Marketing method based on user behavior preference

Abstract

The invention discloses a marketing method based on user behavior preference, which comprises the steps of data collection and expansion, user motivation deep analysis, personalized recommendation optimization, expanded and innovated marketing channel and user loyalty management, and user information in more dimensions is obtained by linking relative accounts under user mobile phone numbers. And after the receipts are obtained, associating the various data by identifying the common identifiers in the different data. According to the marketing method based on the user behavior preference, the information social media data and the Internet of Things data of the user are acquired, and the lottery drawing information, the official account operation and the coupon receiving behavior data of the user are extracted, so that the diversification of data sources is realized, and the marketing efficiency is improved by linking the relative account under the mobile phone number of the user. Therefore, user information in more dimensions is obtained, richer materials are provided for marketing, and deeper user behavior modes and preferences can be mined.

Description

Technical Field

[0001] The present invention relates to the technical field of marketing data processing, and in particular to a marketing method based on user behavior preference. Background Art

[0002] With the rapid development of the Internet and information technology, data has exploded, and companies are able to collect massive amounts of user behavior data. At the same time, artificial intelligence technologies, such as machine learning and deep learning, have made continuous breakthroughs, providing powerful tools and methods for analyzing and mining these data, allowing companies to achieve precise personalized marketing based on user behavior preferences. Traditional marketing methods are mainly product-centric, attracting consumers through large-scale advertising and promotional activities. However, with the intensification of market competition and the diversification of consumer demand, the effectiveness of this marketing method has gradually weakened. Companies have begun to realize that only by deeply understanding user needs and preferences and providing personalized products and services can they improve user satisfaction and loyalty, thereby standing out in market competition.

[0003] Common marketing methods based on user behavior preferences first collect user behavior data through websites and APP platforms, including browsing records, purchase records, search records and collection records, and use data analysis tools and technologies to conduct in-depth analysis of the collected user behavior data to explore user behavior patterns, interest preferences, consumption habits, etc. For example, through association rule mining, the association relationship between the products purchased by users is discovered, and then a label system is established for users based on the results of data analysis, and finally corresponding products are pushed to users based on the analyzed structure; However, when obtaining user data, this method only collects browsing records, purchase records, search records and collection records, and then pushes marketing based on the obtained user behavior data, which can easily lead to misjudgment of user intentions and makes it difficult to deeply explore the user's real needs and preferences. As a result, the recommended content may be broad and superficial, and cannot meet the user's expectations for personalized experience. At the same time, it is impossible to fully consider the various factors that affect the user's purchasing decision, such as emotional factors, social factors, etc. Marketing push based only on these types of behavioral data is difficult to impress users, resulting in a low conversion rate and cannot meet the requirements of marketing data processing. For this reason, a marketing method based on user behavior preferences is proposed. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a marketing method based on user behavior preferences to solve the technical problem of simply collecting browsing records, purchase records, search records and collection records, and then pushing marketing through the obtained user behavior data, which can easily lead to misjudgment of user intentions and makes it difficult to deeply explore the user's real needs and preferences.

