Research And Development On Full Business Points System And Precision Marketing Technology Based On Big Data

Under the overall goal of implementing state-owned enterprise reforms to “continually enhance the vitality,control and influence of the state-owned economy”,telecom operators actively explore market-oriented reforms,carry out “small contracting”,enhance incentives for front-line employees,and stimulate front-line employees.Customers provide better and better services and achieve a win-win situation for employees,businesses and society.At present,only KPI is managed,which leads the frontline employees to work passively in order to complete the KPI,and the development of the inventory customer and the new users cannot be well coordinated.At the same time,the opacity of employee incentives is not timely,and the enthusiasm of employees cannot be promptly motivated.In terms of marketing tools and means,front-line account managers lack the tools of precision marketing,adopting human-sea tactics to carry out the marketing of customers,which is time-consuming and laborious,and forms the current high cost and low efficiency of marketing.In order to solve the above problems existing by operators,operators have started the construction of the points system.The main content and functions of the construction include four subsystems: system management subsystem,integration subsystem,business opportunity marketing subsystem and data ETL subsystem.The system management subsystem mainly provides the basic functions of the system,provides the decentralized and sub-domain authority control for the system,and ensures the security of the system and data.The data ETL subsystem mainly implements data collection,conversion and loading based on Kettle to realize multiple heterogeneous System data access and integration,and monitoring the data collection process to ensure the timeliness,integrity and stability of data collection;the integration subsystem uses data warehouse technology,establishes the integral rule model,and supports the integral calculation rules flexible and configurable Quickly respond to the adjustment of marketing policies,realize and support the multi-dimensional visual analysis and exploration of operators' full-service points,revenues and developments at the company level;the business-marketing subsystem adopts big data technology,based on the original features.To enrich and expand customer characteristics,provide a complete and vivid customer portrait for front-line sales personnel,and use the collaborative filtering calculation model based on customer segmentation to achieve personalized business recommendation,and through the closed-loop management and control of the marketing order,the information feedback and the calculation model of the man-machine cooperation integration are continuously optimized.This thesis uses the techniques of big data,data warehouse,machine learning and visualization technology to design the integral rule model,data warehouse three-layer model and customer portrait feature system,and realizes full service points calculation and responsible field calculation.At the same time,this paper proposes a random forest-collaborative recommendation algorithm based on customer segmentation.The algorithm has been proved to be more efficient than the common collaborative filtering algorithm,and achieves the highest accuracy with a smaller K value,thereby reducing the amount of computation and improving the computational performance.The main contributions of the thesis include: 1)Based on the responsibility field,the points are applied to the performance evaluation of the account manager within the operator,coordinating the coordinated business development and stock management,effectively supporting the operator's market operation management;2)in terms of computing mode Fully combine the advantages of customer portrait features and customer ordering behaviors,“moving” and “quiet” to extract the calculation model of collaborative filtering based on customer segmentation;3)using machine learning,statistics and logical reasoning on customer images Means,derivative of customer characteristics;4)In system engineering,make full use of offline marketing capabilities of operators,business opportunity calculation as a part of precision marketing closed-loop control,realizing timely feedback and input of external signals,conducive to timely adjustment and Improve the characteristics of the learning model.