Infomediaries, such as movie critics, often evaluate product quality using discrete grades, such as thumbs-up and thumbs-down. Their reports are used by potential consumers to decide whether to purchase the product. Our aim is to examine how the grading can be improved by answering research questions such as: What is the quality cutoff for each grade? How many grades should the grading scale use? How should grading be done for other goals than informativeness, such as maximizing reviewer’s traffic? How should grading be done to encourage producers to improve quality? Using an analytical model, we show that optimal grade cutoffs depend on the distributions of product quality and consumer taste, and the resulting scale may have unequal intervals. As few as five grades can be near-optimal for consumer welfare and grades should be made more difficult to achieve when producers can invest in quality improvement. These results should help improve the design of scales for reporting quality.

电影评论家等信息媒体通常使用不连续的等级来评估产品质量，例如赞成和反对。他们的报告被潜在消费者用来决定是否购买该产品。我们的目标是检查如何通过回答研究问题来改进评分，例如：每个等级的质量临界值是多少？评分标准应该使用多少个等级？除了信息量之外，应该如何为其他目标进行评分，例如最大化评论者的流量？应该如何进行分级来鼓励生产者提高质量？使用分析模型，我们表明最佳等级截止值取决于产品质量和消费者口味的分布，并且由此产生的尺度可能具有不相等的区间。低至五个等级对于消费者福利来说可能接近最佳，并且当生产商可以投资于质量改进时，等级应该更难实现。这些结果应该有助于改进报告质量量表的设计。