| It has been nearly 20 years since the emergence of E-commerce,E-commerce has broken the original way of commodity transaction and changed from offline one-to-one cash transaction to online transaction on the Internet.People only need to make simple click operation through mobile phones and computers to complete the purchase of commodities they like.However,the consumer groups of traditional E-commerce platforms are often ordinary users who buy daily necessities.However,B2B(business-to-business Business activities)E-commerce platforms are rare.In 2019,the transaction amount of B2 B FMCG industry has reached nearly 200 billion yuan,and such a huge consumer group is fundamentally different from ordinary E-commerce users.Due to the characteristics of the FMCG industry,such as fast flow,high inventory backlog cost and fixed purchase of commodities,the corresponding E-commerce platforms need to make functional changes according to the characteristics of users.Traditional E-commerce platforms have been unable to meet the needs of such users.According to the real B2 B FMCG E-commerce project and the pain points of the existing FMCG E-commerce platform,the author designed and implemented a B2 B business system based on recommendation algorithm.This paper first describes the status of B2 B FMCG industry,the content of the research paper,the organizational structure of the paper and the related technologies used in the paper,then analyzes the function of the overall e-commerce system.Firstly,the system is divided into front-end requirements,server-side requirements and recommendation related requirements.Then the author describes feasibility analysis and detailed design of the functional points of requirements.The B2 B FMCG industry is mainly targeted at small store owners.In order to prevent public opinion and harm the interests of brand manufacturers,the opacity of commodity prices needs to be guaranteed,so users' store audit process is added outside the login process.In addition,in view of the characteristics of FMCG(such as fixed purchase frequency and fast flow),this paper designed special promotion methods to meet the balance of interests between manufacturers and small shop owners.At the same time,in order to ensure the normal operation of the client side,the server side shall add the related requirements of the CMS(Content Management System)to guarantee the client side's requirements through functions such as commodity center,order setting and promotion configuration.Firstly,the user's operation behavior is elaborated by means of flow chart,use case diagram,page screenshot,etc.,so as to realize all functions of registration,login,audit,browse,add car,order and pay.In addition,the original recommendation algorithm is partially modified,real-time recommendation is introduced and offline recommendation types are enriched,which greatly improved the order conversion rate of users.Finally,a full-process regression test is performed on the entire e-commerce system.From the functional test,the user's order process is verified for problems.From the performance test,the critical value of the system interface is tested.The overall test results are in line with the expected functions and Non-functional requirements.This system has been implemented in the author's internship company,but there are still some needs to be improved,such as simple page level,difficult to modify product information.The optimization of the recommendation module brings the overall CVR(Conversion Rate)close to 3 percentage points.The effect is very significant.The viewpoints and methods proposed in this paper and the final results obtained have certain reference value for the development of B2 B E-commerce platform. |
