Eye Tracking Data Predicts User Discontinuation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining user satisfaction with complex online software applications are limited by relying on indirect metrics like time spent and number of clicks, which may not accurately reflect user experience, leading to potential abandonment of services.
Innovation Solution
The use of eye tracking data to analyze user interactions, including pupil dilation and saccadic movements, to predict the likelihood of user discontinuation and implement targeted interventions to reduce abandonment, such as offering discounts or support, based on calibrated user experiences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional interaction metrics (time spent, number of clicks) are used to measure user satisfaction, then implementation complexity is low, but measurement precision is insufficient leading to inaccurate user experience assessment
Solution Approach 1:
The patent uses eye tracking technology as an intermediary to capture objective physiological data about user attention and engagement. The eye tracking system serves as a mediator between the user's internal cognitive state and the external measurement system, providing more accurate indicators of user experience than traditional interaction metrics alone.
Solution Approach 2:
The patent replaces mechanical interaction metrics (clicks, time spent) with physiological measurements (eye movements, pupil dilation). This substitution transitions from measuring voluntary mechanical actions to capturing involuntary physiological responses, thereby improving measurement precision of genuine user engagement.
2Measurement precision
If eye tracking data is collected and analyzed in real-time to predict user discontinuation, then user experience measurement accuracy improves, but processing complexity and computational requirements increase
Solution Approach 1:
The system performs preliminary calibration of eye tracking data during initial user interactions to establish baseline patterns. This preliminary action prepares the system for more efficient real-time prediction by pre-processing and normalizing data streams, reducing the computational burden during actual prediction operations.
Solution Approach 2:
The patent implements feedback loops where prediction results are continuously refined based on actual user outcomes. The system learns from whether users actually discontinue or continue using the application, adjusting prediction models to improve accuracy over time while adapting to individual user behaviors and patterns.
3Reliability
If proactive interventions are implemented based on eye tracking predictions, then customer retention improves, but system complexity and intervention management burden increase
Solution Approach 1:
The patent applies interventions locally and selectively based on specific user states and contexts rather than universally. Different intervention strategies are tailored to specific situations (e.g., confusion vs. disinterest vs. technical problems), ensuring that interventions are appropriately matched to user needs while managing system complexity through targeted rather than blanket approaches.
Solution Approach 2:
The system implements self-service interventions where possible, such as automatically providing help resources, adjusting interface elements, or offering targeted assistance without requiring manual intervention from support staff. This reduces the operational burden while maintaining retention benefits through automated, data-driven responses.
Data Source
AI summary
Techniques are disclosed for determining application experience of a user. One embodiment presented herein includes a computer-implemented method, which includes receiving, at a computing device, eye tracking data of a user interacting with at least a first page of an application. The computer-implemented method further includes determining, based at least on the eye tracking data, at least a current user experience regarding the first page. The computer-implemented method further includes predicting, based on evaluating the current user experience, that the user is likely to discontinue use of the application. The computer-implemented method further includes determining, based at least on the prediction, an intervention that reduces a likelihood of the user discontinuing use of the application, and interacting with the user according to the intervention.