The Research On Problems And Countermeasures Of B2C Enterprises Service Quality Based On Electronic Word-of-mouth

In social e-commerce environment, with the ever-changing ways of generatinginformation and media, the Internet-based interaction, communication and contact betweenusers and enterprises become extremely widespread and frequent. The relationship betweenusers and businesses tending to equal, interactive and influence each other, which become themost important features of e-commerce. With the rise of a new generation of social media,Micro-blog, the electronic word-of-mouth(eWOM) produced through interaction,communication and contact between customers and enterprises in Micro-blog platform, hasalso become an important source of gaining access to the data for enterprises, which maycontain a huge commercial value. At the same time, with the continuous development of B2Ce-commerce, the subsequent service problems become more and more attention. It is of greatsignificance for the sustainable development of enterprise to improve service quality throughrich eWOM data generated in social media.This paper studies improvement strategy of B2C enterprise service quality based oneWOM. In order to understand objectively and design services, we choose ServiceBlueprinting as the main method to analyze the process of B2C e-commerce service. And thenthe B2C e-commerce services can be divided into six sub-services, including decision-makingservices, access service, transaction service, payment services, logistics services andafter-sales service.By building eWOM classification model, the accessed eWOM could be classifiedaccording to six sub-services, thus helping enterprises to efficiently access eWOM withineach sub-service. The survey takes B2C e-commerce enterprise dangdang as respondent,3,260service-related eWOM via Sina Micro-blog are collected. Then we labels the eWOMand selects the features of different sub-service with CHI approach. Support VectorMachine(SVM) and Navie Bayes(NB) are used to get the classification. The experimentshows that classification builded by SVM has a better result with82.09%accuracy rate.Sentiment analysis module of ROST CM is used to determine the negative eWOM ineach sub-service. It is found that the number of negative eWOM in logistics and after-sales sub-services is more than others. Then some methods of data mining such as word frequencyanalysis are used to identify the major problems of service quality within these twosub-services. Finally, some suggestions and countermeasures are provided for dangdang toimprove service quality by using Service Blueprinting.