A Tobacco Retail Customer Value Hierarchy Classification Method Based on RFMC

Abstract

This invention provides a method for classifying and stratifying the value of tobacco retail customers based on RFMC (Retail Customer Value Management), comprising the following steps: S1, Data extraction and preprocessing: Extracting transaction data and credit data of retail customers within a preset period from the tobacco data center, cleaning missing and outlier values, and aggregating them by customer ID to form a modeling dataset; S2, Construction and quantitative assignment of four RFMC indicators: Based on the modeling dataset, constructing an RFMC model including the most recent purchase (R), purchase frequency (F), purchase amount (M), and credit rating (C), and assigning 0 / 1 binary values ​​to each indicator by setting thresholds; This invention breaks through the limitations of the original purely marketing-oriented evaluation by introducing the monopoly credit rating (C) indicator into the RFM model, deeply binding "trustworthiness / distrustworthiness" with "consumer value," and realizing data exchange and strategy synergy between marketing resource allocation and monopoly supervision.

Description

Technical Field

[0001] This invention relates to the fields of tobacco commercial marketing management and data mining technology, and in particular to a method for classifying and stratifying the value of tobacco retail customers based on RFMC. Background Technology

[0002] In the daily operations of tobacco commercial enterprises, retail customers are the core link between tobacco companies and consumers. To achieve optimal resource allocation and precise marketing, it is crucial to classify and stratify retail customers based on their value. Currently, tobacco commercial enterprises generally use the traditional RFM model (Recency, Frequency, Monetary) to conduct retail customer value analysis. These three marketing indicators are used to classify customers into different levels, thereby implementing differentiated supply and services.

[0003] However, existing customer classification methods based on the traditional RFM model have the following significant drawbacks: Traditional RFM models only cover transaction data in the marketing dimension, missing the tobacco industry's unique dimension of monopoly credit. The tobacco industry implements a strict monopoly system, and customer dishonesty (such as illegal procurement and resale) directly affects channel control. Simply relying on transaction data cannot achieve coordinated management of marketing services and monopoly supervision.

[0004] The 0 / 1 dichotomy of the three RFM indicators can only form 8 customer groups. Under the current demand for refined operation of tobacco with multiple brands, multiple specifications, and many segmented perspectives, the 8-category division seems too coarse, and the group characteristics overlap significantly.

[0005] Existing customer classification methods mostly rely on manually set rules or simple binary classification. The selection of thresholds depends heavily on the experience of business personnel, resulting in some groups having a very low proportion, lacking practical operational significance, and being unable to adapt to market changes.

[0006] Existing technologies typically only output the overall value of a customer across all product categories, failing to form a three-tiered system of "overall value - brand value - product specification value." This makes it impossible to answer the question of "what value a customer possesses in a specific brand or product specification," resulting in poor targeting of brand cultivation strategies.

[0007] Due to the imperfection of the classification system, high-value customers cannot receive preferential treatment in terms of resources, while dishonest customers with risks of violations are not subject to supervision, resulting in low operational efficiency. To address this, a value-stratified classification method for tobacco retail customers based on RFMC is proposed. Summary of the Invention

[0008] In view of this, embodiments of the present invention provide a tobacco retail customer value stratification and classification method based on RFMC to solve or alleviate the technical problems existing in the prior art, and at least provide a beneficial option.

