Method and device for analyzing customer value based on operator business, and medium

Abstract

The present disclosure provides a kind of based on the customer value analysis method, device and medium of operator service, it is related to data analysis technical field, the method comprises: according to the characteristic data of customer using operator service, obtain customer income contribution arpu value, customer online age age value and customer service package use rate service_rate value;In customer value analysis RFM model, customer arpu value is regarded as consumption amount M, customer age age value and customer service_rate value are respectively regarded as consumption frequency F and time interval R one;According to RFM model and its M, F, R value, at least customer is classified into one of high-value customer, key customer, key customer and potential customer one.The present disclosure is according to the characteristic of operator service, selects suitable data as the R, F, M of RFM model, according to RFM model, operator customer is classified, and RFM model is introduced into the field of operator service customer value analysis.

Description

Technical Field

[0001] This disclosure relates at least to the field of data analysis technology, and in particular to a method, apparatus and medium for customer value analysis based on operator services. Background Technology

[0002] Customer value analysis is a crucial method for businesses to identify and assess the value of customers. It helps businesses allocate resources more effectively, improve customer satisfaction and loyalty, thereby increasing sales and profits. Research reveals that commonly used customer value analysis models include the RFM (Recency, Frequency, and Monetary) model.

[0003] However, due to the unique nature of their business and the differences in products and marketing objectives, telecommunications operators do not have customer value analysis models and marketing methods that can be directly copied. They need to combine marketing scenarios to conduct customer value analysis and formulate marketing strategies. Summary of the Invention

[0004] The technical problem to be solved by this disclosure is to provide a customer value analysis method, apparatus and medium based on operator services to address the above-mentioned shortcomings, so as to solve the problem of how to conduct customer value analysis based on operator services.

[0005] Firstly, this disclosure provides a customer value analysis method based on operator services, the method comprising:

[0006] Based on the characteristic data of customers using operator services, obtain the customer revenue contribution ARPU value, customer network time Age value, and customer service package usage rate Service_rate value;

[0007] In the RFM model for customer value analysis, the customer's ARPU value is used as the consumption amount M, and the customer's age value and service rate value are used as one of the consumption frequency F and time interval R, respectively.

[0008] Based on the RFM model and its M, F, R values, customers should be categorized into at least one of the following groups: high-value customers, key retention customers, key target customers, and potential customers.

[0009] Furthermore, based on customer usage data of operator services, the system obtains customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service Rate), specifically including:

[0010] The customer ARPU value is obtained based on at least one of the customer's order amount and customer consumption fee corresponding to each customer information.

[0011] The customer's age value is obtained based on at least one of the following: customer registration time, customer account lifespan, and customer service usage duration, which corresponds to each customer's information.

[0012] The customer's service_rate value is obtained based on at least one of the following: customer subscription amount, customer service remaining amount, customer service usage, and customer service satisfaction, corresponding to each customer's information.

[0013] Furthermore, the method also includes:

[0014] Using a monthly billing cycle, monthly customer data is obtained for all customers using operator services within the scope of analysis. Monthly customer data includes customer information, customer subscription amount, customer consumption expenses, customer registration time, customer account lifecycle, customer service usage duration, customer subscription amount, customer service remaining amount, customer service usage rate, and customer service satisfaction.

[0015] Preprocessing of customer monthly data includes detecting and handling abnormal data, converting character data to numerical values, and standardizing the numerical values ​​of customer monthly data.

[0016] At least a portion of monthly customer data should be selected as the primary characteristic data for customer value analysis, and at least a portion of monthly customer data should be selected as the secondary characteristic data for customer user profiling.

[0017] Furthermore, in the customer value analysis RFM model, the customer's ARPU value is used as the consumption amount M, and the customer's age value and service rate value are used as one of the consumption frequency F and time interval R, respectively. Specifically, this includes:

[0018] Based on the Customer Lifetime Value (CLV) model, calculate the current and / or future monthly averages of customer ARPU, customer age, and customer service rate.

[0019] The current and / or future monthly averages of customer ARPU, customer age, and customer service_rate are used as the consumption amount M, consumption frequency F, and time interval R in the RFM model, respectively.

[0020] Furthermore, based on the RFM model and its M, F, R values, customers are categorized into at least one of the following groups: high-value customers, key retention customers, key target customers, and potential customers. Specifically, these include:

[0021] Customers whose M is greater than the first threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model are marked as high-value customers.

[0022] Customers whose M is greater than the first threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model are marked as key customers to retain.

[0023] Customers whose M is less than the fourth threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model are marked as key customers for further exploration.

[0024] Customers whose M is less than the fourth threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model are marked as potential customers.

[0025] Furthermore, based on the RFM model and its M, F, R values, customers are categorized into at least one of the following groups: high-value customers, key retention customers, key target customers, and potential customers. Specifically, these include:

[0026] Obtain the mean values ​​of M, R, and F for each customer in the RFM model. and and variance σ arpu σ age and σ service_rate ;

[0027] Will Customers are marked as high-income contributing customers, Customers are marked as low-income contributing customers;

[0028] Will and High-income contributing customers are marked as high-value customers;

[0029] Will and High-income contributing customers are marked as key customers to retain;

[0030] Will and Low-income, high-contribution customers are marked as key customers for further development.

[0031] Will and Low-income contributing customers are marked as potential customers.

[0032] Furthermore, the method also includes:

[0033] according to calculate and according to Calculate σ arpu σ age and σ service_rate In the formula, n represents the number of M, R, and F values ​​for each customer, σ is the standard deviation, and x i Let x represent the i-th value of M, R, and F for each customer, where x is the mean.

[0034] Will or Customers who meet these criteria are marked as general customers.

