Video View Value Determination for Guaranteed Delivery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online systems face challenges in predicting and ensuring the number of users who will view videos, as user engagement is unpredictable, making it difficult to guarantee video delivery based on campaign budgets.
Innovation Solution
An online system determines a video view value by pacing video delivery based on user engagement metrics, selecting users who match target audience specifications, and adjusting presentation to ensure a guaranteed number of views by aggregating individual view values and comparing them to predetermined limits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the online system presents videos to users based on campaign budget alone, then the system can control delivery costs, but it cannot guarantee a specific number of video views because user engagement is unpredictable
Solution Approach 1:
The system implements feedback by continuously monitoring actual video view counts and comparing them against predicted view counts. Based on this feedback loop, the system dynamically adjusts the number of video presentations to users, ensuring that the actual views match the guaranteed number while optimizing delivery pacing throughout the campaign period
Solution Approach 2:
The system performs preliminary action by predicting the number of video views before the campaign begins. This prediction is used to determine the guaranteed number of views in advance, allowing the system to plan and control delivery pacing from the start of the campaign rather than reacting after views are delivered
2Reliability
If the online system presents videos to more users to ensure minimum view counts, then video delivery reliability improves, but the campaign budget may be exceeded
Solution Approach 1:
The system applies dynamics by making the number of video presentations adjustable rather than fixed. The presentation count is dynamically modified based on the difference between actual and predicted views, allowing the system to respond to real-time engagement patterns while staying within budget constraints
Solution Approach 2:
The system changes parameters by adjusting the number of video presentations as a variable parameter. This parameter is modified throughout the campaign based on performance data, enabling flexible control over both view delivery reliability and budget consumption without being locked into predetermined values
3Reliability
If the online system presents videos for longer durations to ensure complete viewing, then user engagement quality improves, but the time required for campaign delivery increases
Solution Approach 1:
The system applies partial action by requiring users to view only a portion of the video (e.g., 50% or 75% completion) rather than demanding 100% completion. This partial viewing threshold is sufficient to ensure meaningful engagement while significantly reducing the time required for campaign delivery compared to requiring full video completion
Data Source
AI summary
An online system presents a video from a third party system based on a video view value generate by the online system. Based on a target audience specification from the third party system, the online system selects a plurality of online system users as the audience of the video. The online system presents the video the selected users in a duration of time equal to a video presentation period received from the third party system. A play of the video by a selected user constitutes a historical view. The online system generates a video view value for the video based on historical views of the video by the selected users. The online system further applies the generated video view value for pacing delivery of the video by the online system.