[0009] The technical solution of this invention is implemented as follows: a method for classifying and stratifying the value of tobacco retail customers based on RFMC, comprising the following steps: S1. Data Extraction and Preprocessing: Extract transaction data and credit data of retail customers within a preset period from the tobacco data center, clean missing and outlier values, and aggregate them by customer ID to form a modeling dataset. S2. Construction and Quantification of RFMC Four Indicators: Based on the modeling dataset, an RFMC model is constructed that includes the most recent purchase R, purchase frequency F, purchase amount M, and credit rating C, and thresholds are set for 0 / 1 binarization. S3, RFMC Combination and Preliminary Classification: The results of the four indicator assignments are combined to form 16 initial customer groups. Empty groups without actual samples are eliminated to obtain the effective groups. S4. Determining the optimal cluster number based on the elbow method: using the sum of squared errors within clusters. To evaluate the indicators, draw With the number of clusters The K value at the inflection point of the changing curve is taken as the optimal number of clusters; S5. K-means clustering and customer segmentation: Based on the optimal number of clusters, the K-means algorithm is used to perform unsupervised clustering on the effective group. According to the characteristics of the cluster centers, customers are divided into five categories: important high-quality customers, ordinary potential customers, customers with slight churn, key attention customers, and general attention customers. S6. Three-tier value segmentation: Using the full category, single brand, and single product specification as statistical criteria, repeat the RFMC construction and clustering process from S1 to S5 to output the customer's comprehensive value segmentation, brand value segmentation, and product specification value segmentation results. S7. Differentiated Strategy Output: Based on the three-layer value stratification results, automatically match and output corresponding service, supply, terminal construction, financial support and exclusive sales supervision strategies.

[0010] In some embodiments, the quantization assignment rule in S2 is specifically as follows: The number of days between the last purchase and the statistical date is calculated and denoted as the R value. If the R value is less than or equal to the preset time threshold, it is assigned the value of 1; otherwise, it is assigned the value of 0. The number of non-repeating orders within the statistical period is recorded as the F value. If the F value is greater than or equal to the preset frequency threshold, it is assigned a value of 1; otherwise, it is assigned a value of 0. The cumulative purchase amount within the statistical period is recorded as the M value. If the M value is greater than the preset amount threshold, it is assigned the value of 1; otherwise, it is assigned the value of 0. The credit rating obtained from tobacco monopoly is recorded as a C value. If the credit rating is trustworthy, the value is 1; if it is untrustworthy, the value is 0.

[0011] In some embodiments, the preset time threshold is 7 days, that is, R value ≤ 6 talent value is 1, R value ≥ 7 talent value is 0; The preset frequency threshold is an average frequency of 48.89 times, that is, F value ≥ 49 is assigned a value of 1, and F value ≤ 48 is assigned a value of 0; The preset amount threshold is an average amount of 211,400 yuan, that is, if the M value is greater than 211,400 yuan, it is assigned a value of 1, and if the M value is less than or equal to 211,400 yuan, it is assigned a value of 0. The trustworthiness levels are AA and A, while the untrustworthiness levels are B, C, and D.

[0012] In some embodiments, the sum of squared errors within the cluster in S4 The calculation formula is: in, The number of clusters, For the first Clusters, For sample points within a cluster, For the first Cluster center vectors of each cluster; By calculating different Value Select The inflection point where the rate of decline changes from steep to gradual is taken as the optimal number of clusters.

[0013] In some embodiments, the specific characteristics of the five types of customers in S5 are as follows: Key high-quality customers: Customers with high credit scores and high frequency and amount of spending; Ordinary potential customers: Customers with high credit scores but moderate spending frequency or amount; Customers who are gradually losing contact: This group is at risk of losing customers whose purchase frequency or amount has declined and includes some customers with low credit in the development stage. Customers to focus on: Customers with high spending amounts or frequency but low credit scores and potential for violations; Customers we generally focus on: those with low spending power and low creditworthiness.

[0014] In some embodiments, the differentiation strategy in S7 includes: For key, high-quality clients, priority will be given to allocating new modern terminal resources and providing large-scale financial loan support; Increase the frequency of inspections of key and general customers and strengthen the regulatory mechanism. For customers who are gradually losing customers, strengthen operational guidance and optimize product allocation in terms of brand and product specifications to revitalize consumption.

