Do Fit Opinions Matter? The Impact of Fit Context on Online Product Returns

Abstract

Product fit uncertainty is cited as one of the top reasons for high online product return rates. Fit describes how well a product suits a consumer’s needs. The value of a product drops sharply when it deviates from a customer’s ideal fit. In this study, we focus on ordinal fit, a type of fit attribute that can be ordered on a scale, e.g. the size of apparel, and the difficulty level of courses. By leveraging a change in the product review system at an online retailer, we examine the impacts of two types of fit information – fit valence (an overall evaluation of a product’s ordinal-fit attribute) and fit reference (a reviewer’s ordinal-fit attribute and her choice of the product’s fit attribute) – on returns of apparel goods. Using the lens of advice-taking, we reveal the important role of the context of fit opinions (i.e. fit reference) in facilitating shoppers to better interpret fit valence by enabling effective ordinal-fit adjustment and, consequently, reducing product returns. \n \nWe employ a predictive analytics framework for counterfactual prediction via the Generalized Synthetic Control method to address endogeneity issues and shed light on the dynamic treatment effect. Our findings indicate that fit valence alone can lower product returns only in a limited situation – when the majority of reviewers agree on the fit valence. In other cases – when either the fit valences are inconsistent or far and few between, it is the combination of fit valence and fit reference that lowers product returns. With the availability of both types of fit information, similar reviewers play an important role in helping improve the accuracy in ordinal-fit adjustments. Yet, albeit less effective, information from reviewers with dissimilar body sizes can also help make useful ordinal-fit adjustments. Besides, shoppers appear to benefit from both positive and negative fit valences, as long as they are aided by fit reference. Our empirical insights are relevant to many situations where ordinal-fit attributes dominate consumers’ product evaluation process. Accordingly, we provide useful implications for online sellers grappling with high product return rates.