[0035] Furthermore, the method also includes:

[0036] Based on the operator's business marketing tasks and user profiles, different operator business marketing plans are developed for high-value customers, key retention customers, key target customers, and potential customers. These plans include maintaining high-value customers, preventing the loss of key retention customers, exploring the needs of key target customers, and increasing the stickiness of potential customers.

[0037] Secondly, this disclosure provides a customer value analysis device based on operator services, the device comprising:

[0038] The data module is used to obtain customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service_Rate) based on the characteristics of customer use of operator services.

[0039] The assignment module, connected to the numerical module, is used in the customer value analysis RFM model to assign the customer's ARPU value as the consumption amount M, and the customer's age value and service_rate value as one of the consumption frequency F and time interval R, respectively.

[0040] The classification module, connected to the assignment module, is used to classify customers into at least one of the following categories based on the RFM model and its M, F, and R values: high-value customers, key retention customers, key target customers, and potential customers.

[0041] Thirdly, this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the customer value analysis method based on operator services as described above.

[0042] This disclosure provides a customer value analysis method, apparatus, and medium based on operator services. According to the characteristics of operator services, appropriate data is selected as the R, F, and M components of the RFM model. The selected data includes customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service Rate). Customers using operator services are classified according to the RFM model, and the customer categories include at least one of the following: high-value customers, key retention customers, key target customers, and potential customers. This realizes the introduction of the RFM model into the field of operator service customer value analysis, using the RFM model to evaluate customer value, thereby providing a basis for personalized marketing strategies and customer retention methods for operator services. Attached Figure Description

[0043] Figure 1 This is a flowchart of a customer value analysis method based on operator services according to an embodiment of this disclosure;

[0044] Figure 2 This is a schematic diagram of the structure of a customer value analysis device based on operator services according to an embodiment of this disclosure;

[0045] Figure 3 This is a flowchart of another customer value analysis method based on operator services, according to an embodiment of this disclosure;

[0046] Figure 4 This is an architectural diagram of a customer value analysis device based on operator services according to an embodiment of this disclosure;

[0047] Figure 5 This is a flowchart of a customer value RFM model analysis method according to an embodiment of this disclosure. Detailed Implementation

[0048] To enable those skilled in the art to better understand the technical solutions of this disclosure, the embodiments of this disclosure will be further described in detail below with reference to the accompanying drawings.

[0049] It is understood that the specific embodiments and accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of this disclosure.

[0050] It is understood that, without conflict, the various embodiments and features in the embodiments of this disclosure can be combined with each other.

[0051] It is understood that, for ease of description, only the parts relevant to this disclosure are shown in the accompanying drawings, while parts unrelated to this disclosure are not shown in the drawings.

[0052] It is understood that each module or unit involved in the embodiments of this disclosure may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple modules or units may be integrated into one entity structure.

[0053] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this disclosure may occur in a different order than that marked in the accompanying drawings.

[0054] It is understood that the flowcharts and block diagrams of this disclosure illustrate the architecture, functions, and operations of possible implementations of systems, apparatuses, devices, and methods according to various embodiments of this disclosure. Each block in a flowchart or block diagram may represent a module, unit, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagrams and flowcharts may be implemented using hardware-based devices to implement the specified function, or using a combination of hardware and computer instructions.

[0055] It is understood that the modules and units involved in the embodiments of this disclosure can be implemented by software or by hardware, for example, the modules and units can be located in a processor.

[0056] Example 1:

[0057] like Figure 1 As shown, this disclosure provides a customer value analysis method based on operator services, the method comprising:

[0058] S1. Based on the characteristic data of customers using operator services, obtain the customer revenue contribution (ARPU) value, customer network time (Age) value, and customer service package usage rate (Service_Rate) value.

[0059] S2. In the customer value analysis RFM model, the customer ARPU value is used as the consumption amount M, and the customer age value and customer service_rate value are used as one of the consumption frequency F and time interval R, respectively.

[0060] S3. Based on the RFM model and its M, F, R values, customers should be categorized into at least one of the following groups: high-value customers, key retention customers, key prospective customers, and potential customers.

[0061] In this embodiment, the method selects appropriate data as the R, F, and M components of the RFM model based on the characteristics of the operator's business. The selected data includes customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service Rate). Customers using the operator's business are classified according to the RFM model, into categories including high-value customers, key retention customers, key target customers, and potential customers (at least one of these). This introduces the RFM model into the field of customer value analysis for operator businesses, using it to evaluate customer value and providing a basis for personalized marketing strategies and customer retention methods for operator businesses. Figure 1 The method shown is applicable to, for example, Figure 2 The apparatus shown.

[0062] Specifically, in order to better help frontline staff expand their business, maintain high-quality customers, explore the needs of potential customers, and understand the pain points of customer needs, thereby accurately providing customers with more suitable products and services, improving customer satisfaction, and ultimately enhancing the company's operating performance and promoting high-quality development, this embodiment mainly relies on the operator's internal customer behavior big data to conduct customer value analysis, classify customers according to customer value, formulate personalized marketing strategies for different user groups, and manage marketing tasks.

[0063] Customer value analysis (CVA) methods leverage a company's existing customer behavior and marketing data, utilizing techniques such as data mining, machine learning, and deep learning to assess customer value, generate user profiles, and thus personalize marketing strategies and customer retention methods. Companies can choose appropriate models for CVA based on their own characteristics and needs. Through these models, companies can better understand customer behavior, predict customer value, and formulate corresponding marketing strategies and customer service plans.

[0064] Customer value analysis is an important business strategy that primarily involves assessing and understanding customer value, understanding customer needs and preferences, and comprehending the characteristics of different customer groups. This allows for the provision of more personalized services, optimization of customer relationship management, and improved business performance. Utilizing a pre-defined RFM mathematical model for customer value analysis is one method employed in various marketing scenarios.

