Intelligent marketing decision-making system based on dynamic customer grouping

Abstract

The invention relates to the technical field of bank customer marketing scenes, in particular to an intelligent marketing decision-making system based on dynamic customer grouping. Comprising a data acquisition and preprocessing module, a dynamic customer grouping module, a multi-channel marketing collaboration module, a marketing effect evaluation and optimization module, a data security and privacy protection module and a customer feedback and complaint processing module. According to the intelligent marketing decision-making system based on dynamic customer grouping, the real-time data acquisition and preprocessing module is adopted, customer behavior data are captured in real time through a stream processing engine, real-time cleaning and standardization processing are carried out, meanwhile, the dynamic customer grouping module carries out dynamic grouping according to the customer real-time behavior data by using a deep clustering algorithm, and the dynamic customer grouping efficiency is improved. A traditional static grouping rule is replaced, the data timeliness is improved, marketing decisions can be based on latest customer behavior data, and market changes and customer demands can be captured in time.

Description

Technical Field

[0001] The present invention relates to the technical field of bank customer marketing scenarios, and in particular to an intelligent marketing decision-making system based on dynamic customer grouping. Background Art

[0002] The intelligent marketing decision-making system is a comprehensive system that integrates advanced data analysis, artificial intelligence technology and automated processes to optimize the efficiency and effectiveness of bank marketing activities. Its core lies in utilizing customer data and providing banks with accurate marketing strategies and decision-making support through intelligent algorithms and models.

[0003] Existing intelligent marketing decision-making systems rely on T+1 batch data processing methods, resulting in delayed responses to customer behavior and an inability to capture market changes and customer needs in a timely manner. At the same time, customer segmentation methods based on static tags rely on historical transaction data to generate fixed classification rules, have a low update frequency, and are unable to adapt to dynamic market changes.

[0004] The existing intelligent marketing decision-making system relies on the T+1 batch data processing method, which leads to delayed customer behavior response and inability to capture market changes and customer needs in a timely manner. At the same time, the customer clustering method based on static tags relies on historical transaction data to generate fixed classification rules, has a low update frequency, and cannot adapt to dynamic market changes. This solution is based on an intelligent marketing decision-making system for dynamic customer clustering. It adopts a real-time data acquisition and preprocessing module to capture customer behavior data in real time through a stream processing engine, and performs real-time cleaning and standardization. At the same time, the dynamic customer clustering module uses a deep clustering algorithm to dynamically cluster according to real-time customer behavior data, replacing traditional static clustering rules, improving data timeliness, enabling marketing decisions to be based on the latest customer behavior data and capture market changes and customer needs in a timely manner. At the same time, dynamic clustering rules can adapt to dynamic market changes, improve marketing response rate and business agility, and reduce the impact of the imbalance between data timeliness and business agility. Summary of the Invention

[0005] In order to overcome the problems raised in the above background technology, the present invention proposes an intelligent marketing decision-making system based on dynamic customer grouping.

[0006] The technical solution of the present invention is: an intelligent marketing decision-making system based on dynamic customer grouping, comprising: Data collection and preprocessing module: used to solve the problem of insufficient data timeliness, collect and preprocess customer data in real time; Dynamic customer segmentation module: used to improve marketing response rates and adapt to changes in customer behavior through dynamic segmentation; Multi-channel marketing collaboration module: used to improve multi-channel marketing collaboration and avoid duplicate reach; Marketing Effect Evaluation and Optimization Module: for shortening the evaluation cycle of marketing effect, real-time feedback optimization strategy; Data Security and Privacy Protection Module: for reducing data privacy and compliance risks, ensuring data security, and protecting customer privacy; Customer Feedback and Complaint Handling Module: for reducing customer complaints, improving customer satisfaction, and optimizing customer experience.

[0007] As preferred, the data collection and preprocessing module includes: A11: Real-time data collection unit, including stream processing engine, data interface adapter and data cache component, for real-time capture of customer behavior data through stream processing technology; A12: Data cleaning unit, including data cleaning algorithm library, outlier detection module and data completion tool, for cleaning the collected raw data, removing noise and outliers, and completing missing data; A13: Data standardization unit, including data format converter, data mapping table and data validation rule engine, for converting the cleaned data into a unified format.

