While there is some evidence that review length, review score, and argument frame can impact consumers' perceptions regarding the helpfulness of online consumer reviews, studies have not yet identified the most appropriate levels of such factors in terms of maximizing perceived helpfulness of these reviews. Drawing on Negativity Bias and Cue-Summation theories, we propose a theoretical model that explains online reviews' helpfulness based on specific characteristics of these reviews (i.e., length, score, argument frame). The model is empirically validated using two datasets of online consumer reviews related to products and services from Amazon.com and Insureye.com respectively. We also employ ANOVA analyses to reveal the levels of each of these characteristics that result in maximizing perceived helpfulness of online consumer reviews. Further, we employ an artificial neural network approach to predict the helpfulness of a given review based on its characteristics. Our findings reveal that the most helpful online consumer reviews are those that are associated with medium length, lower review scores, and negative or neutral argument frame. Our results also reveal that there is no major difference between the characteristics of the most helpful online consumer reviews related to products or services. Finally, our findings reveal that the most helpful factor in predicting the helpfulness of an online consumer review is the review length. Theoretical and practical contributions are outlined.