[0065] For customer value analysis methods in operator businesses, approaches such as RFM-based analysis, risk user identification based on value analysis, cluster analysis and customer segmentation, and data mining can be drawn upon. However, given the different product offerings, market segments, and marketing objectives within operator businesses, there is currently no single customer value analysis and marketing management method that is fully adapted to the specific business scenarios of operators, such as home broadband and bundled packages.

[0066] This embodiment primarily provides a solution for how telecom operators can use customer value analysis to guide frontline marketing and customer retention management. For telecom operators, it collects customer behavior and marketing data, performs data extraction, preprocessing, and model analysis, and combines this with specific business operations to analyze the characteristics of various customer groups, identifying customers of different values ​​and types. Different marketing strategies are then developed for these different customer groups. This addresses the problem of identifying high-value users and potential users, and provides a marketing management system for scheduling marketing activities based on customer value analysis results.

[0067] In the RFM (Recency, Frequency, Monetary) model, Recency originally refers to the time of a customer's most recent purchase. The more recent the purchase, the more interested the customer is in the product or service, and the higher the customer value. It's commonly calculated by subtracting the customer's most recent purchase date from the current date. For example, if the current date is October 16, 2024, and a customer's most recent purchase date is October 10, 2024, then the R value is 6 days. Frequency originally refers to the number of times a customer makes a purchase within a certain period. The higher the purchase frequency, the higher the customer loyalty and the greater the customer value. It's generally calculated by counting the number of purchases made by a customer within a specific time period. For example, if a customer made 5 purchases in the past year, then the F value is 5. Monetary originally refers to the total amount a customer spends within a certain period. The larger the purchase amount, the stronger the customer's spending power and the greater the customer value. It's generally calculated by summing the total amount spent by a customer within a specific time period. For example, if a customer's total spending in the past year was 5000 yuan, then the M value is 5000.

[0068] Due to the unique nature of operator services, users typically make periodic purchases. Therefore, the methods for determining the values ​​of R, F, and M differ from those in traditional fields. Specifically, the values ​​for at least the most recent purchase time and purchase frequency need to be determined using a different approach. This embodiment analyzes the characteristics of operator services and selects customer revenue contribution (ARPU) (indicating customer spending power), customer network usage time (Age) (indicating customer loyalty), and customer service package usage rate (Service Rate) (indicating customer interest in goods or services) as equivalent to purchase amount (M), purchase frequency (F), and time interval (R). This allows the RFM model to be introduced into the operator service field. Furthermore, the RFM model is used to categorize customers using operator services into at least four types: high-value customers, key retention customers, key target customers, and potential customers.

[0069] In one implementation, S1, based on the characteristic data of customer's use of operator services, obtain the customer's revenue contribution (ARPU) value, customer's network usage time (Age) value, and customer service package usage rate (Service_Rate) value, specifically including:

[0070] The customer ARPU value is obtained based on at least one of the customer's order amount and customer consumption fee corresponding to each customer information.

[0071] The customer's age value is obtained based on at least one of the following: customer registration time, customer account lifespan, and customer service usage duration, which corresponds to each customer's information.

[0072] The customer's service_rate value is obtained based on at least one of the following: customer subscription amount, customer service remaining amount, customer service usage, and customer service satisfaction, corresponding to each customer's information.

[0073] In this embodiment, the data related to the operator's services and customers has more than ten dimensions. Calculating all of them would waste significant computing power and lead to model overfitting. Therefore, feature selection and dimensionality reduction are necessary. This embodiment selects user ARPU (Average Revenue per User), network usage time, and service package usage rate as the main indicators for model analysis. The remaining feature data is used to construct customer user profiles for marketing personnel to make decisions when conducting marketing activities. ARPU refers to the revenue the operator receives from each customer, also known as revenue contribution. It mainly comes from data such as the fees paid by customers for subscribing to operator services and monthly consumption fees. It can be considered comprehensively or individually, taking into account the customer's contribution to a single service. The age value can represent the length of time a customer has been a user of the operator or a consumer of a particular operator service. The service rate value can be characterized by the customer's usage of service packages, the number of people in their household broadband network, or user satisfaction. Both age and service rate values ​​indicate user loyalty.

[0074] In one embodiment, the method further includes:

[0075] Using a monthly billing cycle, monthly customer data is obtained for all customers using operator services within the scope of analysis. Monthly customer data includes customer information, customer subscription amount, customer consumption expenses, customer registration time, customer account lifecycle, customer service usage duration, customer subscription amount, customer service remaining amount, customer service usage rate, and customer service satisfaction.

[0076] Preprocessing of customer monthly data includes detecting and handling abnormal data, converting character data to numerical values, and standardizing the numerical values ​​of customer monthly data.

[0077] At least a portion of monthly customer data should be selected as the primary characteristic data for customer value analysis, and at least a portion of monthly customer data should be selected as the secondary characteristic data for customer user profiling.

[0078] In this embodiment, to better serve customers, prioritize high-value users, identify potential customers, increase user stickiness, and improve user satisfaction, operators need to utilize big data analytics. This involves analyzing customer behavior and characteristics, as well as pricing patterns, pain points, and satisfaction ratings, to identify different customer groups and provide customized services and products accordingly. Therefore, this embodiment provides a customer value analysis method based on operator business scenarios, such as... Figure 2 and 3 As shown, user data for a city is collected through the operator's big data platform, the data is preprocessed, and then a model is built and applied. The data analyzed by the model is directly applied to front-line business scenarios, and the model is optimized based on the application results and feedback.

[0079] like Figure 2 The data collection shown specifically includes: extracting user information, current user packages and consumption details, user registration time, and account lifecycle (e.g., account cancellation) from a city branch through the operator's big data platform, using a monthly data billing cycle; user satisfaction, home broadband quality rating, number of home terminals, home broadband dependency, current ARPU (Average Revenue per User), number of subscribed service packages (service_num), service package usage rate, and account lifespan, etc., and generating data on a monthly billing cycle. That is, in... Figure 3 The data layer shown collects data from various data sources, including user data, product data, and order data. After certain data processing, the data is collected and sent to the operator's big data platform.