[0008] As preferred, the dynamic customer segmentation module includes: A21: Feature extraction unit, including machine learning algorithm library, feature storage database and feature update module, for extracting customer features using machine learning algorithms; A22: Clustering algorithm unit, including deep clustering algorithm library, algorithm parameter configuration tool and clustering result visualization module, for applying deep clustering algorithms to dynamically segment customers, and adjusting the segmentation results in real time according to customer features and behavior changes; A23: Segmentation update unit, including segmentation update strategy engine, update frequency configuration tool and segmentation version management module, for regularly updating the segmentation results.

[0009] As preferred, the dynamic customer segmentation module includes the following steps when working: S11: The business management personnel uploads the relevant business raw data to the front-end server through the system front-end interface; S12: The server automatically checks the data format and content, if the verification is passed, the data format standardization processing is carried out, and the relevant data is transmitted to the data analysis system; if the verification fails, a pop-up window is prompted on the front-end page to require re-uploading data; S13: The data analysis system scans the file transmission directory in real time, if there is data file, the file is stored in the data processing queue for waiting; S14: According to the preset time task, the first data in the waiting queue is collected and stored in the data lake; S15: clean the data, including outlier removal and missing value filling, normalize the data after cleaning, use mean filling method to handle missing values, the formula is: wherein represents the filled value, represents other non-missing values, represents the number of non-missing values; S16: extract customer features using machine learning algorithms, select important features through feature selection algorithms; S17: apply deep clustering algorithm to dynamically group customers, and divide customers into different groups according to the algorithm results; S18: update the grouping results regularly, update the grouping results every hour or every day, and store the updated grouping results in the database.

[0010] As preferred, the multi-channel marketing synergy module comprises: A31: channel management unit, including channel interface adapter, channel state monitoring module and channel performance evaluation tool, for integrating and managing different marketing channels; A32: synergy strategy unit, including cross-channel marketing synergy strategy library, strategy configuration tool and strategy execution engine, for formulating cross-channel marketing synergy strategies; A33: reach control unit, including reach frequency limit module, reach content management module and reach effect tracking module, for implementing the fuse mechanism, limiting the daily reach frequency, avoiding disturbing the customers, and tracking the reach effect.

[0011] As preferred, the multi-channel marketing synergy module comprises the following steps when working: S21: the intelligent marketing system receives the customer list and the importance score of the features; S22: analyze customer behavior characteristics using data analysis algorithms, and the support degree formula of association rules is: wherein represents the support degree of association rules , and represents the probability of the occurrence of event and event ; S23: according to the analysis results, formulate targeted marketing strategies and channel reach methods, for high-frequency consumption customers, coupon issuing strategies can be formulated; for potential customers, new product promotion strategies can be formulated; S24: set the fuse threshold in advance to limit the daily reach frequency and prevent excessive disturbance to customers, and set the daily reach frequency not more than 3 times; S25: Reach customers through multi-channel coordination mechanism, activity placement considerations include channel priority, reach time and reach content; S26: Track the effect of marketing activities, and adjust marketing strategies and activity placement methods in a timely manner according to the tracking results.

[0012] As preferred, the marketing effect evaluation and optimization module includes: A41: Effect evaluation unit, including marketing effect index library, real-time calculation engine and effect visualization module, for real-time calculation of marketing activity effect index; A42: Strategy optimization unit, including optimization algorithm library, strategy generation module and strategy test environment, for automatically adjusting marketing strategies according to evaluation results; A43: Feedback loop unit, including feedback collection module, feedback analysis module and strategy update module, for forming a closed-loop feedback mechanism to continuously collect and analyze marketing activity feedback.

[0013] As preferred, the marketing effect evaluation and optimization module includes the following steps when working: S31: After the activity ends, the intelligent marketing system automatically collects pre-buried data on related pages, including customer behavior data and transaction data; S32: Generate visual statistical analysis report to show activity effect and customer feedback, statistical analysis methods include descriptive statistics and hypothesis testing; S33: Evaluate the effect of marketing activities according to the results of statistical analysis report, calculate through return on investment formula, according to the evaluation results, judge whether the marketing activities achieve the expected effect, the return on investment formula is: Where, represents the return on investment, represents the income brought by marketing activities, represents the cost of marketing activities; S34: According to the evaluation results, give suggestions for next marketing strategies for different customer groups, for customer groups with high response rate, marketing efforts can be increased; for customer groups with low response rate, marketing strategies or channels can be adjusted, strategy optimization considerations include customer segmentation, channel effect and marketing content effect; S35: Provide a path for quickly deploying related activities to adjust and optimize subsequent marketing activities in a timely manner, activity placement considerations include optimized strategies, channel priority and reach time.

