Token Architecture for Precomputed Content Item Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online systems face challenges in increasing the likelihood of publishers selecting content items from their platform while efficiently using resources, as they compete with third-party systems for display opportunities on digital platforms.
Innovation Solution
The online system implements a token architecture and uses multiple content item value predictors to determine and provide content item values quickly and accurately, reducing latency and improving the chances of selection by publishers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the online system determines content item values in real-time for each impression opportunity, then the accuracy and relevance of content item values are improved, but the latency increases and response speed deteriorates
Solution Approach 1:
The system performs preliminary determination of content item values before the actual impression opportunity arises. By pre-computing values based on user information and publisher identifiers, the system avoids real-time calculation delays while maintaining accuracy. The pre-determined values are stored and retrieved during impression opportunities, resolving the contradiction between accuracy and response speed.
2Reliability
If the online system requests content item values from multiple third-party systems to ensure selection, then the likelihood of publisher selection is improved, but the resource consumption and system complexity increase
Solution Approach 1:
The patent introduces an intermediary mechanism where the online system acts as a mediator between publishers and content providers. Instead of publishers directly querying multiple third-party systems, the online system receives a single request, internally evaluates multiple content items from different providers, and returns one optimized content item value. This intermediary approach maintains high selection likelihood while reducing system complexity and resource consumption.
3Speed
If the online system pre-determines content item values in advance, then the response latency is reduced, but the adaptability to real-time user context and impression opportunities deteriorates
Solution Approach 1:
The system performs preliminary determination of content item values before the actual impression opportunity arises. By pre-computing values based on user information and publisher identifiers, the system avoids real-time calculation delays while maintaining accuracy. The pre-determined values are stored and retrieved during impression opportunities, resolving the contradiction between accuracy and response speed.
Data Source
AI summary
An online system provides content item values for content items to be displayed via publisher servers on client devices of users. For example, the content items include text, images, or video for display on user interfaces such as webpages. The online system may compete with other third party systems that also provide content items for display via the publisher servers. To reduce latency between requests and responses by the online system, the online system may pre-determine a candidate content item value before an opportunity occurs to display a content item. The online system may associate candidate content item values with tokens provided to client devices. Additionally, the online system may use different types of content item value predictors that provide a range of content item values and that require varying amounts of latency to determine the content item values.