Product ratings and customer reviews are vital for e-commerce businesses, as they improve customer satisfaction and achieve competitive advantages over rivals. On the other hand, they also help customers make informed decisions when buying based on the experience and recommendations of previous customers, taking advantage of crowd wisdom. However, the interrelation between product rating numerical scores and its associated text-based feedback about product reviews is only sometimes obvious and needs to be investigated. To achieve this endeavor, this work investigates the predictability of product rating scores from the text-based review in customer feedback using Natural Language Processing (NLP), Text Mining, and Machine Learning algorithms, including (1) Logistic Regression, (2) Random Forests, and (3) KNN. This study uses a web-scraped dataset that represents customer ratings and feedback for thousands of products that span various categories and were sold online over a long time. Yet, while the results obtained highlight a positive correlation between the product rating score and the textual description of customer feedback, it turned out that such an association is not always obvious. This was reflected in the prediction performance score of the applied algorithms. The Logistic regression algorithm achieved the best classification performance score. It predicted product rating with an 81% AUC score and 70% classification accuracy.