[0014] As preferred, the data security and privacy protection module includes: A51: Data encryption unit, including encryption algorithm library, key management module and encryption performance optimization tool, for encryption processing of sensitive data; A52: an access control unit including an identity authentication module, a permission management module, and an audit log module, for implementing strict access control policies while recording data access and operation logs; A53: an audit tracking unit including an audit rule engine, an audit log analysis module, and a violation alarm module, for auditing and tracking data access and operations, and timely discovering and handling violation behaviors.

[0015] As a preferred, the customer feedback and complaint handling module includes: A61: a feedback collection unit including a multi-channel feedback interface, a feedback data integration module, and a feedback classification tool, for collecting customer feedback through multiple channels; A62: a complaint classification unit including a natural language processing algorithm library, a complaint classification model, and a classification result visualization module, for automatically classifying and prioritizing complaints; A63: a complaint handling unit including a complaint allocation module, a processing progress tracking module, and a processing result feedback module, for allocating complaints to corresponding teams for processing, tracking processing progress, and feeding back processing results to customers.

[0016] The beneficial effects of the present application are: 1. Compared with the traditional intelligent marketing decision system relying on T+1 batch data processing method, the customer behavior response is lagged, and the market changes and customer demands cannot be captured in time. Meanwhile, the customer grouping method based on static tags relies on historical transaction data to generate fixed classification rules, and the update frequency is low, which cannot adapt to market dynamic changes. The intelligent marketing decision system based on dynamic customer grouping in the present application adopts a real-time data acquisition and preprocessing module, captures customer behavior data in real time through a stream processing engine, and performs real-time cleaning and standardization processing. Meanwhile, the dynamic customer grouping module uses a deep clustering algorithm to dynamically group customers according to real-time behavior data, replaces the traditional static grouping rules, improves the data timeliness, makes the marketing decision based on the latest customer behavior data, captures the market changes and customer demands in time, and the dynamic grouping rules can adapt to market dynamic changes, improve the marketing response rate and business agility, and reduce the impact of data timeliness and business agility imbalance; 2、Compared with the traditional intelligent marketing decision system, there is a risk of plaintext data leakage in the data sharing process, which violates privacy protection regulations, and at the same time, to protect privacy, the traditional system may over-desensitize the data, resulting in reduced data value and affecting marketing effect, the scheme is based on an intelligent marketing decision system based on dynamic customer segmentation, a data security and privacy protection module uses data encryption technology to encrypt sensitive data, ensuring data security during sharing, and strict access control strategy is implemented through the access control unit to avoid plaintext data leakage. In addition, the system balances data value mining and privacy compliance requirements through fine-grained data desensitization and permission management, avoids over-desensitization, reduces compliance risk, protects customer privacy, maximizes data value, and improves the accuracy and effectiveness of marketing decisions; 3、Compared with the traditional intelligent marketing decision system, multi-channel marketing touch tools lack dynamic coordination mechanism, and marketing channels such as SMS, APP push, and outbound call operate independently, resulting in repeated push or missed customers, and due to the lack of a fuse mechanism, multi-channel customers are repeatedly touched, which can easily cause customer dissatisfaction and complaints. The scheme is based on an intelligent marketing decision system based on dynamic customer segmentation, and a multi-channel marketing coordination module integrates multi-channel marketing touch tools to realize unified coordination of cross-channel marketing activities through intelligent channel coordination and fuse mechanism. The system pre-sets a fuse threshold to limit the daily touch frequency and avoid repeated customer touch, improve multi-channel marketing coordination, and improve customer experience. At the same time, it reduces the probability of customer complaints and enhances customer satisfaction and loyalty. BRIEF DESCRIPTION OF DRAWINGS

[0017] Fig. 1 A dynamic customer segmentation-based intelligent marketing decision system framework process schematic diagram of the application is shown. Fig. 2 A dynamic customer segmentation-based intelligent marketing decision system dynamic customer segmentation module workflow schematic diagram of the application is shown. DETAILED DESCRIPTION

[0018] The application will be further described below in conjunction with the drawings and examples.

