关键词: 推荐奖励计划, 社交推荐, 网络结构, 奖励模式, 奖励分配

Abstract: With the popularity of social media such as Micro-Blog and WeChat, compared with traditional advertising, modern retailers tend to rely on the social referral of existing customers to obtain new ones. The referral reward program is a marketing strategy, by which retailers pay material or non-material rewards to existing customers to stimulate them to send referrals to new customers. Compared with traditional advertising, the program is more targeted and controllable and has lower payment costs. It has become one of the most commonly used marketing models for retailers. When retailers use the program in social networks for marketing, to achieve optimal performance, it is necessary to make unified decisions on reward mode, reward distribution and reward degree on the basis of making full use of social networks. The differentiated social network structure, the complex combination of reward modes with reward distribution types easily lead to retailers' decision-making problems. Optimizing the reward mode, reward distribution and reward degree, fully tapping the potential of social network information dissemination, and designing an efficient referral reward program have become the first problem for retailers when they adopt the program to do marketing. However, we have not been able to find a clear answer from existing research.
In this paper, the interaction among the social utility, referral cost and social network structure is introduced into the program. The dynamic Stackelberg game model of the retailer, referrer and receiver is constructed to analyze whether the retailer adopts the program and how he optimizes it under the consumers' social referral behavior. Specifically, in this paper, the social network structure, social utility and referral cost are introduced into our model construction, and with the social network structure as the clue, the modes with threshold reward and without it are related. Then, considering the consumer behavior and retailer decision-making in the social network, the consumer referral utility and retailers' revenue management models after purchasing products under different referral reward programs are constructed in this paper. Finally, we make a comparative static analysis of the retailer's optimal decision under different reward modes to explore the reward design (the reward mode, the reward distribution and the reward degree) of the retailer's optimal referral reward program.
This paper mainly draws the following conclusions different from previous studies. First, when the social utility is high and the referral cost and price are low, the retailer will not need to use the program. Instead, retailers need to optimize it (the reward mode, the reward distribution and the reward degree) based on factors such as social network structure. Second, (1)when the connectivity of social network is high, the reward threshold will have the inducement effect of promoting the referral of existing consumers. Thus, the performance of the retailer with the threshold mode will be better than that without it when the connectivity of social network is high. (2)The reward mode adopted by retailers has a significant impact on the reward distribution and reward degree setting. The retailer only needs to reward the recommender with the threshold reward mode, while he needs to allocate the reward according to the social utility, referral cost and price without it. The reason is that the inducement effect with the threshold reward mode is higher than the economic incentive effect of direct reward to the recipient. In addition, with the threshold reward mode, the retailer's reward degree decreases with an increase in network connectivity, but remains unchanged without it.
This study hopes to expand and enrich the research on the consumer social referral and referral reward program marketing model, and provide a more comprehensive theoretical guidance for retailers' referral reward program design practice.

Key words: referral-reward model, social referral, network structure, reward mode, reward distribution