[0005] To achieve the above objectives, the present invention provides the following technical solutions: A marketing method based on user behavior preferences, comprising the following steps: S1 Data collection expansion Collect user behavior data through websites and APP platforms, including browsing records, purchase records, search records, and favorite records, and obtain user preferences through questionnaires and user interviews. At the same time, obtain the user's information social media data to reveal the user's social relationships, interest dissemination, and word-of-mouth influence, and obtain the user's Internet of Things data to reflect the user's life scenarios and consumption correlations. Extract the lottery information, public account operations, and coupon-receiving behavior data of user greetings to display the user's consumption tendencies and participation from different perspectives. Finally, link the relative accounts under the user's mobile phone number to obtain more-dimensional user information. After obtaining the receipts, associate various types of data by identifying the common identifiers in different data, and then store the collected data in the enterprise's data storage repository to provide rich materials for subsequent in-depth analysis; S2 In-depth analysis of user motivation Obtain various types of user data from the data storage repository, perform data preprocessing including data standardization and missing value processing, first use association rule mining and classification algorithms, set algorithm parameters and run the algorithms, and then analyze and interpret the algorithm results to find the factors affecting the user's purchase decision, and store the analysis results back to the data storage repository; S3 Optimization of personalized recommendations According to the results of the in-depth analysis of user motivation, obtain multi-form content resources from the content management system, input the user behavior data and content features into the recommendation algorithm, screen and sort the recommended content, and then generate recommended reasons. Optimize the display effect of the recommended content and reasons together, and then display it to the user through the user interface. At the same time, receive user feedback and optimize and adjust the recommendation algorithm; S4 Expansion and innovation of marketing channels For emerging marketing channels including VR, AR, and live streaming, first conduct market research and technical evaluation. When choosing VR and AR, it is necessary to design, develop, and test the product experience scenarios, and deploy the marketing content to VR and AR devices. When choosing live streaming marketing, it is necessary to first determine the live streaming theme, content, and anchor, and then carry out the preparatory work before the live stream. During the marketing activity, adjust the marketing strategy in a timely manner according to user feedback. After the activity ends, conduct effect evaluation and summary. In the programming of VR and AR, use Unity3D and UnrealEngine for programming, create virtual models of products, design interactive scenarios, and implement the interactive functions between users and goods by writing code; S5 User loyalty management Regularly obtain the historical behavior data of users from the user relationship management system, input this data into the user loyalty prediction model, and classify users into different levels of churn risk according to the results of the prediction model. For users in different levels, formulate corresponding loyalty improvement strategies including exclusive offers and personalized services, and through the operation of the user community, promote communication and interaction among users, enhance user belonging and loyalty, and at the same time adjust and optimize the loyalty improvement strategies according to the actual responses of users.

[0006] Preferably, use web crawlers to target web browsing records, search engine APIs to obtain search records, social media platform APIs to collect social media data, Internet of Things device interfaces to obtain Internet of Things data, and software development kits to collect specific user behavior data, and synchronously interact with the enterprise's marketing data management system, transmit the collected data to the data repository, and interact with the data verification and cleaning module in the data receipt extension. When the format of the collected data does not meet the requirements, the data verification module will prompt the collection tool to re-collect and perform format conversion, and then store the cleaned data in the database.

[0007] Preferably, for the relative accounts under the linked user mobile phone number in the data collection extension, the following steps are included: S11 Data collection and integration Collect the relative relationship information under the user's mobile phone number, including parents, spouses and children, through user registration, questionnaire surveys and customer service communication, and obtain their basic information, including age, gender and occupation. Then integrate the collected relative information with the user's other behavior data and consumption data to form a comprehensive user data set; S12 Feature extraction and tagging Extract valuable features based on the basic information and behavior data of relatives, and at the same time perform tagging processing on the extracted features; S13 Portrait construction and analysis Construct a user portrait based on the label information of the user and their relatives, and find out the user's behavior patterns and preferences by analyzing the behavior data of the user and their relatives; S14 Personalized recommendation and marketing Provide personalized product recommendations and services for users based on the user portrait and behavior analysis results, and at the same time formulate personalized marketing plans for different user groups and their relative relationships; S15 Dynamic update and optimization By real-time monitoring the changes in the behavior data and information of the user and their relatives, update the user portrait in a timely manner, and at the same time continuously optimize the construction model and algorithm of the user portrait according to user feedback and data analysis results to improve the accuracy and effectiveness of the portrait.

[0008] Preferably, in the in-depth analysis of user motivation, a data analysis algorithm is implemented through programming. The data is preprocessed and cleaned using the data analysis library of Python and Pandas, while Numpy is used for numerical calculations. For association rule mining, the Apriori algorithm in the mlxtend library can be used to implement it, and for classification algorithms, they are implemented through the Scikit-learn library, including decision trees and logistic regression. Moreover, during the programming process, the parameters of the algorithms need to be adjusted and optimized to improve the accuracy of the analysis results.