[0019] Please refer to Figs. 1-2 The application provides an embodiment: an intelligent marketing decision system based on dynamic customer segmentation, comprising: A data acquisition and preprocessing module is used to solve the problem of insufficient data timeliness, and real-time acquisition and preprocessing of customer data; A dynamic customer segmentation module is used to improve the marketing response rate by dynamically segmenting to adapt to customer behavior changes; A multi-channel marketing coordination module is used to improve multi-channel marketing coordination and avoid repeated touch; Marketing Effect Evaluation and Optimization Module: for shortening the marketing effect evaluation cycle, real-time feedback optimization strategy; Data Security and Privacy Protection Module: for reducing data privacy and compliance risks, ensuring data security, and protecting customer privacy; Customer Feedback and Complaint Handling Module: for reducing customer complaints, improving customer satisfaction, and optimizing customer experience.

[0020] In this embodiment, the data collection and preprocessing module includes: A11: Real-time data collection unit, including stream processing engine, data interface adapter and data cache component, for real-time capture of customer behavior data through stream processing technology; A12: Data cleaning unit, including data cleaning algorithm library, outlier detection module and data completion tool, for cleaning the collected raw data, removing noise and outliers, and completing missing data; A13: Data standardization unit, including data format converter, data mapping table and data validation rule engine, for converting the cleaned data into a unified format.

[0021] As described above, the data collection and preprocessing module uses the real-time data collection unit to capture customer behavior data in real time using the stream processing engine, ensuring the timeliness and freshness of the data, so that marketing decisions can quickly respond to market changes and customer needs. At the same time, the data cleaning unit effectively removes noise and outliers in the data and completes missing data, significantly improving the quality and accuracy of the data, providing a reliable foundation for subsequent analysis. In addition, the data standardization unit converts the cleaned data into a unified format, ensuring data consistency and comparability, greatly facilitating subsequent data processing and analysis work.

[0022] As a preferred, the dynamic customer segmentation module includes: A21: Feature extraction unit, including machine learning algorithm library, feature storage database and feature update module, for extracting customer features using machine learning algorithms; A22: Clustering algorithm unit, including deep clustering algorithm library, algorithm parameter configuration tool and clustering result visualization module, for applying deep clustering algorithms to dynamically segment customers, and adjusting the segmentation results in real time according to customer features and behavior changes; A23: Segmentation update unit, including segmentation update strategy engine, update frequency configuration tool and segmentation version management module, for regularly updating the segmentation results.

[0023] As described above, the dynamic customer clustering module dynamically updates the label by combining the customer multi-dimensional data through the deep clustering algorithm, realizes real-time accurate clustering, solves the problems of data lag and static clustering in the traditional marketing system, can respond to customer behavior changes in time, significantly improves the marketing response rate, at the same time, the customer portrait and feature importance score built by the module provide strong support for formulating personalized marketing strategies, further enhance the marketing effect and customer satisfaction.

[0024] As preferred, the dynamic customer clustering module includes the following steps when working: S11: The business management personnel uploads the relevant business raw data to the front-end server through the system front-end interface; S12: The server automatically checks the data format and content, if the verification is passed, the data format standardization processing is carried out, and the related data is transmitted to the data analysis system; if the verification fails, a pop-up window is prompted on the front-end page to require re-uploading data; S13: The data analysis system scans the file transmission directory in real time, if there is data file, the file is stored in the data processing queue for waiting; S14: According to the preset time task, the first data in the waiting queue is collected and stored in the data lake; S15: The data is cleaned, including outlier rejection and missing value filling processing, the data after cleaning is standardized, the missing value is processed by using the mean filling method, the formula is: Wherein represents the filled value, represents other non-missing values, represents the number of non-missing values; S16: The customer features are extracted by using the machine learning algorithm, and the important features are selected by using the feature selection algorithm; S17: The deep clustering algorithm is applied to dynamically cluster the customers, and the customers are divided into different groups according to the algorithm result; S18: The clustering result is updated regularly, the clustering result is updated every hour or every day, and the updated clustering result is stored in the database.