[0015] A tobacco retail customer value stratification and classification system based on RFMC, used to implement the method described above, the system comprising: Data acquisition module: used to extract retail customer sales data, transaction dates, purchase amounts, credit ratings, and exclusive sales penalty records, and to perform data preprocessing; RFMC indicator calculation module: used to complete the quantitative calculation, threshold division and 0 / 1 assignment of the four indicators R, F, M and C; Clustering Hierarchy Module: Used to determine the optimal K value based on the elbow method, and to complete customer clustering using the K-means algorithm, outputting five-class hierarchical results; The three-tiered value segmentation module is used to switch between statistical calibers for all categories, single brands, and single product specifications, and outputs three tiered results for comprehensive value, brand value, and product specification value, respectively. Strategy Output Module: Used to match and output differentiated service, supply of goods, terminal construction, financial support and exclusive sales supervision strategies based on the stratified results.

[0016] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of any of the methods described above.

[0017] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0018] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions: I. This invention breaks through the limitations of the original purely marketing-oriented assessment by introducing the exclusive sales credit rating C index into the RFM model, deeply binding "trustworthiness / distrustworthiness" with "consumer value", and realizing data exchange and strategy synergy between marketing resource allocation and exclusive sales supervision.

[0019] Second, this invention automatically determines the optimal K value by employing K-means unsupervised clustering combined with the elbow method, replacing the traditional classification method that relies on manual experience to set thresholds. The five clustered groups have reasonable proportions and all have practical operational significance, and the algorithm can adaptively iterate as the data is updated.

[0020] Third, this invention uses a three-tiered value segmentation architecture of comprehensive, brand, and product specification. The same customer may be an ordinary potential customer in the whole category, but may be an important high-quality customer in a specific product specification. This fine-grained penetration provides direct data support for new product selection, personalized display, and precise supply of goods, and fulfills the primary responsibility of brand cultivation.

[0021] Fourth, this invention significantly reduces the cost of manual analysis by automatically completing the entire closed loop from data extraction, indicator calculation, cluster analysis to strategy generation. At the same time, the differentiated strategy ensures that high-quality resources are concentrated on high-value / high-credit customers, while regulatory resources are concentrated on low-credit customers, thereby improving overall channel control.

[0022] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram illustrating the composition of the RFMC model indicators of this invention; Figure 2 This is a flowchart illustrating the overall process of customer value stratification and classification based on RFMC in this invention. Figure 3 A diagram illustrating the distribution of customers by R-index. Figure 4 This is a diagram illustrating the distribution of customer proportions based on the F-index. Figure 5 A schematic diagram illustrating the determination of the optimal K value using the elbow method; Figure 6 A schematic diagram illustrating the clustering distribution of comprehensive customer value; Figure 7 A diagram illustrating the clustering distribution of customer brand value; Figure 8 A schematic diagram illustrating the clustering distribution of customer product value; Figure 9 This is a system architecture module diagram of the present invention. Detailed Implementation

[0025] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0026] It is important to note that terms such as "first," "second," "symmetric," "array," "set in," and "set with" are used only to distinguish between descriptive and positional descriptions and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, features specified with terms such as "first" or "symmetric" may explicitly or implicitly include one or more of that feature; similarly, when the quantity of certain features is not limited by words such as "two" or "three," it should be noted that such features also explicitly or implicitly include one or more features.

[0027] In this invention, unless otherwise explicitly specified and limited, terms such as "installation," "connection," and "fixation" should be interpreted broadly; for example, they can refer to a fixed connection, a detachable connection, or an integral molding; they can refer to a mechanical connection, a direct connection, a welding connection, or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the accompanying drawings and specific circumstances.

