Research examining online reviews and ratings has generally found higher ratings result in positive perceptions and ultimately more sales, but little is known about how polarized ratings and/or reviews influence consumer behavior and purchasing outcomes. We argue that controversy is an important boundary condition for predicting how aggregated ratings and review valences impact user behavior and product outcomes. Using cognitive dissonance theory, particularly research on predecisional distortion and the blemishing effect, we develop and test three hypotheses that examine whether average-rated products that are a result of consensus elicit different consumer behaviors and product outcomes than average-rated controversial products (i.e., products with polarized ratings). Utilizing four years of real-world rating, commenting, and purchasing behaviors pertaining to over 60 thousand designs and two thousand products from a creative crowdsourcing ecommerce site, we find that negative information mixed with positive feedback can produce net positive results. Products which receive more polarized premarket feedback generate more community discussion, community sales, and follow-through purchases. The findings suggest that, in situations where consumers have a developing preference for an item, encountering comparatively negative information can in fact have a positive impact on commenting and purchasing behaviors—more so than if no opposing information were present.

检查在线评论和评级的研究通常发现，较高的评级会带来积极的看法并最终增加销售额，但对两极分化的评级和/或评论如何影响消费者行为和购买结果知之甚少。我们认为，争议是预测综合评分和评论效价如何影响用户行为和产品结果的重要边界条件。使用认知失调理论，特别是对预决失真和瑕疵效应的研究，我们开发并测试了三个假设，以检验作为共识结果的平均评级产品是否会引发与平均评级有争议的产品不同的消费者行为和产品结果（即，具有极化评级的产品）。利用四年的真实世界评级，评论，和来自创意众包电子商务网站的 6 万多款设计和 2000 多款产品的购买行为，我们发现负面信息与正面反馈相结合可以产生净正面结果。收到更多两极分化的售前反馈的产品会产生更多的社区讨论、社区销售和后续购买。研究结果表明，在消费者对某件商品的偏好不断发展的情况下，遇到相对负面的信息实际上会对评论和购买行为产生积极的影响——比不存在相反信息的情况更是如此。收到更多两极分化的售前反馈的产品会产生更多的社区讨论、社区销售和后续购买。研究结果表明，在消费者对某件商品的偏好不断发展的情况下，遇到相对负面的信息实际上会对评论和购买行为产生积极的影响——比不存在相反信息的情况更是如此。收到更多两极分化的售前反馈的产品会产生更多的社区讨论、社区销售和后续购买。研究结果表明，在消费者对某件商品的偏好不断发展的情况下，遇到相对负面的信息实际上会对评论和购买行为产生积极的影响——比不存在相反信息的情况更是如此。