舟山地区公司平台客户价值评估数据

采集销售记录表中舟山地区的数据，通过客户在2018年4月1日距离2025年4月30日间隔的最近一次消费时间天数（R）、客户在2018年4月1日至2025年4月30日之间消费频次(F)和客户在2018年4月1日至2025年4月30日之间消费金额(M)（单位：元）， 采用 RFM 模型对客户进行价值评级，实现精准化运营，通过对舟山地区客户价值管理，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。数据处理：对从销售记录表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模型结合客户在2018年4月1日距离2025年4月30日间隔的最近一次消费时间天数（R）、客户在2018年4月1日至2025年4月30日之间消费频次(F)和客户在2018年4月1日至2025年4月30日之间消费金额(M)（单位：元）的得分排名对客户进行一个综合排名，最终得出一个RFM总评分。a.提取出最近一次消费时间距离当前分析时间的天数（R）、客户在2018年4月1日至2025年4月30日之间消费频次(F)和客户在2018年4月1日至2025年4月30日之间消费金额(M)（单位：元）进行分类，最近一次消费时间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.根据客户在2018年4月1日至2025年4月30日消费频次(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, 根据客户在2018年4月1日至2025年4月30日消费金额(M)，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。 RFM得分=0.3*（R得分）+0.3*（F得分）+0.4*（M得分） 评分大于等于4分的为A级客户，大于等于3小于4的为B级客户，大于等于2小于3的为C 级客户，低于2的为D 级客户

Data was extracted from the sales record table for the Zhoushan region. Three metrics were calculated for each customer: R, the number of days between the customer's most recent purchase date and April 30, 2025 (the end of the analysis period spanning from April 1, 2018 to April 30, 2025); F, the customer's total purchase frequency during this period; and M, the customer's total consumption amount (unit: Yuan) during this period. The RFM model was adopted to conduct customer value rating, enabling precise operational management, and meeting the personalized needs of customers across different value tiers through targeted customer value management in the Zhoushan region. For Level A customers, monthly return visits for maintenance will be conducted; Level B customers will receive quarterly return visits; Level C customers will be visited semi-annually; and Level D customers will be visited annually. Additionally, this dataset can provide data support for local enterprises with highly overlapping customer groups to deliver personalized services tailored to customers of different value types.

Data Processing: The raw data collected from the sales record table underwent desensitization, noise reduction, cleaning, aggregation, and preliminary analysis.

2. Data Refinement: A comprehensive customer ranking was generated using the RFM model, based on the score rankings of the three metrics above, to derive the final overall RFM total score.
a. Classification and scoring of R: Extract the values of R, F, and M for classification. Customers with the shortest interval since their most recent purchase are ranked highest. A 1-5 scoring scale is adopted: the top 20% of customers receive a score of 5, the next 20% receive 4, the subsequent 20% receive 3, the next 20% receive 2, and the bottom 20% receive 1.
b. Classification and scoring of F: Rank customers by their purchase frequency during the analysis period from highest to lowest. The top 20% of customers are assigned a score of 5, with the remaining tiers following the same logic.
c. Classification and scoring of M: Rank customers by their total consumption amount during the analysis period from highest to lowest. The top 20% receive a score of 5, while the bottom 20% (those with the lowest consumption) receive a score of 1.
The overall RFM score is calculated as: RFM Score = 0.3 * (R Score) + 0.3 * (F Score) + 0.4 * (M Score)
Customers are segmented into four tiers based on their overall RFM score: Level A (score ≥4), Level B (3 ≤ score <4), Level C (2 ≤ score <3), and Level D (score <2).

搜集汇总

背景与挑战

背景概述

该数据集为舟山地区公司平台客户价值评估数据，包含550条企业数据，采用RFM模型（基于最近消费天数、消费频次和消费金额）对客户进行价值评级，分为A、B、C、D四个等级，用于支持精准化运营和个性化客户服务。