[0028] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] like Figures 1-9 As shown, this embodiment of the invention provides a method for classifying and stratifying the value of tobacco retail customers based on RFMC, including the following steps: S1. Data Extraction and Preprocessing: Extract transaction data and credit data of retail customers within a preset period from the tobacco data center, clean missing and outlier values, and aggregate them by customer ID to form a modeling dataset. S2, RFMC Four-Indicator Construction and Quantification Assignment: Based on the modeling dataset, an RFMC model is constructed that includes the most recent purchase R, purchase frequency F, purchase amount M, and credit rating C, and thresholds are set for 0 / 1 binarization assignment. S3, RFMC Combination and Preliminary Classification: The results of the four indicator assignments are combined to form 16 initial customer groups. Empty groups without actual samples are eliminated to obtain the effective groups. S4. Determining the optimal cluster number based on the elbow method: using the sum of squared errors within clusters. To evaluate the indicators, draw With the number of clusters The K value at the inflection point of the changing curve is taken as the optimal number of clusters; S5, K-means clustering and customer segmentation: Based on the optimal number of clusters, the K-means algorithm is used to perform unsupervised clustering of the effective group. According to the characteristics of the cluster centers, customers are divided into five categories: important high-quality customers, ordinary potential customers, customers with slight churn, key attention customers, and general attention customers. S6. Three-tier value segmentation: Using the full category, single brand, and single product specification as statistical criteria, repeat the RFMC construction and clustering process from S1 to S5 to output the customer's comprehensive value segmentation, brand value segmentation, and product specification value segmentation results. S7. Differentiated Strategy Output: Based on the three-tier value layering results, automatically match and output corresponding service, product supply, terminal construction, financial support, and exclusive sales supervision strategies.

[0030] In this embodiment, the specific quantization assignment rule in S2 is as follows: The number of days between the last purchase and the statistical date is calculated and denoted as the R value. If the R value is less than or equal to the preset time threshold, it is assigned the value of 1; otherwise, it is assigned the value of 0. The number of non-repeating orders within the statistical period is recorded as the F value. If the F value is greater than or equal to the preset frequency threshold, it is assigned a value of 1; otherwise, it is assigned a value of 0. The cumulative purchase amount within the statistical period is recorded as the M value. If the M value is greater than the preset amount threshold, it is assigned the value of 1; otherwise, it is assigned the value of 0. The credit rating obtained from tobacco monopoly is recorded as a C value. If the credit rating is trustworthy, the value is 1; if it is untrustworthy, the value is 0.

[0031] In this embodiment, specifically, the preset time threshold is 7 days, that is, R value ≤ 6 talent value is 1, R value ≥ 7 talent value is 0; The preset frequency threshold is an average frequency of 48.89 times, that is, F value ≥ 49 is assigned a value of 1, and F value ≤ 48 is assigned a value of 0; The preset threshold for the amount is an average amount of 211,400 yuan. That is, if the M value is greater than 211,400 yuan, it is assigned a value of 1, and if the M value is less than or equal to 211,400 yuan, it is assigned a value of 0. The trustworthy level is AA and A, while the untrustworthy level is B, C and D.

[0032] In this embodiment, specifically, the sum of squared errors within the cluster in S4 The calculation formula is: in, The number of clusters, For the first Clusters, For sample points within a cluster, For the first Cluster center vectors of each cluster; By calculating different Value Select The inflection point where the rate of decline changes from steep to gradual is taken as the optimal number of clusters.

[0033] In this embodiment, the specific characteristics of the five types of customers in S5 are as follows: Key high-quality customers: Customers with high credit scores and high frequency and amount of spending; Ordinary potential customers: Customers with high credit scores but moderate spending frequency or amount; Customers who are gradually losing contact: This group is at risk of losing customers whose purchase frequency or amount has declined and includes some customers with low credit in the development stage. Customers to focus on: Customers with high spending amounts or frequency but low credit scores and potential for violations; Customers we generally focus on: those with low spending power and low creditworthiness.

[0034] In this embodiment, the specific differentiation strategy in S7 includes: For key, high-quality clients, priority will be given to allocating new modern terminal resources and providing large-scale financial loan support; Increase the frequency of inspections of key and general customers and strengthen the regulatory mechanism. For customers who are gradually losing customers, strengthen operational guidance and optimize product allocation in terms of brand and product specifications to revitalize consumption.

