Research On The Effect Of Recommendation System In Sharing Platform On User Behavior

In recent years,with the rapid development of the sharing economy,it is gradually penetrating into the life of consumers.In the sharing economy,the suppliers are generally defined as individuals with idle resources.Their idle resources are limited.So in the face of huge demand,they may give up providing services and exit the platform.This will have a significant impact on the operation of the platform.In order to solve this problem,the general solution of sharing platforms is to introduce personalized recommendation to improve the matching efficiency of suppliers and users.In the sharing platform,personalized recommendation can help users to solve the problem of information overload,improve the efficiency of user decision-making,increase the continuous search and purchase behavior of users,so as to improve the platform revenue.However,recent studies have found that personalized recommendation has no significant positive impact on users' willingness to repeat purchase.Even after the introduction of personalized recommendation service,users have reduced their stay time on the platform,thus reducing the(potential)access opportunities to upstream service providers.From this,we begin to pay attention to the upstream and downstream two-way effectiveness of personalized recommendation technology on the sharing platform.In order to solve this problem,this paper obtained a large amount of users' behavior data of the actual sharing platform APP.The data contains users' behavior when the sharing platform before and after providing personalized recommendation.Also,this paper combined with related theories to build models,and mainly explored the following two problems.From the perspective of solving information overload,this paper explores whether personalized recommendation has a significant effect on the users' decisionmaking efficiency.From the perspective of a service to improve users' satisfaction,this paper explores whether it has a significant effect on the user's continuous using behavior.For the first issue that this paper focuses on,we used the DID method to study users' click log data.The results show that personalized recommendation has no significant impact on the users' decision-making efficiency.This paper explored the following two conjectures by using DID model.One is that the decision-making efficiency of users is related to the total time spent on the platform and the total number of clicks.Personalized recommendations may have an impact on these two behaviors.The other is that personalized recommendations may affect the user's decision-making efficiency by affecting the user's search diversity behavior.Therefore,the first question of this paper is answered.Personalized recommendation does not have a positive impact on the users' decision-making efficiency.Because it has a significant positive impact on the users' searching diversity,users are more inclined to search for stores which they have not searched before,which affects their decision-making efficiency.It is found that personalized recommendation improves the activity of the upstream in the sharing platform by affecting the increase of users' searching diverse behaviors,thereby ensuring the upstream diversity of the sharing platform.For the second problem,we used the DID model to study the users' click logs data and purchase data from two aspects: users' continuous searching behavior and continuous purchasing behavior.The results show that personalized recommendation has a significant positive impact on the users' continuous searching behavior and continuous purchasing behavior.This answers the second question of this paper,that is,personalized recommendation has a significant positive impact on the users' continuous using behavior,which shows that as a service,personalized recommendation can bring long-term benefits for the sharing platform.To quantify the improvement brought by personalized recommendations,it is found that personalized recommendation has a greater impact on users' continuous searching behavior than their continuous purchasing behavior.Therefore,we suggest that when measuring the impact of services on the platform,the platforms should not only pay attention to the indicators such as purchase volume,which generate direct benefits for the platform,but also pay attention to the indicators such as platform activity.Through the analysis of the experimental conclusions,it is found that personalized recommendation improves the downstream activity of the sharing platform by increasing the users' continuous using behavior on the platform.Combined with the characteristics of the sharing platform,cross-network externalities,it can be seen that the downstream activity can stimulate the upstream activity,which is very meaningful for the sharing platform.This paper mainly studies the effectiveness of personalized recommendation on the sharing platform.The study shows that personalized recommendation has a significant positive impact on the upstream and downstream of the platform.And it has brought a benign improvement to the customer value of the sharing platform.The study This study overturns the previous research conclusion that personalized recommendation can solve the problem of user information overload,and fills the gap in the research on the personalized recommendation effectiveness in the aspects of the sharing platform and users' objective behavior.In addition,the research method in this paper provides a more rigorous measurement method for the platform,which can help the platform to reduce the measurement bias caused by endogenous issues as much as possible when measuring the actual service effect.In general,this study can help the sharing platform to further clarify the significance and value of personalized recommendation services.