Analytical Model Training for Customer Experience Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for quantifying customer experience rely on limited survey responses and fail to incorporate diverse customer input sources, such as warranty data, telematics, and social media, lacking optimization tools for improving model reliability and accuracy.
Innovation Solution
A computer-implemented method for training an analytical model that combines machine repair records, survey records, and machine operation records to estimate customer experience by associating machines with defects and repair characteristics, and identifying time-based operation parameters, creating a model that outputs customer experience estimates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If survey-based methods are used to quantify customer experience, then implementation simplicity is maintained, but measurement precision and reliability deteriorate due to limited response proportions and small sample sizes
Solution Approach 1:
The patent merges multiple data sources including survey responses, machine repair records, warranty claims, telematics data, and social media information into a unified customer experience model. This combination allows the system to overcome the limitations of survey-only approaches by incorporating diverse data streams that collectively provide a more comprehensive and precise measurement of customer experience while maintaining model implementability through systematic integration.
Solution Approach 2:
The analytical model is designed to process multiple types of input data serving different functions - survey responses provide direct customer feedback, repair records indicate product reliability issues, telematics data offers operational insights, and social media captures broader sentiment. This multi-functional approach enables a single model to deliver comprehensive customer experience measurement that surpasses any single data source alone.
2Ease of manufacture
If conventional survey methods are used, then data collection simplicity is maintained, but information completeness deteriorates due to inability to incorporate diverse customer input sources
Solution Approach 1:
The system combines diverse information sources including quantitative survey data, structured repair and warranty records, telematics operational data, and unstructured social media content. This merging approach recovers and integrates information that would be lost in survey-only methods, providing a complete view of customer experience across multiple touchpoints and data types.
Solution Approach 2:
The patent introduces an analytical model as an intermediary that processes and synthesizes data from multiple sources. This intermediary component translates diverse input formats into unified customer experience metrics, enabling the system to incorporate information from various sources while maintaining data coherence and interpretability.
3Device complexity
If single-source data methods are used, then model complexity is minimized, but reliability deteriorates due to inability to incorporate multiple customer touchpoints
Solution Approach 1:
The patent merges data from multiple customer touchpoints including direct surveys, product repair interactions, warranty claims, telematics monitoring, and social media engagement. This combination of diverse sources enhances the reliability of customer experience estimates by cross-validating information across different channels and reducing dependence on any single potentially biased source.
Solution Approach 2:
The system incorporates feedback loops where the analytical model processes multi-source data and generates customer experience estimates that can be continuously refined. The model learns from patterns across different data types and touchpoints, improving its reliability over time through iterative processing of feedback from multiple sources rather than relying on static single-source data.
Data Source
AI summary
An analytical model training method includes generating a modified set of machine repair records by associating a machine with at least one of i) one or more defects associated with the machine, ii) a repair of the one or more defects, and iii) one or more repair characteristics. An analytical model training system generates a training model based from the modified set of repair records and a set of survey records indicative of reported experience values. The system isolates at least one time-based machine operation parameter associated with machine operation from a set of machine operation records. A modified set of machine operation records is generated that includes a plurality of values associated with the at least one time-based machine operation parameter. An analytical model is configured to output an estimate of customer experience based at least partly on a second set of machine repair records.