[0009] Preferably, the personalized recommendation optimization obtains content resources of images, videos, and audios from the content management system, and according to the results of in-depth analysis, sends requests to the CMS to obtain content that meets the user's interests. At the same time, it interacts with the recommendation algorithm module, inputs the user behavior data and content features into the recommendation algorithm, obtains the recommendation results and reasons, and then feeds back this information to the CMS for display to the user. The content management system interacts with the user interface software to display the personalized recommended content and reasons to the user, and receives the feedback from the user on the UI, including clicks and browsing durations, and transmits this feedback information to the recommendation algorithm module for optimizing subsequent recommendations.

[0010] Preferably, for user loyalty management, it includes the following steps: S51 Data collection and integration First, collect user data from multiple channels, including the user's browsing history, purchase behavior, participation in activities, and social media interactions on the platform, and integrate them to form a comprehensive user behavior dataset; S52 Feature extraction and engineering Extract key features related to loyalty from the dataset, including purchase frequency, purchase amount, the time of the last purchase, and user activity, and perform data transformation and engineering processing for constructing new features; S53 Model selection and training Divide the processed user data into a training set, a validation set, and a test set for model training and optimization; S54 Loyalty prediction and evaluation Input the behavior data of new users or old users into the trained model to obtain the loyalty prediction results. At the same time, use indicators such as accuracy, recall rate, F1-score, and the area under the ROC curve to evaluate the model performance, and optimize the model according to the feedback.

[0011] Preferably, in the operation of the user community, an incentive mechanism is designed to encourage all users to participate in community activities, including holding lottery activities for ordinary users, establishing a user growth system, etc., to improve the participation enthusiasm of ordinary users, and establish clear community rules, strengthen the supervision of the community, and promptly handle violations.

