广东地区鞋类清洗消费者基于RFM模型的客户分级评价数据

采集广东地区鞋类清洗消费者最近半年的消费记录，使用RFM客户价值模型分析方法，对用户最近一次消费活动时间（Recency）、消费活动频率（Frequency）和消费金额（Monetary）三个维度进行评分，识别不同价值的客户群体。通过对消费者进行分级管理，满足不同等级消费者的个性化需求，并为同行业企业管理不同等级的消费者，实现精准个性化服务提供数据支持。1、数据处理：对从广东地区采集到的鞋类清洗消费者数据进行脱敏、降噪、清洗、聚集、分析。
2、数据加工：RFM模型通过计算客户最近一次活动(R)天数、活动频率(F)次数和消费总金额(M)这三个维度来评估用户价值。对于R维度,根据客户最近一次活动R（天数）,划分为5个等级: 0≤R≤30得分为5分，30＜R≤60得分为4分，60＜R≤90得分为3分，90＜R≤120得分为2分，D＞120得分为1分；对于F维度,根据用户在最近半年内活动频率F（次数）,划分为5个等级: F≥8得分为5分、6≤F≤7得分为4分、4≤F≤5得分为3分、2≤F≤3得分为2分、0≤F≤1得分为1分；对于M维度,根据用户在最近半年内的消费总金额（M）,划分为5个等级，M≥800得分为5分，600≤M＜800得分为4分，400≤M＜600得分为3分，200≤M＜400得分为2分，M＜200得分为1分。 RFM得分=0.3*（R得分）+0.3*（F得分）+0.4*（M得分） ，再根据RFM综合评分对客户进行分类，RFM ≥4为高粘度客户，3≤RFM＜4为重要维系客户，2≤RFM＜3 为潜力深耕客户，RFM＜2为一般发展客户。
3、数据使用：基于消费频次、消费金额等不同维度获得的聚类分组成果，对聚类分组数量和分组阀值、以及维度权重进行人为干预，使客户分类趋于合理。通过对消费者进行分级管理，满足不同等级消费者的个性化需求，并为同行业企业管理不同等级的消费者，实现精准个性化服务提供数据支持。

This dataset collects consumer spending records of shoe cleaning services in Guangdong Province over the past six months. Using the RFM customer value model, we score customers across three dimensions: Recency (time since last consumption), Frequency (consumption frequency), and Monetary (total consumption amount), to identify customer groups with different value levels. This work provides data support for enterprises in the same industry to manage customers at different tiers and deliver precise personalized services by implementing graded customer management to meet the personalized demands of customers at various levels.

1. Data Processing: Anonymize, denoise, clean, aggregate and analyze the collected shoe cleaning consumer data from Guangdong Province.

2. Data Modeling & Processing: The RFM model evaluates user value by calculating three core dimensions over the past six months: days since the customer’s last activity (R), number of consumption activities (F), and total consumption amount (M).
- For the R dimension: Customers are divided into 5 tiers based on R values: 5 points for 0≤R≤30, 4 points for 30<R≤60, 3 points for 60<R≤90, 2 points for 90<R≤120, and 1 point for R>120.
- For the F dimension: Customers are divided into 5 tiers based on F values: 5 points for F≥8, 4 points for 6≤F≤7, 3 points for 4≤F≤5, 2 points for 2≤F≤3, and 1 point for 0≤F≤1.
- For the M dimension: Customers are divided into 5 tiers based on M values: 5 points for M≥800, 4 points for 600≤M<800, 3 points for 400≤M<600, 2 points for 200≤M<400, and 1 point for M<200.
The comprehensive RFM score is calculated as: RFM Score = 0.3*(R score) + 0.3*(F score) + 0.4*(M score). Customers are then classified based on their comprehensive RFM scores:
- High-viscosity customers: RFM ≥4
- Important retention customers: 3≤RFM<4
- Potential in-depth development customers: 2≤RFM<3
- General development customers: RFM<2

3. Data Application: Based on the clustering results derived from dimensions such as consumption frequency and total consumption amount, manual adjustments are made to the number of clustering groups, grouping thresholds and dimension weights to optimize the rationality of customer classification. By implementing graded customer management to meet the personalized demands of customers at various tiers, this dataset provides data support for enterprises in the same industry to manage customers at different levels and deliver precise personalized services.