[0025] As described above, the working steps of the dynamic customer clustering module significantly improve the accuracy and timeliness of marketing decisions. First, the management personnel uploads the raw data to the front-end server, and the system automatically checks and standardizes the data format to ensure data quality, laying a solid foundation for subsequent analysis. Then, the data analysis system scans and stores data in real time, cleans and normalizes the data, effectively removes noise and outliers, and fills in missing data, further improving data quality. On this basis, the system uses machine learning algorithms to extract customer features and selects important features through feature selection algorithms to provide strong support for dynamic clustering. Next, the deep clustering algorithm is applied to dynamically cluster customers, and the clustering results are adjusted in real time according to customer characteristics and behavior changes to ensure the accuracy and timeliness of clustering. Finally, the system updates the clustering results regularly and stores the updated clustering results in the database to provide timely and accurate customer clustering information for subsequent marketing activities.

[0026] As a preferred embodiment, the multi-channel marketing collaboration module comprises: A31: a channel management unit, including a channel interface adapter, a channel state monitoring module, and a channel performance evaluation tool, for integrating and managing different marketing channels; A32: a collaborative strategy unit, including a cross-channel marketing collaboration strategy library, a strategy configuration tool, and a strategy execution engine, for formulating cross-channel marketing collaboration strategies; A33: a reach control unit, including a reach frequency limitation module, a reach content management module, and a reach effect tracking module, for implementing a fuse mechanism to limit daily reach frequency and avoid disturbing customers, while tracking reach effects.

[0027] As described above, the multi-channel marketing collaboration module integrates and manages different marketing channels, formulates and executes cross-channel marketing collaboration strategies, and implements a fuse mechanism to limit daily reach frequency, effectively avoiding repeated customer reach and reducing customer disturbance during marketing, while tracking reach effects to adjust marketing strategies in a timely manner, significantly improving the collaboration and efficiency of multi-channel marketing, optimizing customer experience, reducing customer complaint probability, and providing a more scientific and efficient marketing method for banks and other financial institutions.

[0028] As a preferred embodiment, the multi-channel marketing collaboration module performs the following steps when working: S21: the intelligent marketing system receives a customer list and feature importance scores; S22: analyze customer behavior characteristics using data analysis algorithms, and the association rule support degree formula is: wherein represents the support degree of the association rule , represents the event and events Probability of co-occurrence; S23: According to the analysis results, formulate targeted marketing strategies and channel touch methods. For high-frequency consumption customers, coupon distribution strategies can be formulated; for potential customers, new product promotion strategies can be formulated; S24: Pre-set fuse threshold to limit the frequency of daily touch to prevent excessive disturbance to customers. Set the daily touch frequency not more than 3 times; S25: Touch customers through multi-channel coordination mechanism. Activity placement considerations include channel priority, touch time and touch content; S26: Track the touch effect of marketing activities, and adjust marketing strategies and activity placement methods in a timely manner according to the tracking results.

[0029] As described above, the working steps of the multi-channel marketing coordination module significantly improve the coordination of marketing activities and customer experience. First, the intelligent marketing system receives customer lists and feature importance scores to provide data support for subsequent formulation of targeted marketing strategies. Then, the system uses data analysis algorithms to analyze customer behavior characteristics, helping marketing personnel better understand customer needs and preferences. Based on this, targeted marketing strategies and channel touch methods are formulated according to the analysis results to ensure that marketing activities can accurately touch target customers. At the same time, a fuse threshold is pre-set to limit the frequency of daily touch, effectively avoiding excessive disturbance to customers. Through multi-channel coordination mechanism to touch customers, ensure the coordination of marketing activities among different channels, improve the marketing effect. Finally, track the touch effect of marketing activities, and adjust marketing strategies and activity placement methods in a timely manner according to the tracking results, forming a closed-loop feedback mechanism to continuously optimize marketing activity effectiveness.