[0080] like Figure 2 The data preprocessing shown includes data cleaning and processing, specifically: detecting and processing outlier data; outlier detection mainly involves null value (missing value) detection and outlier detection; different methods are used for different indicator characteristics for anomaly detection and data processing. For example, methods such as average imputation are used to handle possible missing and outlier values. MinMaxScaler is used for data standardization or normalization. One-hot encoding is used for character data, and recursive feature elimination is used for numerical data. For missing values, the `isnull()` or `isna()` functions from Python's Pandas library are used directly to detect missing values. For outliers, it is necessary to analyze whether the data format is correct, whether the numerical value range is within a reasonable range, and whether outliers exist. For example, a customer's ARPU value is greater than 0, but generally will not exceed 1000; values ​​outside this range may be outliers. If the data conforms to a normal distribution, the numerical range is generally within... arrive Data exceeding this range is generally priced as outliers. For outlier handling, if the data is missing key information, such as ARPU value or network duration, it will be removed. If the data format is incorrect, such as a home network quality rating of "Excellent," "Good," "Medium," or "Poor," the value will be converted and a numerical price will be used. Figure 3 The data processing performed by the model layer shown mainly includes anomaly data processing and data quantization.

[0081] like Figure 2 The feature selection shown is as follows: The data collected in this project includes more than ten dimensions such as customer codes, ordered service package amounts, average monthly spending, and ARPU (average revenue per user). Calculating all of these dimensions would waste significant computing power and lead to model overfitting. Therefore, feature selection is necessary to reduce the dimensionality of the data. This solution selects user ARPU, online time, and service package usage rate as the main indicators for model analysis. The remaining feature data is used to build customer user profiles for marketing personnel to make decisions when conducting marketing activities. Figure 3 The feature engineering performed on the model layer shown is primarily aimed at selecting features suitable for classifying the operator's customers and achieving data dimensionality reduction.

[0082] In one implementation, S2, in the customer value analysis RFM model, the customer ARPU value is used as the consumption amount M, and the customer age value and customer service rate value are used as one of the consumption frequency F and time interval R, respectively, specifically including:

[0083] Based on the Customer Lifetime Value (CLV) model, calculate the current and / or future monthly averages of customer ARPU, customer age, and customer service rate.

[0084] The current and / or future monthly averages of customer ARPU, customer age, and customer service_rate are used as the consumption amount M, consumption frequency F, and time interval R in the RFM model, respectively.

[0085] In this embodiment, the solution fully utilizes the RFM and CLV models to assess customer value and classify customers, segmenting customer types and identifying customer pain points. It then uses personalized customer service solutions to retain high-value users and uncover the needs of potential users. The RFM model is a customer value analysis model that measures customer value using three indicators: Recency, Frequency, and Monetary Amount. When classifying customers, the RFM model categorizes them based on the calculated R, F, and M values. For example, R, F, and M values ​​can be classified into high, medium, and low levels. Customers are then comprehensively classified based on these categories, and corresponding operational strategies are formulated accordingly. The CLV model, short for Customer Lifetime Value, is an important tool for predicting the total profit a customer can bring to a company in the future. The CLV model not only considers the current profit contribution of a customer but also predicts the value changes of a customer throughout their lifecycle. Therefore, it not only calculates the value of current customers but also predicts future value. Other models that can be used in conjunction with it include the Pareto model and the customer social value model.

[0086] Because the business scenarios of telecom operators differ somewhat from general shopping scenarios—such as airline ticket purchases and supermarket / department store purchases being one-off purchases—telecom operator services are generally cyclical, such as monthly fees. Therefore, the metrics of the RFM model from other application areas cannot be directly applied. This solution combines Customer Lifetime Value (CEV) with model analysis. Based on the impact of metric characteristics on customer value assessment, customer ARPU, recent spending amount, network duration, subscribed service packages, and remaining package status are selected for customer value analysis. Referring to the RFM and CLV models, ARPU is used to define spending amount (Monetary). Regarding recency and frequency, due to the special nature of telecom operator services, monthly data is collected. Based on the impact of metric characteristics on customer value analysis, the characteristic metrics of service package usage rate (service_rate) and network duration (user_id_age) are used to comprehensively assess customer consumption.

[0087] In one implementation, S3, based on the RFM model and its M, F, R values, customers are categorized into at least one of the following groups: high-value customers, key retention customers, key target customers, and potential customers, specifically including:

[0088] Customers whose M is greater than the first threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model are marked as high-value customers.

[0089] Customers whose M is greater than the first threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model are marked as key customers to retain.

[0090] Customers whose M is less than the fourth threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model are marked as key customers for further exploration.

[0091] Customers whose M is less than the fourth threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model are marked as potential customers.

[0092] In this embodiment, the RFM model is used to classify customers. There are classification methods that can be referenced. One feasible method is to set two thresholds for the M, R, and F values ​​respectively, and divide the M, R, and F into three levels: high, medium, and low, focusing mainly on customers in the high and low levels.

[0093] In one implementation, S3, based on the RFM model and its M, F, R values, customers are categorized into at least one of the following groups: high-value customers, key retention customers, key target customers, and potential customers, specifically including:

[0094] Obtain the mean values ​​of M, R, and F for each customer in the RFM model. and and variance σ arpu σ age and σ service_rate ;

[0095] x arpu >(x arpu +σ arpu Customers marked as high-income contributors will be assigned x arpu <(x arpu -σ arpu Customers of this type are marked as low-income contributing customers;

[0096] x age >x age +σ age And x service_rate >x service_rate +σ service_rate High-income contributing customers are marked as high-value customers;

[0097] x age <x age -σ age And x service_rate <x service_rate -σ service_rate High-income contributing customers are marked as key customers to retain;

[0098] x age >x age +σage And x service_rate >x service_rate +σ service_rate Low-income, high-contribution customers are marked as key customers for further development.

