Research On Promption Strategies Of Short-Term Repeat Products Based On Social Network

This paper analyzes the products diffusing speed and the benefits of three marketing strategies, seeding to revenue leaders(high CLV), seeding to opinion leaders(high Node Degree), and seeding to random customers, with methods of agent-based model of a seeding program in the social networks, in order to achieve goals in the context.Based on reviewing literatures, this paper is to solve some problems as followings:(1) Choosing the structure of social networks, meeting the purpose of the research.This paper chooses Jackson-Rogers algorithm to set up social networks which is controlled by two parameters, CLV assortativity coefficient and network clustering coefficient. The fundamentals of Jackson-Rogers algorithm are same with Small-Word models and Scale-Free models, but its generative rules are different. The purpose of this research is to get social networks with characters of revenue leaders, opinion leaders, and random customers. CLV assortativity is the Pearson correlation between the CLV of a customer and the CLV of the others to whom that customer is connected. It can representative the wealth concentrations of groups. Network clustering coefficient representatives the cluster degree and scales of groups. Meeting the two parameters is close to the characters of this paper’s social networks.(2) Choosing the rationality of computing methods of CLV.After reviewing literatures, this paper introduces the customer life cycle to calculating CLV. Customer buying behaviors include four phases, exploration, buildup, stability and decline. The customer life cycle is made up of the four phases. This paper chooses this method in order to make customers buying behavior more real.(3) Choosing the diffusion model.This paper analyzes an agent-based model of a seeding program and contrasts the analysis results of using different seeding strategies, such as, revenue leader seeding, opinion leader seeding and random seeding.After finishing the simulation for different strategies, this research gets some conclusions and management inspirations. The first is that clustering effect can promote the product diffusion. Companies should make more concentrate on circle marketing and maintaining customers in these circles. The second is, higher the concentration of wealth is, higher the diffusion rate of products is. Furthermore, lower the standard deviation of customer values is, higher the profits of companies are. That means companies should take more resources into getting groups with high customer values. The last conclusion is, the cost of getting revenue leaders is higher than the others’. But the net profit of revenue leader seeding is higher than others’. So, companies should focus on revenue leader seeding.