Beyond Means: A Dynamic Framework for Predicting Customer Satisfaction

2025-11-18

AI Summary

Online rating systems commonly aggregate scores using sample means, failing to capture temporal evolution of service quality and heterogeneity in reviews—such as sentiment polarity and perceived helpfulness—leading to biased customer satisfaction predictions. To address this, we propose a Gaussian process–based dynamic rating aggregation framework that jointly models review time series, fine-grained sentiment scores, and user-voted helpfulness weights, thereby co-characterizing rating dynamics and review information quality. Evaluated on 120,000 real Yelp reviews, our model reduces mean absolute error by 10.2% over conventional mean aggregation, significantly improving future rating prediction accuracy and platform credibility signaling. Our key contribution is the first application of Gaussian processes to heterogeneous temporal review aggregation, uniquely balancing temporal dynamics with semantic heterogeneity.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Reasoning under Uncertainty: Relational Probabilistic Models
• Planning, Routing, and Scheduling: Temporal Planning

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Online ratings influence customer decision-making, yet standard aggregation methods, such as the sample mean, fail to adapt to quality changes over time and ignore review heterogeneity (e.g., review sentiment, a review's helpfulness). To address these challenges, we demonstrate the value of using the Gaussian process (GP) framework for rating aggregation. Specifically, we present a tailored GP model that captures the dynamics of ratings over time while additionally accounting for review heterogeneity. Based on 121,123 ratings from Yelp, we compare the predictive power of different rating aggregation methods in predicting future ratings, thereby finding that the GP model is considerably more accurate and reduces the mean absolute error by 10.2% compared to the sample mean. Our findings have important implications for marketing practitioners and customers. By moving beyond means, designers of online reputation systems can display more informative and adaptive aggregated rating scores that are accurate signals of expected customer satisfaction.

Problem

• Predict customer satisfaction using dynamic rating aggregation
• Address limitations of static mean ratings over time
• Incorporate review heterogeneity in rating prediction models

Innovation

• Uses Gaussian process framework for rating aggregation
• Captures rating dynamics over time and review heterogeneity
• Reduces prediction error compared to sample mean method