Who should pay for online reviews? Design of an online user feedback mechanism

Highlights

• Modeling the influence of online reviews on consumers’ perception of sellers’ reputations.

• Analyzing the ineffectiveness of reputation system when sellers offer rebates.

• Illustrating feasibility of e-commerce platforms encouraging honest online reviews.

Abstract

Online reputation systems provide consumers important references before their purchase decisions. So designing a mechanism to encourage consumers to leave honest online reviews becomes very important for e-commerce platforms. We establish a Bayesian model to simulate the formation of consumers’ perceptions of sellers’ reputations in a C2C e-commerce platform. We find that both truthfulness of reviews and number of reviews influence consumers’ perceptions of sellers’ reputation, and they are mutually substitutable. Consumers may have no faith in the truthfulness of the reviews if sellers offer rebates for more online reviews. To obtain honest reputation information, the platform should encourage consumers to provide honest opinions about experiences and feelings consistent with of social-exchange theory. In addition, to obtain a certain level of perceived reputation, the system does not need all consumers to submit their opinions. We also provide upper and lower bounds for rebates offered by the platform.

Introduction

Consumer trust is much more important in online transactions than it is in real-world transactions due to the serious information asymmetry in a cyber-environment. However, the trust-fraud problem has long been a large issue in online C2C markets, and buyers face great risk in these transactions. With continuous improvement in the trading environment, the main risk is consumers’ uncertainty about the quality of goods (Westland, 2002, Zhu and Zhang, 2010). Studies have shown that online reputation systems (user-feedback mechanisms) can effectively solve this problem (Dellarocas, 2003, Lin et al., 2015, Resnick et al., 2000). Reputation systems have been widely used in many e-commerce platforms, such as Taobao(Taobao.com), eBay, Amazon, and others. Such a system collects the transaction and user-feedback information to help consumers make their purchase decisions. It is believed that more than 70% of consumers refer to other buyers’ comments before making purchases (Hu et al., 2014).

Taobao, the largest C2C online market in China (with more than a 90% market share), captivates researchers and practitioners. The reputation system used by Taobao is similar to the eBay reputation system, but it faces a real challenge. Taobao faces three dilemmas: limited consumer participation, seller review manipulation, and consumer positive bias. First, online reviews can help potential consumers make purchase decisions. However, few consumers leave feedbacks after completing the deal; after a certain period of time, the system generates positive reviews by default.

Second, online reviews, as a type of word of mouth, are very important for sellers (Chen and Xie, 2008, Kwark et al., 2014). Therefore, some sellers commit trust fraud to boost their reputation and attract more consumers (Hendrikx et al., 2015; Xiong and Zhong, 2012). For example, sellers employ professional scammers to artificially increase their reputation scores through fake transactions, or sellers entice consumers to generate positive reviews by offering rebates (Xu et al., 2015).

Third, consumers rarely submit negative reviews because they fear sellers might repeatedly pester them by telephone to soften a negative review (Li et al., 2013; Zhang et al., 2005). Consumers know about these types of system abuse, so they are leery of the entire feedback system. User-feedback information is the input of the reputation system, and it determines whether the output of the system (calculated from the reputation model) is credible. Therefore, only by solving these problems can the reputation system work effectively.

Many researchers have tried to solve these problems. Some have concluded that to receive more online reviews, some type of incentive strategy should be implemented(Jurca and Faltings, 2003, Zhao et al., 2012). However, what is the effect of this strategy in practice? The intuition of this paper comes from the reality of the effect of this incentive strategy: A Taobao consumer sometimes receives a request for positive reports from the seller who offers a 5- or 10-yuan cash rebate. If the customer is not satisfied with the product, but it is not worth returning the product due to the freight costs, then the customer may leave a positive review to receive the 5- or 10-yuan rebate.

In contrast, the Meituan (Meituan.com) platform solicits the customer’s opinion in exchange for reward points. If a consumer is not satisfied with the provided service, the seller may receive no more than three points out of five. At the same time, the customer receives reward points from Meituan. Therefore, we can see that in reality both seller and platform solicit online reviews. However, they have distinctly different objectives. A seller, honest or dishonest, wants to increase the reputation score to attract more consumers. Therefore, a dishonest seller has an incentive to only provide rebates for positive reports and may retaliate for negative reports. A platform, however, needs to create an honest trading environment for long-term viability, so it should ensure that subsequent consumers can successfully distinguish good sellers in the online market.

Sometimes sellers’ manipulations are aligned with the platform’s interest (honest sellers pay to promote their products through online reviews), but other manipulations are not (dishonest sellers pretended to be honest, but buy only positive reviews).

In this paper, we propose that the e-commerce platform should provide the incentive to get honest online reviews. To reach this judgment, we established a Bayesian model to explain the process of perceived-reputation formation and came to the following conclusions. First, the truthfulness of online reviews will positively influence consumers’ perceptions of the seller’s reputation. Second, if the truthfulness of a review is given, consumers’ perceptions of the seller’s reputation can be increased by addition positive reviews. So, some incentives for more reviews are necessary. After comparing the analyses, we find that consumers are more likely to submit honest reviews when rebates are offered by the platform than when rebates are offered by sellers. In addition, truthfulness and the number of online reviews are substitutable. Therefore, to reach a certain level of perceived reputation, it is not necessary to offer all consumers rebates, and this is a cost saving for the incentives. As two-sided markets, e-commerce platforms are responsible for providing an honest trading environment. Sellers often pay per-transaction fees for a platform’s services, so a fraction of these fees can be used to pay incentives. We give upper and lower bounds for these rebates.

The rest of the paper is organized as follows. In the next section, we will discuss the related literature. Then, we create a Bayesian model to explain how online reviews affect consumers’ perceptions of sellers’ reputations. In Section 4, we compare the effectiveness of the reputation system when online reviews are encouraged by different subject and propose that the e-commerce platform should encourage consumers to provide online reviews. Section 5 discusses some extensions of the model and the feasibility of this rebate mechanism, and Section 6 concludes.