[0012] In summary, compared with the prior art, the present invention provides a marketing method based on user behavior preferences, which has the following beneficial effects: 1. This marketing method based on user behavior preferences realizes the diversification of data sources by obtaining users' information social media data and Internet of Things data and extracting the behavior data of lottery information, official account operations, and coupon collection in users' greetings. It can comprehensively understand users' behavior habits, interests, social relationships, and consumption scenarios. Compared with traditional marketing methods, the data collected is no longer limited to the part where users directly interact with products, thus providing richer materials for subsequent precision marketing. By exploring the underlying logic behind users' multiple browsing, collection, and search but non-purchase behaviors, it can accurately identify the factors affecting users' purchase decisions, including price, product function, brand image, and promotional activities, which helps enterprises adjust their marketing strategies and provide more targeted products or services according to users' concerns and needs. For example, if it is found that users have browsed a certain product multiple times but did not purchase due to price, the enterprise can launch targeted price discounts or promotional activities to increase the purchase conversion rate; 2. This marketing method based on user behavior preferences obtains more dimensional user information by linking the relative accounts under the user's mobile phone number, enabling the collection of relevant data of their relatives in addition to the user's own behavior data, including consumption habits, interests, and social behaviors. These data can more comprehensively depict the user's family consumption portrait and social relationship network, providing richer materials for marketing. This way of analyzing the associated behaviors and interaction data between relative accounts can explore deeper user behavior patterns and preferences, including discovering the mutual influence of users and their relatives in purchase decisions, common interests, and consumption tendencies, thus providing more targeted marketing strategies. At the same time, through the analysis of relative account data, new marketing opportunities and potential user needs will be found, and combined with the characteristics of relative accounts and user behavior preferences, new marketing methods and means can be explored, such as carrying out family-themed marketing activities, launching exclusive preferential packages or services for relatives, attracting the participation of users and their relatives, and improving the innovation and attractiveness of marketing. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 is the flow chart of the marketing method of the present invention; Figure 2 is the extended flow chart of data collection of the present invention; Figure 3 is the flow chart of user loyalty management of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0015] Please refer to Figure 1 , a marketing method based on user behavior preferences, comprising the following steps: S1 Data collection expansion Collect users' behavioral data through websites and APP platforms, including browsing records, purchase records, search records, and favorite records. First, collecting users' behavioral data from websites and APP platforms is an important start of the entire marketing process. Among them, browsing records are the footprints left by users in the digital world. Every page view, whether it is a quick glance or a long stay, contains potential information. For example, if a user stays on a product page for a long time, it may indicate a strong interest in the product; frequently browsing specific types of articles or pages can also reflect the user's preferences in this regard. Purchase records are a direct manifestation of users' actual consumption behaviors. It not only shows the types of products or services purchased by users, but also reflects the frequency, amount, and time nodes of consumption, etc. And obtain users' preferences through questionnaires and user interviews. At the same time, obtain users' information. Social media data is used to reveal users' social relationships, interest dissemination, and word-of-mouth influence, and obtain users' Internet of Things data, which is used to reflect users' life scenarios and consumption correlations. And extract the lottery information, public account operations, and coupon-receiving behavioral data of users' greetings, which are used to show users' consumption tendencies and participation from different angles. Data receipt expansion uses web crawlers for web browsing records, search engine APIs to obtain search records, social media platform APIs to collect social media data, Internet of Things device interfaces to obtain Internet of Things data, and software development kits to collect specific user behavioral data, and synchronously interact with the enterprise's marketing data management system, transfer the collected data to the data repository, interact with the data verification and cleaning module in the data receipt expansion. When the format of the collected data does not meet the requirements, the data verification module will prompt the collection tool to re-collect and perform format conversion, and then store the cleaned data in the database. Finally, by linking the relative accounts under the user's mobile phone number, more dimensional user information can be obtained. After obtaining the receipts, associate various types of data by identifying the common identifiers in different data, and then store the collected data in the enterprise's data repository to provide rich materials for subsequent in-depth analysis. In addition to collecting these basic behavioral data from the platform, users' preferences can also be obtained through questionnaires and user interviews. Questionnaires can cover a large number of user groups. Through carefully designed questions, investigate from multiple dimensions such as users' expectations for product functions, requirements for service quality, and perceptions of brand images. For example, questions like "When choosing a mobile phone, which functions do you value the most (such as camera, battery life, performance, etc.)" can be set in the questionnaire to understand users' preferences for product functions. User interviews are more in-depth and personalized, and conduct targeted exchanges for different types of users. Some loyal users can be selected to understand the reasons for their continuous support for the brand, or some potential users can be selected to explore the concerns that prevent them from purchasing the product. And linking the relative accounts under the user's mobile phone number includes the following steps: Please refer toFigure 2 , S11 Data Collection and Integration Collect kinship information under the user's mobile phone number through user registration, questionnaire surveys, and customer service communication, including parents, spouses, and children, and obtain their basic information, including age, gender, and occupation. Then integrate the collected kinship information with other behavioral data and consumption data of the user to form a comprehensive user data set. In this process, collecting kinship information under the user's mobile phone number through various means is an important step in building a comprehensive user data set. First, the user registration process is an opportunity to collect kinship information. Relevant fields can be designed on the registration page to guide users to fill in kinship information such as parents, spouses, and children, and at the same time encourage users to provide basic information about these relatives, such as age, gender, and occupation. For example, a drop-down menu can be set for users to select kinship, and corresponding text boxes can be provided for users to fill in basic information. Questionnaire surveys are also an effective way to obtain kinship information. Questions about family structure and kinship can be set in the questionnaire. Such questions can not only understand the situation of the elderly in the user's family but also obtain potential consumption-influencing information such as relevant age and health. S12 Feature Extraction and Labeling Extract valuable features based on