[0035] A tobacco retail customer value stratification and classification system based on RFMC, used to implement any of the above methods, the system comprising: Data acquisition module: used to extract retail customer sales data, transaction dates, purchase amounts, credit ratings, and exclusive sales penalty records, and to perform data preprocessing; RFMC indicator calculation module: used to complete the quantitative calculation, threshold division and 0 / 1 assignment of the four indicators R, F, M and C; Clustering Hierarchy Module: Used to determine the optimal K value based on the elbow method, and to complete customer clustering using the K-means algorithm, outputting five-class hierarchical results; The three-tiered value segmentation module is used to switch between statistical calibers for all categories, single brands, and single product specifications, and outputs three tiered results for comprehensive value, brand value, and product specification value, respectively. Strategy Output Module: Used to match and output differentiated service, supply of goods, terminal construction, financial support and exclusive sales supervision strategies based on the stratified results.

[0036] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of any of the methods described above.

[0037] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.

[0038] Example 1: Comprehensive Value Hierarchical Classification Using 22,593 retail customers of a municipal-level tobacco commercial enterprise (such as Suqian Tobacco) as the analysis sample, sales and credit data for one calendar year were extracted.

[0039] S1. Data Extraction and Preprocessing: Remove customers who are closed or suspended from business, clean and repair abnormal records with missing transaction dates or negative purchase amounts, and aggregate transaction records by customer license number (ID).

[0040] The S2, RFMC four-indicator construction and quantification assignment adopts the following rules for binarization: Most recent purchase R: Calculates the number of days between the last purchase and the statistical date.

[0041] Set threshold .

[0042] Assignment function: like ; like .

[0043] Purchase frequency F: The number of times a purchase is not repeated within the statistical period, with a set threshold. (Average value of all customers).

[0044] Assignment function: like .

[0045] Purchase amount M: Cumulative purchase amount during the statistical period.

[0046] Set threshold (Average value of all customers).

[0047] Assignment function: like ; like .

[0048] Credit rating C: Retrieve credit data from the monopoly management system.

[0049] Assignment function: If the credit rating is AA or A; If the credit rating is B, C, or D.

[0050] The combination of S3 and RFMC, a four-indicator combination, theoretically forms 16 customer categories. In actual data, empty categories with a sample size of 0 are removed, resulting in 14-15 effective initial groups.

[0051] S4. Elbow method for determining the optimal number of clusters. Calculate different Sum of squared intra-cluster errors under the value The formula is: in For the customer's four-dimensional vector, For the first The centroid of a cluster.

[0052] draw Curve, Discover The curve shows a clear inflection point (elbow point). The rate of decline decreased sharply, so the optimal number of clusters was determined to be 5.

[0053] S5, K-means clustering and five-layer naming, using Euclidean distance as the metric, yielded 5 clusters after iterative convergence.

[0054] The clusters are named according to the characteristics of their centroid values ​​in the R, F, M, and C dimensions: Key high-quality customers (centroid tending towards 1,1,1,1): 33.24%; Ordinary potential customers (centroid tending towards 1,0,0,1 or 0,1,0,1): accounting for 56.81%; Customers experiencing slight churn (with low centroids in the F and M dimensions and C values ​​of 1 or 0): 8.01%; Key clients (centroids at F and M are high, but C is 0): 0.96%; Customers with a general focus (centroid tending towards 0,0,0,0): 0.98%.

[0055] S6-S7, Strategy Output: Prioritize the allocation of resources for the construction of new modern terminals for important and high-quality customers, and provide large-scale "Tobacco Merchant Loan" financial support; For key customers, the supply of goods will be prudently controlled, and the exclusive sales department will include them in the key targets of random inspections and increase the frequency of market inspections.

[0056] Example 2: Brand Value Stratification (Taking the "Nanjing" brand as an example) Unlike Example 1, this example only counts transaction data of retail customers under the single brand "Nanjing" in terms of data extraction criteria.