[0030] As a preferred embodiment, the marketing effect evaluation and optimization module includes: A41: Effect evaluation unit, including marketing effect index library, real-time calculation engine and effect visualization module, for real-time calculation of marketing activity effect index; A42: Strategy optimization unit, including optimization algorithm library, strategy generation module and strategy test environment, for automatically adjusting marketing strategies according to evaluation results; A43: Feedback loop unit, including feedback collection module, feedback analysis module and strategy update module, for forming a closed-loop feedback mechanism to continuously collect and analyze marketing activity feedback.

[0031] As described above, the marketing effect evaluation and optimization module provides real-time feedback to marketing personnel by calculating marketing campaign effectiveness indicators in real time, enabling them to quickly adjust strategies to respond to market changes. Meanwhile, the module uses an optimization algorithm library and strategy generation module to achieve automated optimization of marketing strategies, significantly improving the efficiency and effectiveness of marketing campaigns. In addition, by forming a closed-loop feedback mechanism and continuously collecting and analyzing marketing campaign feedback, the module helps to continuously optimize marketing strategies, making marketing decisions more scientific and precise, and thus improving the marketing input-output ratio.

[0032] As a preference, the marketing effect evaluation and optimization module includes the following steps when working: S31: After the campaign ends, the intelligent marketing system automatically collects pre-buried data on relevant pages, including customer behavior data and transaction data; S32: Generate a visual statistical analysis report to show campaign effectiveness and customer feedback, using descriptive statistics and hypothesis testing for statistical analysis methods; S33: Evaluate the effectiveness of the marketing campaign based on the results of the statistical analysis report, calculate the return on investment formula, and determine whether the marketing campaign has achieved the expected results based on the evaluation results. The return on investment formula is: Where, represents the return on investment, represents the income generated by the marketing campaign, represents the cost of the marketing campaign; S34: Based on the evaluation results, provide suggestions for the next marketing strategy for different customer groups. For customer groups with high response rates, marketing efforts can be increased. For customer groups with low response rates, marketing strategies or channels can be adjusted. Strategy optimization factors include customer segmentation, channel effectiveness, and marketing content effectiveness; S35: Provide a path for quickly deploying related activities to adjust and optimize subsequent marketing campaigns in a timely manner. Activity deployment factors include optimized strategies, channel priorities, and touch time.

[0033] As mentioned above, the working steps of the marketing effectiveness evaluation and optimization module significantly improve the effectiveness and efficiency of marketing activities. First, after the end of the activity, the system automatically collects pre-buried data from relevant pages, including customer behavior data and transaction data, providing a comprehensive and accurate basis for subsequent evaluation. Then, the system generates a visual statistical analysis report to intuitively display the activity effect and customer feedback, and uses statistical analysis methods such as descriptive statistics and hypothesis testing to conduct an in-depth analysis of the activity effect. On this basis, the system evaluates the effectiveness of the marketing activities based on the results of the statistical analysis report, and quantifies it through the return on investment formula, providing a scientific basis for strategy adjustments. Finally, based on the evaluation results, the system gives recommendations for the next marketing strategy for different customer groups, and provides a path for quickly deploying related activities to facilitate timely adjustment and optimization of subsequent marketing activities.

[0034] Preferably, the data security and privacy protection module includes: A51: Data encryption unit, including encryption algorithm library, key management module and encryption performance optimization tool, used to encrypt sensitive data; A52: Access control unit, including identity authentication module, rights management module and audit log module, is used to implement strict access control policies and record data access and operation logs; A53: Audit tracking unit, including audit rule engine, audit log analysis module and violation alarm module, is used to audit and track data access and operations, and promptly detect and handle violations.

[0035] As mentioned above, the data security and privacy protection module ensures the security of sensitive data during sharing and use through data encryption, access control, and audit tracking. The module uses advanced encryption algorithms to encrypt sensitive data to prevent data leakage and illegal access. At the same time, through strict access control policies, the module restricts access to sensitive data to authorized personnel and records data access and operation logs for easy auditing and tracing. In addition, the audit tracking unit conducts comprehensive audit tracking of data access and operations, promptly discovers and handles violations, ensures compliance of data use, reduces data privacy and compliance risks, protects customer privacy, and provides reliable data security for intelligent marketing decision-making systems.