the basic information and behavioral data of relatives, and at the same time perform labeling processing on the extracted features. After having a comprehensive user data set, it is necessary to mine valuable features based on the basic information and behavioral data of relatives. From the perspective of basic information, age is an important feature. People of different age groups may have obvious differences in consumption preferences. For example, young people may be more inclined to fashionable and trendy products, while the elderly may be more concerned about health and wellness products. Gender is also a factor affecting consumption. Men and women often have different preferences in clothing, beauty products, electronic products, etc. Occupation also affects consumption. For example, people engaged in outdoor work may have a higher demand for sunscreen and wear-resistant products. From the aspect of behavioral data, valuable features can also be extracted from the purchase frequency and product types purchased by relatives. If a relative often buys organic food, this may indicate their pursuit of a healthy lifestyle. S13 Portrait Construction and Analysis Based on the label information of the user and their relatives, construct a user profile, and by analyzing the behavior data of the user and their relatives, find out the user's behavior patterns and preferences. Constructing a user profile based on the label information of the user and their relatives is a process of transforming abstract data into a specific user image. By comprehensively analyzing various labels of the user and their relatives, a three-dimensional user image is depicted. For example, a user labeled as a "fashionable youth" has relatives engaged in the fashion industry, and they all have a relatively high consumption frequency on fashion brands. Then this user profile may be a young consumer living in a family with a strong fashion atmosphere, having a keen perception and strong pursuit of fashion trends, with relatively strong consumption ability, and may be influenced by relatives in fashion consumption decisions; S14 Personalized Recommendation and Marketing Based on the user profile and the results of behavior analysis, provide personalized product recommendations and services for the user. At the same time, formulate personalized marketing plans for different user groups and their kinship; S15 Dynamic Update and Optimization By real-time monitoring of the behavior data and information changes of the user and their relatives, update the user profile in a timely manner. At the same time, according to user feedback and the results of data analysis, continuously optimize the construction model and algorithm of the user profile to improve the accuracy and effectiveness of the profile; S2 In-depth Analysis of User Motivation Obtain various types of user data from the data repository, perform data preprocessing including data standardization and missing value processing, and first use association rule mining and classification algorithms, set algorithm parameters and run the algorithms, and then analyze and interpret the algorithm results to find out the factors affecting the user's purchase decision, and store the analysis results back into the data repository. In the in-depth analysis of user motivation, programming is used to implement data analysis algorithms. Through the data analysis library of Python and Pandas for data preprocessing and cleaning, while Numpy is used for numerical calculations. For association rule mining, the Apriori algorithm in the mlxtend library can be used to implement it, and for classification algorithms, they are implemented through the Scikit-learn library, including decision trees and logistic regression. And during the programming process, the parameters of the algorithms need to be adjusted and optimized to improve the accuracy of the analysis results; S3 Optimization of Personalized Recommendation Based on the results of in-depth user motivation analysis, obtain multi-form content resources from the content management system. Input user behavior data and content features into the recommendation algorithm for screening and ranking of recommended content, and then generate recommendation reasons. Optimize the display effect of the recommended content and reasons together, and then display them to the user through the user interface. At the same time, receive user feedback and optimize and adjust the recommendation algorithm. Personalized recommendation optimization obtains image, video, and audio content resources from the content management system, and according to the results of in-depth analysis, sends requests to the CMS to obtain content that meets the user's interests. At the same time, interact with the recommendation algorithm module, input user behavior data and content features into the recommendation algorithm, obtain recommendation results and reasons, and then feedback this information to the CMS for display to the user. The content management system interacts with the user interface software to display the personalized recommended content and reasons to the user, and receives user feedback on the UI including clicks and browsing durations, and transmits this feedback information to the recommendation algorithm module for optimizing subsequent recommendations; S4 Expanding and Innovating Marketing Channels For emerging marketing channels including VR, AR, and live streaming, first conduct market research and technology assessment. When choosing VR and AR, it is necessary to design, develop, and test product experience scenarios, and deploy marketing content to VR and AR devices. When choosing live streaming marketing, it is necessary to first determine the live streaming theme, content, and host, and then carry out preparatory work before the live stream. During the marketing activity, adjust the marketing strategy in a timely manner according to user feedback. After the activity ends, conduct effect evaluation and summary. In the programming of VR and AR for expanding and innovating marketing channels, use Unity3D and UnrealEngine for programming, and create virtual models of products, design interactive scenarios, and implement the interactive functions between users and goods by writing code; S5 User Loyalty Management Regularly obtain the historical behavior data of users from the user relationship management system, input these data into the user loyalty prediction model, and according to the results of the prediction model, classify users into different levels of churn risk. And user loyalty management includes the following steps: Please refer to Figure 3 , S51 Data Collection and Integration First collect user data from multiple channels, including the user's browsing history, purchase behavior, participation in activities, and social media interactions on the platform, and integrate them to form a comprehensive user behavior dataset; S52 Feature Extraction and Engineering Extract key features related to loyalty from the dataset, including purchase frequency, purchase amount, the time of the last purchase, and user activity, and perform engineering processes such as data transformation and construction of new features; S53 Model Selection and Training Divide the processed user data into training set, validation set and test set for model training and optimization; S54 Loyalty Prediction and Evaluation Input the behavior data of new or existing users into the trained model to obtain loyalty prediction results. At the same time, use metrics such as accuracy, recall, F1-score and area under the ROC curve to evaluate the model performance, optimize the model according to the feedback. For users of different levels, formulate corresponding loyalty improvement strategies including exclusive offers and personalized services, and through the operation of the user community, promote communication and interaction among users, enhance user sense of belonging and loyalty. At the same time, adjust and optimize the loyalty improvement strategies according to the actual reactions of users. In the operation of the user community, design an incentive mechanism to encourage all users to participate in community activities, including holding lucky draw activities for ordinary users, establishing a user growth system, etc., to improve the participation enthusiasm of ordinary users, and establish clear community rules, strengthen the supervision of the community, and promptly handle violations.