[0099] Will and Low-income contributing customers are marked as potential customers.

[0100] In this embodiment, as Figure 2 The detailed process of model analysis shown is as follows: Figure 5 As shown, it specifically includes:

[0101] A. To calculate the monthly average user score, missing and outlier values ​​were removed to ensure data authenticity and accuracy. Only normal data was used for the average calculation. The average calculation method is as follows:

[0102]

[0103] Here, the average values ​​of customer ARPU, network usage time, and service package utilization are calculated and denoted as follows: and

[0104] B. Calculate the standard deviation. The standard deviation represents the degree of deviation of the values ​​in the dataset from the mean. The variance of ARPU is calculated as follows:

[0105]

[0106] Where σ is the standard deviation, x i This represents the current ARPU value. It is the mean;

[0107] Here, the standard deviations of customer ARPU, online duration, and service package usage rate are calculated and denoted as σ. arpu σ age and σ service_rate .

[0108] C. Determine whether a user is high-value. Since there is a large amount of existing user data, it is generally not necessary to process every type of user data. Just focus on high-value users and low-value users according to the business scenario, and set the rest of the users as general users (such as mid-value users).

[0109] Users whose ARPU is greater than a custom threshold are designated as high-value users, and those whose ARPU is lower than the custom threshold are designated as low-value users. Users whose ARPU falls between the two thresholds are not processed for the time being.

[0110] The high-value user threshold defined in this scheme is: That is, x i Users with a value greater than or equal to this are marked as high-value users; the threshold for low-value users defined in this scheme is... That is, x i Users whose value is less than this amount are considered low-value users.

[0111] D. Assess user spending patterns. In order to better assess customer value and classify them, after understanding the customer's income value, it is also necessary to further segment customer types based on their package spending patterns throughout their lifecycle.

[0112] Generally, for users who contribute high monthly revenue, account lifespan (registration time) and service package usage affect their loyalty; for users who contribute low revenue, their potential is assessed by focusing on account lifespan (registration time) and service package usage. Service package usage and account lifespan also influence customer loyalty.

[0113] This plan sets a threshold for high package usage rates. The threshold for low usage is set to Data outside this range is automatically ignored; this solution sets the threshold for high online duration to [value missing]. The threshold for low network duration is set as follows: Data outside this range will be automatically ignored.

[0114] For high-income users, those with high plan usage and long online time are marked as high-value users and require focused maintenance and excellent service. Those with low plan usage and short online time are marked as users to be retained, as historical data suggests a higher risk of churn. For low-income users, those with high plan usage and long online time are marked as users to be explored, as they exhibit high loyalty and relatively good user stickiness, but their user value hasn't increased significantly. They require close attention, and efforts should be made to identify their pain points and needs to promote customer value growth. Those with low plan usage and short online time are marked as users with potential for further development.

[0115] In one embodiment, the method further includes:

[0116] according to calculate and according to Calculate σ arpu σ age and σ service_rate In the formula, n represents the number of M, R, and F values ​​for each customer, σ is the standard deviation, and x i This represents the i-th value of M, R, and F for each customer. It is the mean;

[0117] Will or Customers who meet these criteria are marked as general customers.

[0118] In one embodiment, the method further includes:

[0119] Based on the operator's business marketing tasks and user profiles, different operator business marketing plans are developed for high-value customers, key retention customers, key target customers, and potential customers. These plans include maintaining high-value customers, preventing the loss of key retention customers, exploring the needs of key target customers, and increasing the stickiness of potential customers.

[0120] In this embodiment, as Figure 2 The model application shown corresponds to, for example: Figure 3 The application layer shown identifies four types of customers through model analysis: high-value customers, key retention customers, key prospective customers, and potential customers. Customer information is then entered into the system and pushed to frontline staff for marketing and task management through marketing tasks along with other customer profile information.

[0121] like Figure 3 The measures for retaining high-value customers include: These customers contribute high monthly revenue, spend a lot of time online, subscribe to multiple service packages, and have good user loyalty, requiring focused service and maintenance. The retention strategy involves avoiding excessive telemarketing, promptly obtaining satisfaction surveys, and promptly following up and resolving any issues raised.

[0122] like Figure 3 The user churn warning measures shown include: targeting key retention customers, as their monthly revenue contribution is high and churn would result in significant performance losses. Based on historical data, these customers typically have shorter subscription periods, larger remaining balances on their plans, and lower frequency of product and service usage. It is necessary to promptly understand their service situation, identify pain points, and provide solutions to achieve user retention.

[0123] like Figure 3 The user demand mining measures shown include: targeting key customers for demand mining. These customers contribute less revenue per month, but have been on the network for a considerable time, use products and services extensively, and have relatively high loyalty. However, their user needs have not been fully explored, resulting in stagnant business revenue. Therefore, the primary marketing strategy is to identify their needs and business growth points based on their product and service usage and individual network dependence, encouraging them to upgrade service packages, fully exploring their needs, and thereby increasing business revenue.

[0124] like Figure 3The measures shown for identifying potential users include: targeting potential customers who contribute low monthly revenue, have short network tenures, use products and services infrequently, and have low loyalty. The marketing strategy is to find ways to get them to use the products and services, guide them to upgrade their services, thereby increasing user stickiness and fully tapping their potential.

[0125] In addition, such as Figure 3 As shown, the model layer can also be optimized in reverse based on the effect of the application layer model to improve the accuracy and effectiveness of model classification.