[0036] Preferably, the customer feedback and complaint handling module includes: A61: Feedback collection unit, including a multi-channel feedback interface, a feedback data integration module, and a feedback classification tool, used to collect customer feedback through multiple channels; A62: a complaint classification unit, including a natural language processing algorithm library, a complaint classification model, and a classification result visualization module, for automatically classifying and prioritizing complaints; A63: a complaint processing unit, including a complaint assignment module, a processing progress tracking module, and a processing result feedback module, for assigning complaints to corresponding teams for processing, tracking processing progress, and feeding back processing results to customers.

[0037] As described above, the customer feedback and complaint processing module responds to customer feedback and handles customer complaints in a timely manner, improving customer satisfaction and loyalty. The module collects customer feedback through multi-channel feedback interfaces, including online surveys and customer service systems, ensuring the comprehensiveness and timeliness of customer feedback. Meanwhile, the natural language processing algorithm automatically classifies and prioritizes complaints, helping customer service teams quickly locate and handle problems. In addition, the complaint processing unit assigns complaints to corresponding teams for processing and tracks processing progress, ensuring that problems are resolved in a timely manner, optimizing customer experience, reducing customer complaints, and improving customer satisfaction.

[0038] The embodiments of the present application are described in detail above in conjunction with the accompanying drawings, but the present application is not limited to the above-described embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the spirit of the present application.

Claims

1. A dynamic customer segmentation based intelligent marketing decision system; characterized by: Comprise: Data collection and preprocessing module: to solve the problem of data timeliness, real-time collection and preprocessing of customer data; Dynamic customer segmentation module: to improve marketing response rate, through dynamic segmentation to adapt to customer behavior changes; Multi-channel marketing synergy module: to improve the synergy of multi-channel marketing, avoid repeated touch; Marketing effect evaluation and optimization module: to shorten the marketing effect evaluation cycle, real-time feedback optimization strategy; Data security and privacy protection module: to reduce the risk of data privacy and compliance, ensure data security, protect customer privacy; Customer feedback and complaint handling module: to reduce customer complaints, improve customer satisfaction, optimize customer experience.

2. The intelligent marketing decision system based on dynamic customer segmentation of claim 1, wherein: Data collection and preprocessing module includes: A11: Real-time data collection unit, including stream processing engine, data interface adapter and data cache component, used to capture customer behavior data in real time through stream processing technology; A12: Data cleaning unit, including data cleaning algorithm library, outlier detection module and data completion tool, used to clean the collected raw data, remove noise and outliers, and complete missing data; A13: Data standardization unit, including data format converter, data mapping table and data validation rule engine, used to convert the cleaned data into a unified format.

3. The intelligent marketing decision system based on dynamic customer segmentation of claim 1, wherein: Dynamic customer segmentation module includes: A21: Feature extraction unit, including machine learning algorithm library, feature storage database and feature update module, used to extract customer features using machine learning algorithms; A22: Clustering algorithm unit, including deep clustering algorithm library, algorithm parameter configuration tool and clustering result visualization module, used to apply deep clustering algorithm to dynamically segment customers, and adjust the segmentation results in real time according to customer features and behavior changes; A23: Segmentation update unit, including segmentation update strategy engine, update frequency configuration tool and segmentation version management module, used to update the segmentation results regularly.

4. The intelligent marketing decision system based on dynamic customer segmentation of claim 3, wherein: The dynamic customer segmentation module includes the following steps when working: S11: The business management personnel uploads the relevant business raw data to the front-end server through the system front-end interface; S12: The server automatically checks the data format and content, if the verification is passed, the data format standardization processing is carried out, and the related data is transmitted to the data analysis system; if the verification fails, a pop-up window is prompted on the front-end page to require re-uploading data; S13: The data analysis system scans the file transmission directory in real time, if there is data file, the file is stored in the data processing queue for waiting; S14: According to the pre-set time task, the first data in the waiting queue is collected and stored in the data lake; S15: Clean the data, including outlier rejection and missing value filling processing, standardize the data after cleaning, and use the mean filling method to handle missing values; S16: Extract customer features using machine learning algorithms, select important features through feature selection algorithms; S17: Apply deep clustering algorithm to dynamically segment customers, divide customers into different groups according to algorithm results; S18: Regularly update the clustering results, update the clustering results every hour or every day, and store the updated clustering results in the database.

