Two-Stage Content Selection for Brand Value Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online content distribution systems face challenges in effectively pairing content item inventories with projected demand for video views, particularly in determining the brand value of video sources to optimize content placement and user engagement.
Innovation Solution
A two-stage selection process is implemented, where the online system first determines the brand value threshold based on user interactions and video views, and then performs an auction to decide whether to inject content items into videos, considering both projected benefits and potential engagement losses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If the online system presents content items in all videos to maximize revenue, then revenue generation is improved, but user engagement deteriorates due to potential disruption of high-brand-value content
Solution Approach 1:
The patent segments the video content library into different tiers based on brand value scores. High-brand-value videos are identified and separated from regular videos, allowing the system to apply different content insertion strategies to each segment. This segmentation enables selective preservation of high-quality content while still maximizing revenue from other videos.
Solution Approach 2:
The system dynamically changes the parameter of content insertion probability based on the brand value score of each video. For high-brand-value videos, the content insertion probability is reduced or set to zero, while for lower-brand-value videos, content items are inserted at normal or higher rates. This parameter adjustment resolves the contradiction by adapting revenue optimization to content quality.
2Manufacturing precision
If the online system performs comprehensive auction and selection for every video to optimize content placement, then content placement optimization is improved, but system complexity increases
Solution Approach 1:
The system performs preliminary action by pre-calculating and storing brand value scores for all videos in advance. This pre-computation allows the auction and selection processes to use ready-made metrics rather than calculating brand value in real-time for each video, significantly reducing system complexity while maintaining optimization quality.
Solution Approach 2:
The patent applies partial action by performing comprehensive auction and selection processes only for videos below a certain brand value threshold. For high-brand-value videos, the system skips the complex auction process entirely and either preserves the original content or applies simplified rules, reducing overall system complexity while maintaining optimization for the majority of videos.
3Reliability
If the online system filters out videos based on brand value threshold to preserve quality, then user engagement is improved, but revenue generation worsens due to reduced content item presentations
Solution Approach 1:
The system dynamically adjusts the brand value threshold parameter based on various factors including time of day, user preferences, content item inventory, and revenue targets. This dynamic parameter adjustment allows the system to optimize the balance between user engagement and revenue generation flexibly, raising the threshold when engagement is prioritized and lowering it when revenue optimization is the goal.
Solution Approach 2:
The patent introduces dynamics by making the brand value threshold adaptive rather than static. The threshold can change based on real-time conditions, allowing the system to respond to varying demands for content insertion. This dynamic approach enables the system to maximize revenue during periods when user engagement is less sensitive to content interruptions while preserving quality during peak engagement times.
Data Source
AI summary
An online system presents content in videos to users. Content providers may value having their content injected into videos from certain sources more than others. This preferences is quantified as a brand value score. The brand value score is determined as a function of user engagement with a source of the video and, to account for brand value, the system performs a two-stage auction. First, the system determines whether to inject any content into a video by determining a distribution of brand value of videos per demand for videos in a previous period and filling a projected demand for the content in a current period to determine a brand value threshold. Then, any videos having a brand value above the threshold are eligible for the second stage of the selection process where the system performs an auction where projected benefit of presenting the content is compared to projected loss.