[0126] A specific implementation case is as follows:

[0127] Taking the marketing management module of a certain operator's integrated support platform in a certain city as a case study, this paper specifically illustrates the design method and implementation process of the solution.

[0128] The marketing management module's main task is to identify customer groups of different values ​​through data analysis and other methods, tag these groups, and simultaneously push other user information to frontline marketing personnel to facilitate business promotion. This application module connects to a telecom operator's big data platform to collect data such as terminal identification, user profiling (application behavior), quality defect identification, product data, and order details. Through data processing and analysis, it calculates ARP (Average Revenue Per User), the number of service packages ordered (service_num), service package usage rate, account lifespan, and other data. It comprehensively applies a customer value analysis method based on telecom operator business to identify and filter four types of customers: high-value customers, key retention customers, key target customers, and potential customers. This customer data is then pushed to frontline staff for customer maintenance and value extraction.

[0129] This system is developed based on a certain telecom operator's cloud platform, utilizing container technology and a cloud-native architecture for development and deployment. It also employs OSS, Redis, MQ, and RDS components for distributed microservice development. The main workflow is as follows:

[0130] 1) Data Acquisition: This case study utilizes a telecom operator's big data platform for data collection. The source data includes user data, product data, and order data. The big data platform has already processed the source data, calculating user ARPU, activation time, revenue information, etc., according to the billing cycle. Data is retrieved individually within a specified range using the operator's data request query interface, with monthly data selected. Data from August 2024 was used in this model comparison test.

[0131] 2) Data Preprocessing: Data preprocessing mainly involves the following tasks for data cleaning and processing: detecting and handling outliers; for missing values, the `isnull()` or `isna()` functions from Python's Pandas library are used to detect missing values ​​in the data. Key checks include whether the billing period is empty, and whether data such as customer ARPU, online duration, and service package usage rate are complete. For outliers, it is necessary to analyze whether the data format is correct, whether the numerical value range is within a reasonable range, and whether outliers exist. For example, a customer's ARPU value is greater than 0, but generally will not exceed 1000; values ​​outside this range may be outliers. If the data conforms to a normal distribution, the numerical range is generally within... arrive Values ​​exceeding this range are generally marked as outliers. For outlier handling, if the data is missing key information, such as ARPU value or network duration, it will be removed. If the data format is incorrect, such as a home network quality rating of "Excellent," "Good," "Medium," or "Poor," we will perform a numerical conversion and use a numerical price.

[0132] 3) Feature Selection: The data collected in this case includes customer codes, service package subscription amounts, average monthly spending, ARPU (Average Revenue Per User), and more than ten other dimensions. Calculating all of these dimensions would waste significant computing power and lead to model overfitting. Therefore, feature selection is necessary to reduce the dimensionality of the data. Five common dimensionality reduction methods are used: univariate feature selection, recursive feature elimination (RFE), optimal subset selection, stepwise regression, and regularization. To better evaluate customer value, stepwise regression was primarily used for feature selection, obtaining the three sets of labels that have the greatest impact on the overall trend: customer ARPU, online time, and service package usage rate.

[0133] 4) Model Analysis: The model analysis method primarily involves setting thresholds (which are customizable) to categorize customers based on the range of key customer metrics. The metric features and threshold settings are mainly based on the few metrics that have the greatest impact on customer value segmentation, extracted and analyzed during feature engineering. Threshold settings are primarily calculated based on the mean and standard deviation, but can also be customized. Threshold settings affect the accuracy of the evaluation data. Detailed analysis methods are described above.

[0134] 5) Model Application: Taking data from a project in a certain city as an example, out of more than 60,000 collected information entries in the city, 556 entries were identified as belonging to the project area, involving 39 indicator tags. Three key indicator characteristics were analyzed and calculated. Through data analysis, 45 high-value customers were identified; 18 key customers were identified for retention; 83 key potential customers were identified; and 26 potential customers were identified. A total of 172 customers were identified and marked, accounting for approximately 31% of all users. This avoided the need for mass customer acquisition and marketing. These customers were then targeted to frontline marketing personnel through the system for targeted marketing and maintenance services. In a broadband marketing campaign for the project, 83 users made appointments, of which 67 successfully completed speed tests. The remaining unsuccessful attempts were due to users being temporarily out or not answering calls. A total of 23 new users were acquired over two days, including 10 monthly subscription FTTR users, 9 new dual-new FTTR users, 1 new dual-new 70-bit network user, 1 existing number new broadband line, and 2 existing broadband new migration users.

[0135] In summary, customer value analysis is an important business strategy. It primarily involves assessing and understanding customer value, understanding customer needs and preferences, and comprehending the characteristics of different customer groups. This allows for more personalized services, optimized customer relationship management, and improved business performance. Customer value analysis methods for telecom operators mainly include RFM model-based analysis, Customer Lifetime Value (CLV) analysis, risk user identification based on value analysis, cluster analysis and customer segmentation, and data mining. However, given the different products, market segments, and marketing objectives, there is currently no single customer value analysis and marketing management method that is fully adapted to the specific business scenarios of telecom operators, such as home broadband and bundled packages.

[0136] This embodiment 1 primarily provides a solution for how operator services can use customer value analysis to guide frontline marketing and customer retention management. It addresses how to identify high-value and potential users and provides a marketing management system for scheduling marketing activities based on customer value analysis results. Specifically, this embodiment 1 has the following advantages:

[0137] 1) It comprehensively references the RFM model and CLV model, analyzes multiple feature data, and quantifies and manages customer value.

[0138] 2) The collected data is in monthly units, and the monthly customer behavior data and business data are output, which are suitable for time series evaluation and prediction. A customer value analysis method based on operator business is proposed, which can be effectively applied in front-line marketing.

