Audience Measurement System for User Attention Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Content providers face challenges in accurately measuring user engagement and attention, leading to incorrect billing and content selection decisions, as existing systems assume all users are engaged when content is delivered, despite some users being distracted or absent.
Innovation Solution
An audience measurement system estimates user presence and attention by analyzing interactions, comparing them to high-quality engagement data from small panels or using a temporal engagement curve, to distinguish between engaged and non-engaged users, thereby adjusting billing and content selection algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If content delivery systems assume all delivered content is actively consumed, then billing accuracy improves, but measurement precision deteriorates
Solution Approach 1:
The patent introduces an audience measurement system as an intermediary between content delivery and billing. This system uses panel data and engagement metrics to estimate actual user attention, serving as a mediator that translates raw delivery data into accurate engagement measurements without requiring direct complex monitoring of each user's attention state
Solution Approach 2:
The system implements feedback loops where engagement metrics from panelists and interaction data are continuously fed back into the measurement model. This feedback mechanism allows the system to refine its estimates of user attention over time, improving measurement precision through iterative learning rather than through complex real-time monitoring
2Measurement precision
If all users are assumed to be engaged with delivered content, then content selection accuracy improves, but measurement precision deteriorates
Solution Approach 1:
The patent segments the audience into different groups: panelists who provide detailed engagement data, general users whose engagement is inferred, and users who are estimated to be disengaged. This segmentation allows the system to preserve information about actual user interest by treating different user groups differently in the measurement model
Solution Approach 2:
The measurement system acts as an intermediary that recovers lost information about user interest by using panel data as a proxy. The panelists serve as intermediaries who provide detailed engagement information that is then used to infer the behavior of the broader audience, preventing complete loss of interest information
3Productivity
If interaction rates are used to measure engagement, then productivity improves, but measurement precision deteriorates due to over-representation of engaged users
Solution Approach 1:
The system uses feedback from panelist interactions to correct the bias in general user interaction rates. By comparing panelist behavior (who are known to be engaged) with general user behavior, the system can feedback-adjust the engagement estimates to remove the over-representation of engaged users and achieve more accurate measurements
Solution Approach 2:
The patent changes the parameters used to calculate engagement rates by introducing attention gap calculations and adjusting interaction rate weights based on panel data. This parameter transformation converts the biased interaction rates into corrected engagement metrics that more accurately reflect true user interest while maintaining productivity
Data Source
AI summary
The present disclosure provides systems and methods for estimating audience member engagement with content, or distinguishing between users consuming content and users who have become distracted or have left. User presence or attention may be estimated from user interactions with or skipping of content, with the user interactions either compared to high quality engagement data from small audience measurement panels, or extrapolated based on a temporal-engagement curve. An attention gap may be estimated, representing users that were not present for or not engaged with or paying attention to a presentation of content at a client device. This allows the measurement system to distinguish between users who consumed and potentially enjoyed the content, and users who did not, even as client devices of both sets of users receive and present the content items.