An empirical research of customers' transaction evaluation system at C2C e-commerce website

Generally, a service provider only gets negative assessments from his/her customers due to bad services. However, this is not always the case. Recently, a new type of e-commerce swindle called “malicious rating attack” has disordered the evaluation system of Taobao, which offers the largest virtual platform in mainland. Different from earlier fraud cases aiming shoppers, this time “profitable” owners of online stores fall victims to the tricky cheat. For fear that negative assessments from malicious raters may lead to customer drain, most poor retailers can do nothing but be blackmailed. Therefore the paper tries to design a new assessment mechanism to prevent online retailers from such swindles and to normalize the operation of online business. In the modified evaluation system, online shoppers' credits, which can be measured by the length of register time and the number of transactions, is included. Since most malicious raters change their accounts frequently in case of detected by network surveillance, they cannot be customers of high credibility. And thus, to launch an attack in the evaluation system considering the weight of credit will be extremely hard and costly. As the sensitivity analysis in the paper turns out that it will call for more than 60 times the size of the former gang to realize the same attack effect. It is believed that such obstacles for malicious raters to making profit will eventually bring an end to the “treat or trick” game.