[0139] 3) Abnormal data identification and abnormal data processing methods were adopted in data cleaning, namely null value identification and outlier identification, as well as data integrity identification methods.

[0140] 4) In practical applications, different types of target customers are identified based on the time, location, and marketing strategy of the actual marketing task, and the customer information and marketing strategies are pushed to front-line staff for marketing management.

[0141] 5) Based on the customer analysis results, four types of customers are identified: high-value customers, key retention customers, key target customers, and potential customers. Customer maintenance and customer value mining are carried out for each type of customer.

[0142] Example 2:

[0143] like Figure 2 As shown, this disclosure provides a customer value analysis device based on operator services, the device comprising:

[0144] Data module 1 is used to obtain customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service_Rate) based on the characteristic data of customers using operator services.

[0145] Assignment module 2, connected to numerical module 1, is used in the customer value analysis RFM model to assign the customer ARPU value as the consumption amount M, and the customer age value and customer service_rate value as one of the consumption frequency F and time interval R, respectively.

[0146] Classification module 3, connected to assignment module 2, is used to classify customers into at least one of the following categories based on the RFM model and its M, F, and R values: high-value customers, key retention customers, key target customers, and potential customers.

[0147] In one embodiment, data module 1 specifically includes:

[0148] The revenue data unit is used to obtain the customer ARPU value based on at least one of the customer's order amount and customer consumption fee corresponding to each customer information.

[0149] The duration data unit is used to obtain the customer's age value based on at least one of the following: customer registration time, customer account lifecycle, and customer usage time of each customer's information.

[0150] The data unit is used to obtain the customer's service_rate value based on at least one of the following: customer order volume, customer service remaining volume, customer service usage, and customer service satisfaction, corresponding to each customer's information.

[0151] In one embodiment, the apparatus further includes:

[0152] The data acquisition unit is used to acquire monthly customer data for all customers using operator services within the scope to be analyzed, with a monthly data billing period. The monthly customer data includes customer information, customer subscription amount, customer consumption fee, customer registration time, customer account life cycle, customer service usage time, customer subscription amount, customer service remaining amount, customer service usage rate, and customer service satisfaction.

[0153] The data preprocessing unit, connected to the data acquisition unit, is used to preprocess customer monthly data, including detecting and processing abnormal data, converting character data into numerical data, and standardizing the numerical values ​​of customer monthly data.

[0154] The feature selection unit, connected to the data preprocessing unit, is used to select at least a portion of monthly customer data as the first feature data for customer value analysis, and at least a portion of monthly customer data as the second feature data for customer user profiling.

[0155] In one embodiment, the assignment module 2 specifically includes:

[0156] The CLV unit is used to calculate the current and / or future monthly averages of customer ARPU, customer age, and customer service_rate based on the Customer Lifetime Value (CLV) model.

[0157] The RFM unit, connected to the CLV unit, is used to take the current and / or future monthly averages of the customer's ARPU value, customer's age value, and customer's service_rate value as the consumption amount M, consumption frequency F, and time interval R of the RFM model, respectively.

[0158] In one embodiment, the classification module 3 specifically includes:

[0159] The first type of unit is used to mark customers whose M is greater than the first threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model as high-value customers.

[0160] The second type of unit is used to mark customers whose M is greater than the first threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model as key customers to retain.

[0161] The third type of unit is used to mark customers whose M is less than the fourth threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model as key customers to be explored.

[0162] The fourth type of unit is used to mark customers whose M is less than the fourth threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model as potential customers.

[0163] In one embodiment, the classification module 3 specifically includes:

[0164] The mean and variance calculation unit is used to obtain the mean values ​​of M, R, and F for each customer in the RFM model. and and variance σ arpu σ age and σ service_rate ;

[0165] The revenue contribution calculation unit, connected to the mean and variance calculation unit, is used to calculate... Customers are marked as high-income contributing customers, Customers are marked as low-income contributing customers;

[0166] The first type of unit, connected to the revenue contribution calculation unit, is used to... and High-income contributing customers are marked as high-value customers;

[0167] The second type of unit, connected to the revenue contribution calculation unit, is used to... and High-income contributing customers are marked as key customers to retain;

[0168] The third type of unit, connected to the revenue contribution calculation unit, is used to... and Low-income, high-contribution customers are marked as key customers for further development.

[0169] The fourth type of unit, connected to the revenue contribution calculation unit, is used to... and Low-income contributing customers are marked as potential customers.

[0170] In one embodiment, wherein:

[0171] The mean and variance calculation unit is based on calculate and according to Calculate σ arpu σ age and σ service_rate In the formula, n represents the number of M, R, and F values ​​for each customer, σ is the standard deviation, and x i This represents the i-th value of M, R, and F for each customer. It is the mean;

[0172] The classification module 3 also includes a fifth category unit, connected to the income contribution calculation unit, used to classify... or Customers who meet these criteria are marked as general customers.

[0173] In one embodiment, the apparatus further includes:

[0174] The application module, connected to the classification module 3, is used to combine operator business marketing tasks and user profiles to develop different operator business marketing plans for high-value customers, key retention customers, key target customers, and potential customers. These plans include maintaining high-value customers, preventing the loss of key retention customers, exploring the needs of key target customers, and increasing the stickiness of potential customers.

[0175] Example 3:

[0176] Embodiment 3 of this disclosure provides a computer-readable storage medium storing a computer program. When the computer program is run by a processor, it implements the customer value analysis method based on operator services as described in Embodiment 1, or the customer value analysis device based on operator services as described in Embodiment 2.

[0177] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program units, or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0178] Additionally, this disclosure may provide a computer device including a memory and a processor. The memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the customer value analysis method based on carrier services as described in Embodiment 1. This computer device may be the customer value analysis device based on carrier services as described in Embodiment 2.

