Dynamic Short-Form Video Traversal for Ecommerce
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ecommerce platforms face challenges in effectively engaging users with short-form videos that lead to product purchases, as existing methods lack the ability to dynamically adapt and personalize video traversal based on user behavior and sales goals.
Innovation Solution
A computer-implemented method using a machine learning model to access and customize a graph structure of short-form videos, rendering them with interactive overlays, and analyzing user behavior to determine and synthesize next videos for enhanced engagement and sales, incorporating photorealistic 3D models and dynamic content generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional static video presentation methods are used, then implementation simplicity is maintained, but user engagement and personalization capability deteriorate
Solution Approach 1:
The patent implements dynamic video traversal where the system transitions from static, pre-defined video sequences to dynamically generated video paths based on real-time machine learning model predictions. The ML model continuously analyzes user behavior data and adjusts video selection, duration, and sequencing dynamically, allowing the system to adapt to individual user preferences while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The system employs self-service mechanisms where the machine learning model automatically generates personalized video traversal paths without requiring manual curation for each user. The model serves itself by using its own predictions to guide video selection, automatically optimizing engagement based on accumulated user behavior data, thereby reducing the need for complex manual personalization processes.
2Productivity
If generic video content is presented to all users, then content production efficiency is maintained, but user engagement deteriorates
Solution Approach 1:
The patent applies local quality by delivering customized video content to different user segments based on their specific behavior patterns and preferences. Instead of uniform generic content, the system analyzes individual user data and serves locally optimized video selections, durations, and sequences tailored to each user's engagement characteristics, thereby improving overall productivity through targeted personalization.
Solution Approach 2:
The system utilizes parameter changes by dynamically adjusting video content parameters (selection, duration, sequencing) based on machine learning predictions of user behavior. The ML model modifies these parameters in real-time based on user interactions, allowing the same video library to serve multiple purposes with optimized engagement outcomes without requiring separate content production for each user.
3Quantity of substance
If video duration is extended to provide more product information, then information completeness is improved, but user attention span and engagement deteriorate
Solution Approach 1:
The patent implements partial action by delivering only the necessary portion of video content based on real-time ML predictions of user engagement. Rather than presenting complete extended videos to all users, the system serves partial video segments tailored to individual user needs and attention patterns, optimizing information delivery without exceeding user tolerance thresholds for video duration.
Solution Approach 2:
The system applies segmentation by dividing video content into manageable segments that can be dynamically selected and sequenced based on user behavior predictions. The ML model analyzes user responses to different video segments and reconstructs personalized traversal paths that deliver comprehensive product information through strategically selected segment combinations rather than requiring users to watch complete extended videos.
4Adaptability or versatility
If manual video curation and sequencing is performed, then content quality control is maintained, but scalability and adaptability deteriorate
Solution Approach 1:
The patent replaces the mechanical system of manual video curation with an automated machine learning-based system. The ML model substitutes human curators by automatically analyzing user behavior data, predicting engagement patterns, and generating personalized video traversal paths at scale. This substitution maintains content quality through algorithmic optimization while enabling dynamic adaptation to individual user preferences without the scalability constraints of manual processes.
Data Source
AI summary
Disclosed embodiments provide techniques for dynamic short-form video traversal with machine learning in an ecommerce environment. A graph structure associated with a library of short-form videos is accessed and customized in a back-end environment based on products for sale on a website. One or more of the customized short-form videos from the library are rendered to one or more users, along with an interactive overlay and an ecommerce environment. As the video is viewed, video consumption behavior data is collected and analyzed by a machine learning model. The machine learning model determines one or more next short-form videos from the graph structure for the user to view, based on sales goals, video consumption behavior data, and interaction with the user. The machine learning model can synthesize additional short-form videos and insert them into the graph structure in order to enhance viewer engagement and product sales.