NPS Driver Analysis Using LightGBM and SHAP for Digital Platforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current solutions fail to effectively utilize Net Promoter Score (NPS) data to ensure service quality and customer satisfaction on digital platforms, as they lack efficient analysis and recommendation generation, particularly in identifying NPS drivers and providing localized explanations for customer feedback.
Innovation Solution
A method and system that analyze customer feedback data to identify key features influencing NPS, fine-tune a subsystem using LightGBM models, and apply SHAP values to determine the contribution of each feature to the promoter rating, generating insights and governance parameters to provide targeted recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional survey methods are used to collect customer feedback, then data collection is straightforward, but the ability to generate actionable recommendations and identify underlying relationships in the data is insufficient
Solution Approach 1:
The patent introduces an intermediary analysis system comprising a processor and memory that acts as a mediator between raw customer feedback data and actionable recommendations. This intermediary layer applies machine learning models and SHAP value analysis to uncover underlying relationships, thereby recovering information that would otherwise be lost in traditional analysis approaches.
Solution Approach 2:
The patent replaces traditional mechanical/manual analysis methods with automated machine learning systems. The analysis unit employs trained models to automatically identify patterns, relationships, and drivers in customer feedback data, substituting manual analytical processes with computational algorithms that can detect complex underlying relationships.
2Measurement precision
If comprehensive customer feedback analysis is performed to identify all NPS drivers, then recommendation accuracy improves, but computational time and processing complexity increase
Solution Approach 1:
The patent implements preliminary action by pre-training machine learning models on historical customer feedback data before actual analysis. The analysis unit has access to pre-computed features and trained models, allowing it to quickly analyze new feedback without performing exhaustive analysis from scratch. This preliminary preparation significantly reduces processing time while maintaining high identification accuracy.
Solution Approach 2:
The patent applies partial action by focusing the analysis on the most influential NPS drivers identified through feature importance analysis. Rather than analyzing all possible factors equally, the system concentrates computational resources on the key features that have the greatest impact on NPS, achieving high precision with reduced processing requirements.
3Ease of operation
If detailed feature analysis is conducted to provide localized explanations for each customer, then customer satisfaction improvement increases, but system complexity and computational resources required increase
Solution Approach 1:
The patent applies segmentation by dividing the customer feedback analysis into distinct functional modules: data collection unit, analysis unit with machine learning models, explanation generation unit using SHAP values, and recommendation unit. This modular segmentation allows each component to specialize in specific tasks, making the overall complex system more manageable and maintainable while enabling detailed personalized analysis.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting the level of analysis detail based on customer segments and feedback types. The system can modify analysis parameters such as feature selection criteria, model complexity, and explanation depth to balance personalized recommendation quality with system resource requirements, making the system adaptable to different operational contexts.
Data Source
AI summary
A system and method for providing one or more recommendations. The method encompasses identifying, target feature(s) from a set of features influencing a customer feedback data associated with customer(s) of a digital platform. The method thereafter comprises fine-tuning, a sub-system based on the target feature(s). Further the method encompasses determining, a probability of a promoter rating for the customer(s) based on the fine-tuned sub-system and the customer feedback data. The method thereafter encompasses determining, a contribution of each target feature in the probability of the promoter rating. The method further comprises generating, at least one of a customer insight and a governance parameter, based on the contribution of each target feature in the probability of the promoter rating. Further the method encompasses providing, the one or more recommendations on the digital platform based on at least one of the customer insight and the governance parameter.