The paper addresses an opinion mining problem: how to find the helpful reviews from online consumer reviews via the quality of the content. Since there are too many reviews, efficiently identifying the helpful ones earlier can benefit both consumers and companies. Consumers can read only the helpful opinions from helpful reviews before they purchase a product, while companies can acquire the true reasons a product is liked or hated. A system is built to assess the difficulty of the problem. The experimental results show that helpful reviews can be distinguished from unhelpful ones with high precision.