[0179] The memory is connected to the processor. The memory can be flash memory, read-only memory or other types of memory. The processor can be a central processing unit or a microcontroller.

[0180] Embodiments 1-3 of this disclosure provide a customer value analysis method, apparatus, and medium based on operator services. According to the characteristics of operator services, appropriate data is selected as the R, F, and M of the RFM model. The selected data includes customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service Rate). Customers using operator services are classified according to the RFM model, and the customer categories include at least one of high-value customers, key retention customers, key target customers, and potential customers. This realizes the introduction of the RFM model into the field of operator service customer value analysis, using the RFM model to evaluate customer value, thereby providing a basis for personalized marketing strategies and customer retention methods for operator services.

[0181] It is understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of this disclosure, and this disclosure is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this disclosure, and these modifications and improvements are also considered to be within the scope of protection of this disclosure.

Claims

1. A customer value analysis method based on operator services, characterized in that, The method includes: Based on the characteristic data of customers using operator services, obtain the customer revenue contribution ARPU value, customer network time Age value, and customer service package usage rate Service_rate value; In the RFM model for customer value analysis, the customer's ARPU value is used as the consumption amount M, and the customer's age value and service rate value are used as one of the consumption frequency F and time interval R, respectively. Based on the RFM model and its M, F, R values, customers should be categorized into at least one of the following groups: high-value customers, key retention customers, key target customers, and potential customers.

2. The method according to claim 1, characterized in that, Based on customer service usage characteristics, obtain customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service Rate). Specifically, this includes: The customer ARPU value is obtained based on at least one of the customer's order amount and customer consumption fee corresponding to each customer information. The customer's age value is obtained based on at least one of the following: customer registration time, customer account lifespan, and customer service usage duration, which corresponds to each customer's information. The customer's service_rate value is obtained based on at least one of the following: customer subscription amount, customer service remaining amount, customer service usage, and customer service satisfaction, corresponding to each customer's information.

3. The method according to claim 2, characterized in that, The method further includes: Using a monthly billing cycle, monthly customer data is obtained for all customers using operator services within the scope of analysis. Monthly customer data includes customer information, customer subscription amount, customer consumption expenses, customer registration time, customer account lifecycle, customer service usage duration, customer subscription amount, customer service remaining amount, customer service usage rate, and customer service satisfaction. Preprocessing of customer monthly data includes detecting and handling abnormal data, converting character data to numerical values, and standardizing the numerical values ​​of customer monthly data. At least a portion of monthly customer data should be selected as the primary characteristic data for customer value analysis, and at least a portion of monthly customer data should be selected as the secondary characteristic data for customer user profiling.

4. The method according to claim 3, characterized in that, In the RFM (Return on Purchase) model for customer value analysis, customer ARPU (Average Revenue Per User) is used as the purchase amount M, and customer age and service rate are used as one of the purchase frequency F and time interval R, respectively. Specifically, this includes: Based on the Customer Lifetime Value (CLV) model, calculate the current and / or future monthly averages of customer ARPU, customer age, and customer service rate. The current and / or future monthly averages of customer ARPU, customer age, and customer service_rate are used as the consumption amount M, consumption frequency F, and time interval R in the RFM model, respectively.

5. The method according to any one of claims 1-4, characterized in that, Based on the RFM model and its M, F, R values, customers should be categorized into at least one of the following groups: high-value customers, key retention customers, key acquisition customers, and potential customers. Specifically, these include: Customers whose M is greater than the first threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model are marked as high-value customers. Customers whose M is greater than the first threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model are marked as key customers to retain. Customers whose M is less than the fourth threshold, R is greater than the second threshold, and F is greater than the third threshold in the RFM model are marked as key customers for further exploration. Customers whose M is less than the fourth threshold, R is less than the fifth threshold, and F is less than the sixth threshold in the RFM model are marked as potential customers.

6. The method according to claim 5, characterized in that, Based on the RFM model and its M, F, R values, customers should be categorized into at least one of the following groups: high-value customers, key retention customers, key acquisition customers, and potential customers. Specifically, these include: Obtain the mean values ​​of M, R, and F for each customer in the RFM model. and and variance σ arpu σ age and σ service_rate ; Will Customers are marked as high-income contributing customers, Customers are marked as low-income contributing customers; Will and High-income contributing customers are marked as high-value customers; Will and High-income contributing customers are marked as key customers to retain; Will and Low-income, high-contribution customers are marked as key customers for further development. Will and Low-income contributing customers are marked as potential customers.

7. The method according to claim 6, characterized in that, The method further includes: according to calculate and according to Calculate σ arpu σ age and σ service_rate In the formula, n represents the number of M, R, and F values ​​for each customer, σ is the standard deviation, and x i This represents the i-th value of M, R, and F for each customer. It is the mean; Will or Customers who meet these criteria are marked as general customers.

8. The method according to claim 5, characterized in that, The method further includes: Based on the operator's business marketing tasks and user profiles, different operator business marketing plans are developed for high-value customers, key retention customers, key target customers, and potential customers. These plans include maintaining high-value customers, preventing the loss of key retention customers, exploring the needs of key target customers, and increasing the stickiness of potential customers.

9. A customer value analysis device based on operator services, characterized in that, The device includes: The data module is used to obtain customer revenue contribution (ARPU), customer network usage time (Age), and customer service package usage rate (Service_Rate) based on the characteristics of customer use of operator services. The assignment module, connected to the numerical module, is used in the customer value analysis RFM model to assign the customer's ARPU value as the consumption amount M, and the customer's age value and service_rate value as one of the consumption frequency F and time interval R, respectively. The classification module, connected to the assignment module, is used to classify customers into at least one of the following categories based on the RFM model and its M, F, and R values: high-value customers, key retention customers, key target customers, and potential customers.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the customer value analysis method based on operator services as described in any one of claims 1-8.

Images