Research on social e-commerce reputation formation and state-introduced model

Purpose - This paper aims to solve the problem of information overload and reduce search costs. It proposes a social e-commerce online reputation formation model and community state-introduced model. A system dynamics trend simulation has been run to capture the relationship among the sellers, buyers, social ecommerce platforms and external environment to obtain an online reputation. Design/methodology/approach - Empirical research relating to social e-commerce reputation has been used to confirm the influencing factors in social e-commerce, and a conceptual framework is developed for social e-commerce reputation formation. Thereafter, a trend simulation is generated to classify the relationship among the factors based on system dynamics. Also, the improved algorithm for community detection and a state-introduced model based on a Markov network are proposed to achieve better network partition for better online reputation management. Findings - The empirical model captures the interaction effect of social e-commerce reputation and the state-introduced model to guide community public opinion and improve the efficiency of social e-commerce reputation formation. This helps minimize searching cost thereby improving social e-commerce reputation construction and management. Research limitations/implications - There is no appropriate online reputation system to be constructed to test the relationship proposed in the study for a field experiment. Also, deeper investigation for the nodes' attributes in social networks should be made in future research. Besides, researchers are advised to explore measurement for the reputation of a given seller by using social media data as from Twitter or micro blogs. Originality/value - Investigations that study online reputation in the social e-commerce are limited. The empirical research figured out the factors which can influence the formation of online reputation in social e-commerce. An SD model was proposed to explain the factors interaction and trend simulation was run. Also, a state-introduced model was proposed to highlight the effect of nodes' attributes on communities' detection to give a deeper investigation for the online reputation management.