Ad Selection Scoring via Video Engagement Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for presenting ads to users in video hosting services often fail to show ads that are of interest to users, as they are not effectively tailored to user interactions or preferences.
Innovation Solution
A method is developed to calculate a score for ads based on user interaction metrics with videos, using an estimated cost per mille (eCPM) that incorporates a predicted click-through rate (pCTR) and cost per click (CPC), where the pCTR is influenced by a net interest score derived from viewer interactions, to select and present relevant ads to users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional ad presentation methods are used, then ads can be displayed to users, but the ads are not of interest to users and lack relevance
Solution Approach 1:
The patent changes the parameters used for ad selection from conventional methods to a scoring system based on eCPM that incorporates pCTR and video engagement metrics. This parameter change enables the system to select ads that are more relevant to user interests while maintaining manageable complexity through standardized metric calculations.
Solution Approach 2:
The system implements feedback loops where user interactions with videos (views, likes, comments, shares) are continuously measured and fed back into the scoring system. This feedback mechanism allows the ad selection to adapt and improve relevance over time by learning from actual user behavior patterns.
2Measurement precision
If ads are selected based on user interaction metrics, then ad relevance improves, but calculation complexity increases
Solution Approach 1:
The patent segments the complex measurement task into distinct components: video engagement metrics (views, likes, comments, shares), pCTR calculation, and eCPM computation. Each component is calculated separately using specific formulas, which simplifies the overall system architecture while achieving precise user interest measurement through the aggregation of these segmented metrics.
3Productivity
If eCPM scoring is used to select ads, then profitability and relevance improve, but processing time increases
Solution Approach 1:
The system performs preliminary calculations of video engagement metrics and stores them in advance. When an ad selection decision is needed, the pre-computed metrics are readily available, eliminating the need for real-time analysis of raw interaction data. This preliminary action significantly reduces processing time while maintaining the effectiveness of eCPM-based ad selection.
Data Source
AI summary
The present invention provides methods for determining which ads to present to a user. An embodiment of the method comprises identifying one or more ads, at least one of the identified ads associated with a video. For each identified ad, a first score is calculated. The first score for the identified ad associated with the video is calculated based on one or more metrics representing user interactions associated with viewing the video. In one embodiment, the score for the ad associated with the video will be better (e.g., higher) the more viewers of the video hosting service interact with the associated video, since such interactions thereby indicate a higher over level of viewer interest in the video. One or more of the identified ads are selected to be presented to the user based at least in part on the first score of each of the identified ads. The one or more selected ads are transmitted to a device for presenting to the user.