographic/geographic information). The MPRS uses the respective selection criteria to fill each of the requests (obtaining the respective selected responses) from one of a plurality of response providers, where the response providers are optionally selected in an order determined by the respective other information and/or other factors such as value returned from the response providers. Optionally, a content server receives content requests for respective specified content, generates requests for the MPRS, and returns the specified contents along with respective selected responses. In one example, a content request is from a cellular telephone, respective specified content includes video content, and respective selected responses include advertisements.