# A two-dimensional framework for profiling online reviewer behavior

## TL;DR

This paper introduces a two-dimensional framework to analyze online reviewer behavior using indices that track extreme ratings and their direction, helping consumers and platforms better understand and trust reviews.

## Contribution

The novel two-dimensional framework introduces the Reviewer Extremeness and Polarity Indices to profile reviewer behavior and identify archetypes.

## Key findings

- The framework identifies nine archetypal reviewer profiles based on their extreme rating behaviors.
- Application to three million Amazon book reviews demonstrates the framework's practical utility in real-world contexts.
- The model helps detect potentially problematic rating patterns, such as those from incentivized reviewers.

## Abstract

Consumers frequently rely on extreme online reviews—highly positive or highly negative—for clarity and detailed insights. However, conflicting extremes can generate confusion and erode trust in rating systems, highlighting the need for additional metrics that provide deeper insight into reviewer behavior. To address this, we introduce a novel and intuitive two-dimensional framework for profiling reviewer behavior through two complementary indices: the Reviewer Extremeness Index (REI), which quantifies the frequency of extreme ratings, and the Reviewer Polarity Index (RPI), which measures the directional imbalance between positive and negative extremes, along with its intensity. The framework maps each reviewer onto a two-dimensional plane whose axes are REI and RPI, identifying nine archetypal profiles of reviewers’ historical extreme behaviors. As a case study, we applied this approach to three million Amazon book reviews, demonstrating its practical value in a real-world context. This framework provides dual utility. For consumers, it offers crucial contextual information: knowing a reviewer’s archetype allows for a more nuanced interpretation of their feedback. For online retail platforms, the framework serves as a scalable tool to monitor reviewer behavior and identify systematic rating patterns that may warrant further scrutiny, such as those potentially associated with incentivized reviewing. By making reviewer tendencies transparent, our model contributes to a more reliable and trustworthy digital marketplace ecosystem.