Comparison of aggregation methods used in online reviews: a critical analysis.

With digitalization, online reviews have become a crucial component of consumer decisions and are an important source of feedback on products and services. Users can make informed choices using online review systems such as e-commerce platforms, social media, and others. However, the collection and analysis of these reviews raises a number of questions regarding reliability, representativeness and accuracy. Since they directly influence users' choices, the methods used to aggregate review scores are extremely important. In this article, we explore the most common aggregation methods for rating online reviews on the Amazon platform, with a particular focus on the arithmetic mean. Although these methods are easy to implement, they do not always accurately reflect the overall quality of a product. We then critically analyze how more sophisticated approaches can minimize some of the limitations of the mean, such as the so-called recency bias; propose more robust solutions and alternative approaches that can provide a more accurate representation of the overall rating, improving the quality of aggregated ratings.