Selective Video Item Displays Under Bandwidth Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users overlook items they are unfamiliar with or do not know how to use due to static images lacking context, leading to fewer sales on online concierge systems, while dynamic content requires more bandwidth and can impact performance.
Innovation Solution
An online concierge system uses a trained machine learning model to determine display and bandwidth scores for videos, selectively replacing static images with dynamic content to maximize lift within acceptable bandwidth thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If static images are used to represent items, then bandwidth usage is low and system performance is maintained, but users cannot obtain contextual information about item usage leading to fewer sales
Solution Approach 1:
The system dynamically replaces static item images with video content based on real-time predictions of user interest and contextual relevance. The machine learning model predicts which users are likely to benefit from video content and selectively activates video displays only for those users, transforming the static display system into a dynamic one that adapts to user needs while controlling bandwidth consumption.
Solution Approach 2:
The system changes the information content parameter by introducing video content that demonstrates item usage, preparation, and context. This parameter change provides users with actionable information about how to use items, directly addressing the information loss problem and potentially increasing sales conversion.
2Productivity
If dynamic video content is used to replace static images, then user interaction and sales are enhanced, but bandwidth usage increases impacting system performance
Solution Approach 1:
Instead of uniformly applying video content across all users and items, the system applies video content locally and selectively only to specific user-item combinations where the machine learning model predicts high benefit. This localized application of dynamic content maximizes sales impact while minimizing overall bandwidth consumption.
Solution Approach 2:
The system implements partial action by replacing static images with videos only for a subset of items and users where the predicted lift is sufficient, rather than applying video content universally. This selective approach achieves meaningful sales improvement without the excessive bandwidth consumption that would result from full video deployment.
3Loss of information
If multiple videos are displayed in the user interface, then item information and promotion are enhanced, but user navigation performance deteriorates
Solution Approach 1:
The system segments the user interface by replacing only specific static images with video content based on predicted user interest, rather than converting all images. This segmentation allows the interface to maintain its original navigation speed for most items while providing enhanced information selectively where it will have the greatest impact on user decision-making.
Data Source
AI summary
An online concierge system selectively replaces default static item displays with dynamic item displays to represent items. The dynamic item displays encourage a viewing user of the online concierge system to purchase the items and may be selected based on item or user preferences or characteristics. The online concierge system applies a machine learning model to determine display scores describing the expected benefit of dynamic item displays and bandwidth scores describing resource usage of dynamic item displays. The online concierge system selectively replaces default static item displays with dynamic item displays based on the display and bandwidth scores so as to maximize benefit while ensuring that performance of the online concierge system is not negatively impacted by the resource usage.