5. The intelligent marketing decision system based on dynamic customer segmentation of claim 1, wherein: The multi-channel marketing synergy module includes: A31: A channel management unit including a channel interface adapter, a channel state monitoring module, and a channel performance evaluation tool, for integrating and managing different marketing channels; A32: A synergy strategy unit including a cross-channel marketing synergy strategy library, a strategy configuration tool, and a strategy execution engine, for formulating cross-channel marketing synergy strategies; A33: A reach control unit including a reach frequency limit module, a reach content management module, and a reach effect tracking module, for implementing a fuse mechanism to limit single-day reach frequency and avoid disturbing customers, while tracking reach effects.

6. The intelligent marketing decision system based on dynamic customer segmentation of claim 5, wherein: The multi-channel marketing synergy module includes the following steps when working: S21: The intelligent marketing system receives customer lists and feature importance scores; S22: Analyze customer behavior characteristics using data analysis algorithms; S23: Based on the analysis results, formulate targeted marketing strategies and channel reach methods. For high-frequency consumption customers, coupon distribution strategies can be developed; for potential customers, new product promotion strategies can be developed; S24: Pre-set a fuse threshold to limit single-day reach frequency to prevent excessive disturbance to customers, and set the single-day reach frequency to no more than 3 times; S25: Reach customers through multi-channel synergy mechanisms, considering factors such as channel priority, reach time, and reach content; S26: Track the reach effects of marketing activities and adjust marketing strategies and activity placement methods in a timely manner based on tracking results.

7. The intelligent marketing decision system based on dynamic customer segmentation of claim 1, wherein: The marketing effect evaluation and optimization module includes: A41: An effect evaluation unit including a marketing effect index library, a real-time calculation engine, and an effect visualization module, for real-time calculation of marketing activity effect indicators; A42: A strategy optimization unit including an optimization algorithm library, a strategy generation module, and a strategy test environment, for automatically adjusting marketing strategies based on evaluation results; A43: A feedback loop unit including a feedback collection module, a feedback analysis module, and a strategy update module, for forming a closed-loop feedback mechanism to continuously collect and analyze marketing activity feedback.

8. The intelligent marketing decision system based on dynamic customer segmentation of claim 7, wherein: The marketing effect evaluation and optimization module includes the following steps when working: S31: After the activity ends, the intelligent marketing system automatically collects pre-buried data on relevant pages, including customer behavior data and transaction data; S32: Generate a visual statistical analysis report to show activity effects and customer feedback, using statistical analysis methods including descriptive statistics and hypothesis testing; S33: Evaluate the effectiveness of marketing activities based on the results of the statistical analysis report, using the return on investment formula, and determine whether the marketing activities have achieved the expected results based on the evaluation results; S34: Based on the evaluation results, provide suggestions for the next marketing strategy for different customer groups. For high-response groups, marketing efforts can be increased; for low-response groups, marketing strategies or channels can be adjusted, with strategy optimization considering factors such as customer segmentation, channel effectiveness, and marketing content effectiveness; S35: Provide a path for quick deployment-related activities to adjust and optimize subsequent marketing activities in a timely manner. Activity deployment considerations include optimized strategies, channel priorities, and reach times.

9. The intelligent marketing decision system based on dynamic customer segmentation of claim 1, wherein: The data security and privacy protection module includes: A51: Data encryption unit, including encryption algorithm library, key management module and encryption performance optimization tool, used for encryption processing of sensitive data; A52: Access control unit, including identity authentication module, permission management module and audit log module, used to implement strict access control strategy and record data access and operation logs; A53: Audit tracking unit, including audit rule engine, audit log analysis module and violation alarm module, used for audit tracking of data access and operation, timely discovery and handling of violation behavior.

10. The intelligent marketing decision system based on dynamic customer segmentation of claim 1, wherein: The customer feedback and complaint handling module includes: A61: Feedback collection unit, including multi-channel feedback interface, feedback data integration module and feedback classification tool, used to collect customer feedback through multiple channels; A62: Complaint classification unit, including natural language processing algorithm library, complaint classification model and classification result visualization module, used for automatic classification and priority sorting of complaints; A63: Complaint handling unit, including complaint allocation module, processing progress tracking module and processing result feedback module, used to allocate complaints to corresponding teams for processing, track processing progress, and feedback processing results to customers.

Images