[0016] Through a series of planning steps, this solution has achieved that the collection of user data is no longer limited to the part where users directly interact with the product, thus providing richer materials for subsequent precision marketing. By exploring the underlying logic behind users' behaviors of browsing, collecting and searching but not purchasing multiple times, it is possible to accurately identify the factors affecting users' purchase decisions, including price, product function, brand image and promotional activities, which helps enterprises adjust their marketing strategies and provide more targeted products or services according to users' concerns and needs. For example, if it is found that a user has browsed a certain product multiple times but did not purchase due to the price, the enterprise can launch targeted price discounts or promotional activities to increase the purchase conversion rate.

[0017] The core of this solution lies in obtaining users' information from social media data and Internet of Things data, extracting the behavior data of users' lottery information, public account operations and coupon collection, and linking the relative accounts under the user's mobile phone number, achieving the diversification of data sources, which helps enterprises adjust their marketing strategies and thus obtain more dimensional user information. In addition to the behavior data of the users themselves, relevant data of their relatives can also be collected, including consumption habits, hobbies and social behaviors. These data can more comprehensively depict the user's family consumption portrait and social relationship network, providing richer materials for marketing. This way of analyzing the associated behaviors and interaction data between relative accounts can explore deeper user behavior patterns and preferences, including discovering the mutual influence of users and their relatives in purchase decisions, common hobbies and consumption tendencies, thus providing more targeted strategies for marketing.

