Reviewer 4 (chair) Expertise Expert The Meta-Review The paper presents an approach to predict customers' fit preferences by exploiting their previous feedback. A latent factor formulation is proposed to cope with the fit problem. Although the paper has many strong points such as a sound formulation and evaluation, the main concern raised by the reviewers is that the proposed approach is not really related to recommendation but it looks more like a predictive and clustering method. The highly sparse nature of the dataset puts into questions the degree to which the recommendation can be evaluated. Final Recommendation after Meta-Review Probably reject: I would argue for rejecting this paper. Reviewer 1 (PC) Review Rating Probably reject: I would argue for rejecting this paper. Expertise Knowledgeable Contribution The authors of this paper are claiming of proposing a new predictive framework for product fit problem which captures the semantic of the customer feedback regarding the product fit instead of a binary feedback. Relevance to RecSys The proposed framework is rather a predictive model and a clustering method rather than a recommender system. I am not convinced that this paper has sufficient relevance to the RecSys focus of research. Your Review Positives: The problem is well defined and providing relevant real examples clearly describe the fit problem. Related literature is fairly presented. The performance of the proposed framework is well presented. A real case dataset is employed. Methodologies are well described in mathematical terms. Negatives: There are no future research directions provided for the paper. Reviewer 2 (PC) Review Rating Definite accept: I would argue strongly for accepting this paper. Expertise Knowledgeable Contribution This short paper presents an incremental refinement in approaches for product size recommendation to address discrepancies between product and customer “true” sizes. Relevance to RecSys This paper addresses algorithm scalability, performance, and implementations, as well as case studies of real-world implementations Your Review This short paper presents an approach for product size recommendation to address discrepancies between product and customer “true” sizes. The approach employs latent factor formulation for customer fit and metric learning for product size imbalance. The approach is evaluated on two datasets that were extracted from online clothing retailers as part of the research (and which are being made available). The problem of fit management for online clothing retailers is a very important issue for applied recommendation. The issue and approach are well motivated and described, and the evaluation shows the promise for this kind of approach being integrated for recommendation where fit is a key part of the user context. Reviewer 3 (PC) Review Rating Borderline: Overall I would not argue for accepting this paper. Expertise Expert Contribution The paper discusses a new method to provide size recommendations for clothing on online marketing sites. Relevance to RecSys relevant for Recsys Your Review The paper tackles an interesting problem, i.e. to suggest product sizes for buying / renting online clothing. The solution provided is original, as it takes into account costumer feedback and thus subjective product / size preferences. It is also technically sound, as far one can tell from a short paper, including an evaluation. It is unclear, however, how much the proposed solution really helps to take subjective customer feedback into account, given that the datasets used are really sparse, with most customers having only one transaction, for which the quality of the recommendation cannot be well evaluated.