Collaborative filtering is specialized in suggesting appropriate products and services to the users concerning personal characteristics and past preferences without requiring any effort of users. It might be more efficient to collect preferences of users based on multiple subcriteria of products and services. For this purpose, researchers propose multicriteria recommender systems that are convenient for more accurate and useful evaluation of items. In such systems, it might be preferable to collect binary ratings instead of numerical ones due to the large number of subcriteria. However, there is a gap in the literature to satisfy a binary preferences-based multicriteria recommender system. In this study, the applicability of multicriteria recommender systems based on binary ratings is investigated. Firstly, recommendations for users on the overall criterion are produced by employing naive Bayes classifier. In order to improve the quality of recommendations, userand item-based similarity models are proposed enabling the formation of more successful neighborhoods. Such models are further improved by integrating a concordance measure between overall preference and subcriteria ratings, which helps to provide more personalized and meaningful similarities among users. Finally, a hybrid model is proposed employing userand item-based models together and real data-based experimental outcomes demonstrate that the quality of estimated binary referrals is improved statistically significantly.