[0057] Due to the narrowing of the statistical scope to single brands, the F-value and M-value of each client decreased overall. The mean threshold under the brand-specific criteria was recalculated. and Then, RFMC values ​​are assigned, and elbow method and K-means clustering (K=5) are performed.

[0058] Clustering results show that ordinary potential customers account for 57.26%, important high-quality customers account for 32.64%, and customers who have been gradually lost account for 8.16%.

[0059] Strategy Output: For key high-quality clients of the "Nanjing" brand, prioritize the selection of new product display locations for the Nanjing brand and provide personalized display cabinet support; for ordinary potential clients, analyze their surrounding consumer profiles and appropriately increase the distribution of Nanjing brand cigarettes in high demand to enhance their enthusiasm for brand cultivation.

[0060] Example 3: Product Value Stratification (Taking "Su Tobacco (Soft Five-Star Red Cedar)" as an example) Unlike Example 1, the statistical scope of this example is further refined to a single product specification, "Su Tobacco (Soft Five-Star Red Cedar)".

[0061] Because the ordering scope of a single product specification is relatively narrow, the RFMC threshold specific to that product specification is recalculated and clustered.

[0062] Clustering results: Important high-quality customers account for 42.04%, ordinary potential customers account for 27.53%, and customers who have been lost account for 28.71% (the loss rate is significantly higher than that of the entire product category).

[0063] Strategy Output: For customers who are gradually losing interest in this product, account managers need to provide on-site business guidance, analyze the reasons for the slow sales of this product, and reactivate and retain them by improving gross profit margin or guiding the display position; for important and high-quality customers, ensure sufficient investment in this product to stabilize their profitability.

[0064] System Implementation Examples Corresponding to the above method, the present invention also provides a tobacco retail customer value hierarchical classification system based on RFMC, such as... Figure 9 As shown, it includes: Data acquisition module: Connects to the tobacco ERP and monopoly system to extract transaction and credit data; RFMC indicator calculation module: executes the indicator quantification and 0 / 1 assignment program; Clustering hierarchical module: Embedded elbow method and K-means algorithm engine, automatically outputs five-level classification; Three-tier value segmentation module: Provides a caliber switching interface and parallel calculation of comprehensive / brand / specification value; Strategy Output Module: Built-in strategy mapping rule library, which converts category tags into business action instructions and pushes them to the marketing and franchise operation platform.

[0065] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for classifying and stratifying the value of tobacco retail customers based on RFMC, characterized in that, Includes the following steps: S1. Data Extraction and Preprocessing: Extract transaction data and credit data of retail customers within a preset period from the tobacco data center, clean missing and outlier values, and aggregate them by customer ID to form a modeling dataset. S2. Construction and Quantification of RFMC Four Indicators: Based on the modeling dataset, an RFMC model is constructed that includes the most recent purchase R, purchase frequency F, purchase amount M, and credit rating C, and thresholds are set for 0 / 1 binarization. S3, RFMC Combination and Preliminary Classification: The results of the four indicator assignments are combined to form 16 initial customer groups. Empty groups without actual samples are eliminated to obtain the effective groups. S4. Determining the optimal cluster number based on the elbow method: using the sum of squared errors within clusters. To evaluate the indicators, draw With the number of clusters The K value at the inflection point of the changing curve is taken as the optimal number of clusters; S5. K-means clustering and customer segmentation: Based on the optimal number of clusters, the K-means algorithm is used to perform unsupervised clustering on the effective group. According to the characteristics of the cluster centers, customers are divided into five categories: important high-quality customers, ordinary potential customers, customers with slight churn, key attention customers, and general attention customers. S6. Three-tier value segmentation: Using the full category, single brand, and single product specification as statistical criteria, repeat the RFMC construction and clustering process from S1 to S5 to output the customer's comprehensive value segmentation, brand value segmentation, and product specification value segmentation results. S7. Differentiated Strategy Output: Based on the three-layer value stratification results, automatically match and output corresponding service, supply, terminal construction, financial support and exclusive sales supervision strategies.

