南京市地区购买媒体投流客户价值评估数据

采集erp系统中南京市地区的数据，通过客户最近一次消费时间的天数R、消费频次F和消费金额M（元）， 采用 RFM 模型对客户进行价值评级，实现精准化运营。通过对南京市地区客户价值管理，满足不同价值客户的个性化需求。对A级客户，每个月进行一次回访维护；对B级客户，每个季度进行一次回访维护；对C级客户，每半年进行一次回访维护；对D级客户，每年进行一次回访维护。另外可以为本地区客户群体高度重叠企业提供不同价值类型的客户个性化服务的数据支持。1、数据处理：对从erp系统中采集到的数据进行脱敏、降噪、清洗、聚集、分析。2、数据加工：运用RFM模型结合客户最近一次消费时间的天数R、消费频次F和消费金额M（元）的得分排名对客户进行一个综合排名，最终得出一个RFM综合分。3、规则评分：a.提取出距离最近一次消费时间的天数R、消费频次F和消费金额M进行分类，距离最近一次消费时间的天数R间隔最短的客户排在最上面。按照从1-5评分，前20%的客户获得5分，接下来的20%用户获得4分,再下来20%的客户为3分，再下来20%的客户为2分，最后20%的客户为1分。b.根据客户消费频次F从高到底依次对用户进行分类，前20%的客户在用户活动频率的分数为5，以此类推。C,根据客户消费金额M（元），前20%的客户在消费金额的分数为5，以此类推。消费金额最少的20%客户则分数为1。4、综合评分计算：RFM综合分=0.3*（R评分）+0.3*（F评分）+0.4*（M评分），评分大于等于4分的客户等级为A，大于等于3小于4的客户等级为B，大于等于2小于3的客户等级为C，低于2的客户等级为D。

This dataset collects data from the ERP system for the Nanjing region. It adopts the RFM model based on Recency (R: days since a customer's last purchase), Frequency (F) and Monetary value (M, yuan) to grade customer value, enabling precise operational management. Through customer value management in the Nanjing region, personalized demands of customers with different value tiers are met. For Class A customers, follow-up visits and maintenance are conducted once a month; for Class B customers, once every quarter; for Class C customers, once every six months; and for Class D customers, once a year. Additionally, this dataset can provide data support for personalized services targeting customers of different value tiers for enterprises with highly overlapping customer groups in this local region.

1. Data Processing: Perform desensitization, noise reduction, cleaning, aggregation and analysis on the data collected from the ERP system.
2. Comprehensive Ranking via RFM Model: Use the RFM model combined with the score rankings of Recency (R), Frequency (F) and Monetary value (M, yuan) to conduct a comprehensive ranking of customers, and finally derive an overall RFM score.
3. Rule-based Scoring:
a. Extract Recency (R: days since last purchase), Frequency (F) and Monetary value (M) for classification. Customers with the shortest R interval are ranked highest. Scoring is conducted from 1 to 5: the top 20% of customers receive 5 points, the next 20% receive 4 points, the subsequent 20% receive 3 points, the next 20% receive 2 points, and the last 20% receive 1 point.
b. Classify customers in descending order based on their consumption frequency F. The top 20% of customers get 5 points for their activity frequency, and the remaining groups follow the same grading rule.
c. For customer consumption amount M (yuan), the top 20% of customers receive 5 points for their consumption amount, and the grading follows the same pattern as above. The 20% of customers with the lowest consumption amount get 1 point.
4. Comprehensive Score Calculation: The overall RFM score = 0.3*(R score) + 0.3*(F score) + 0.4*(M score). Customer tiers are defined as follows: Class A for scores ≥4, Class B for 3 ≤ score < 4, Class C for 2 ≤ score < 3, and Class D for scores < 2.