The Wholesale Customer Analysis System For Petroleum Company

Effective customer relationship management is of great significance for companies to develop the sales market,improve profits and work efficiency.Most of the energy sales companies are state-owned enterprises.They have unique market sales characteristics because of their commitment to national strategic tasks.And there is no deep development in customer management.Due to the competition of a large number of private enterprises in recent years,it is urgently necessary to stabilize the market and increase profits through customer analysis.In particular,it is necessary to deeply analyze consumption behavioral data and accurate grasp of customer value for wholesale sales customers with large differences,strong volatility and obvious seasonal characteristics.It helps to provide decision support for wholesale product pricing,promotion and customer service.This paper addresses the marketing needs of petroleum wholesale sales customers.The core functions of wholesale customer analysis are designed based on the characteristics of the industry,user trading attributes and external influence factors.The main tasks include:establishing an evaluation metrics for oil wholesale sales customer transaction characteristics.On the basis of traditional metrics,we propose targeted metrics such as trading activity,trading distribution uniformity,trading quantity stability and average profit and loss.It fully characterizes the different granular trading features of wholesale sales customers.Based on the characteristic importance learning of small batches of expert mark samples,the value segmentation for long-tail characteristic group customers and the group customer evolution analysis for dynamic consumption behaviors are performed.We build an individual balue tag learning model for specific dimensions.It can not only reflect customers relative attributes in the group but also identify dynamically evolving consumer behavior.And we also construct a customer arrival time and purchase forecast model,as well as trading frequency forecasting and customer loss early warning models.Finally,we designed and implemented a customer value analysis system for petroleum wholesale sales based on the big data analysis platform of a provincial oil sales company.It includes such functions as evaluation metrics management,group customer management,individual customer management,marketing strategy management and user management.The system is based on the Oracle database,uses Java to implement data processing and analysis functions,and utilizes React technology and E-charts technology to build a visualization platform.There are 150 users which includes provincial and municipal company managers,office staff and client managers.The system analyzes nearly 1.53 million consumer behavior data of 6573 wholesale customers from 2015 to 2017.And it effectively assists enterprise users in formulating relevant marketing strategies and evaluating marketing results.