2. The method for classifying and stratifying tobacco retail customer value based on RFMC according to claim 1, characterized in that, The quantization assignment rules in S2 are as follows: The number of days between the last purchase and the statistical date is calculated and denoted as the R value. If the R value is less than or equal to the preset time threshold, it is assigned the value of 1; otherwise, it is assigned the value of 0. The number of non-repeating orders within the statistical period is recorded as the F value. If the F value is greater than or equal to the preset frequency threshold, it is assigned a value of 1; otherwise, it is assigned a value of 0. The cumulative purchase amount within the statistical period is recorded as the M value. If the M value is greater than the preset amount threshold, it is assigned the value of 1; otherwise, it is assigned the value of 0. The credit rating obtained from tobacco monopoly is recorded as a C value. If the credit rating is trustworthy, the value is 1; if it is untrustworthy, the value is 0.

3. The method for classifying and stratifying tobacco retail customer value based on RFMC according to claim 2, characterized in that, The preset time threshold is 7 days, that is, R value ≤ 6 talent value is 1, R value ≥ 7 talent value is 0; The preset frequency threshold is an average frequency of 48.89 times, that is, F value ≥ 49 is assigned a value of 1, and F value ≤ 48 is assigned a value of 0; The preset amount threshold is an average amount of 211,400 yuan, that is, if the M value is greater than 211,400 yuan, it is assigned a value of 1, and if the M value is less than or equal to 211,400 yuan, it is assigned a value of 0. The trustworthiness levels are AA and A, while the untrustworthiness levels are B, C, and D.

4. The method for classifying and stratifying tobacco retail customer value based on RFMC according to claim 1, characterized in that, S4 contains the sum of squared intra-cluster errors. The calculation formula is: in, The number of clusters, For the first Clusters, For sample points within a cluster, For the first Cluster center vectors of each cluster; By calculating different Value Select The inflection point where the rate of decline changes from steep to gradual is taken as the optimal number of clusters.

5. The method for classifying and stratifying tobacco retail customer value based on RFMC according to claim 1, characterized in that, The specific characteristics of the five types of customers in S5 are as follows: Key high-quality customers: Customers with high credit scores and high frequency and amount of spending; Ordinary potential customers: Customers with high credit scores but moderate spending frequency or amount; Customers who are gradually losing contact: This group is at risk of losing customers whose purchase frequency or amount has declined and includes some customers with low credit in the development stage. Customers to focus on: Customers with high spending amounts or frequency but low credit scores and potential for violations; Customers we generally focus on: those with low spending power and low creditworthiness.

6. The method for classifying and stratifying tobacco retail customer value based on RFMC according to claim 1, characterized in that, The differentiation strategy in S7 includes: For key, high-quality clients, priority will be given to allocating new modern terminal resources and providing large-scale financial loan support; Increase the frequency of inspections of key and general customers and strengthen the regulatory mechanism. For customers who are gradually losing customers, strengthen operational guidance and optimize product allocation in terms of brand and product specifications to revitalize consumption.

7. A tobacco retail customer value stratification and classification system based on RFMC, characterized in that, The system for implementing the method according to any one of claims 1 to 6, the system comprising: Data acquisition module: used to extract retail customer sales data, transaction dates, purchase amounts, credit ratings, and exclusive sales penalty records, and to perform data preprocessing; RFMC indicator calculation module: used to complete the quantitative calculation, threshold division and 0 / 1 assignment of the four indicators R, F, M and C; Clustering Hierarchy Module: Used to determine the optimal K value based on the elbow method, and to complete customer clustering using the K-means algorithm, outputting five-class hierarchical results; The three-tiered value segmentation module is used to switch between statistical calibers for all categories, single brands, and single product specifications, and outputs three tiered results for comprehensive value, brand value, and product specification value, respectively. Strategy Output Module: Used to match and output differentiated service, supply chain, terminal construction, financial support and exclusive sales supervision strategies based on the stratified results.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images