[0018] In summary, through a series of innovative data acquisition processes and management strategies, this solution has successfully achieved the diversification of data sources, enabling a more comprehensive understanding of users' behavior patterns, interests, social relationships, and consumption scenarios. Compared with traditional marketing methods, the data collected is no longer limited to the part where users directly interact with the product, thus providing richer materials for subsequent precision marketing. At the same time, by analyzing the data of related accounts, new marketing opportunities and potential user needs can be discovered. Moreover, combined with the characteristics of related accounts and users' behavior preferences, new marketing methods and means can be explored, including carrying out family-themed marketing activities, launching exclusive preferential packages or services for relatives, attracting the participation of users and their relatives, and improving the innovation and attractiveness of marketing.

[0019] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0020] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A marketing method based on user behavior preferences, characterized by: The following steps are involved: S1 Data Collection Extension Collect user behavior data through websites and APP platforms, including browsing history, purchase history, search history and collection history, and obtain user preferences through questionnaires and user interviews. At the same time, obtain user information social media data to reveal users' social relationships, interest dissemination and word-of-mouth influence, and obtain users' IoT data to reflect users' life scenarios and consumption associations. By extracting users' lucky draw information, public account operations and coupon collection behavior data, it is used to show users' consumption tendencies and participation from different angles. Finally, by linking the relatives' accounts under the user's mobile phone number, more dimensional user information can be obtained. After obtaining the receipt, various types of data are associated by identifying common identifiers in different data, and then the collected data is stored in the company's data repository to provide rich materials for subsequent in-depth analysis; In-depth analysis of S2 user motivations Obtain various types of user data from the data repository, perform data preprocessing including data standardization and missing value processing, and first use association rule mining and classification algorithms and set algorithm parameters and run the algorithm, then analyze and interpret the algorithm results to find out the factors that affect user purchase decisions, and store the analysis results back to the data repository; S3 personalized recommendation optimization Based on the results of in-depth analysis of user motivations, we obtain multi-format content resources from the content management system, input user behavior data and content features into the recommendation algorithm, filter and sort the recommended content, and then generate recommendation reasons. We optimize the display effect of the recommended content and reasons together, and then present them to users through the user interface. At the same time, we receive user feedback and optimize the recommendation algorithm. S4 expands and innovates marketing channels For emerging marketing channels including VR, AR and live broadcast, first conduct market research and technical evaluation. When choosing VR and AR, it is necessary to design, develop and test product experience scenarios and deploy marketing content on VR and AR devices. When choosing live broadcast marketing, it is necessary to determine the live broadcast theme, content and anchor first, and then conduct preparatory work before the live broadcast. During the marketing activities, timely adjust the marketing strategy according to user feedback, and conduct effect evaluation and summary after the activities. S5 User Loyalty Management Regularly obtain users' historical behavior data from the user relationship management system, input this data into the user loyalty prediction model, and classify users into different churn risk levels based on the results of the prediction model. Develop corresponding loyalty enhancement strategies for users of different levels, including exclusive discounts and personalized services. Through the operation of the user community, promote communication and interaction between users, enhance user sense of belonging and loyalty, and adjust and optimize the loyalty enhancement strategy based on the actual response of users.

2. The marketing method based on user behavior preference according to claim 1, characterized in that: The data collection extension uses web crawlers to target web browsing records, search engine APIs to obtain search records, social media platform APIs to collect social media data, IoT device interfaces to obtain IoT data, and software development kits to collect specific user behavior data, and synchronously interact with the enterprise's marketing data management system to transfer the collected data to a data repository.

3. The marketing method based on user behavior preference according to claim 1, characterized in that: The data collection extension interacts with the data verification and cleaning module. When the format of the collected data does not meet the requirements, the data verification module will prompt the collection tool to re-collect and convert the format, and then store the cleaned data in the database.

4. The marketing method based on user behavior preference according to claim 1, characterized in that: The linking of the relative accounts under the user's mobile phone number comprises the following steps: S11 Data Collection and Integration Through user registration, questionnaires and customer service communication, we collect the family relationship information under the user's mobile phone number, including parents, spouse and children, and obtain their basic information, including age, gender and occupation. We then integrate the collected family information with the user's other behavioral data and consumption data to form a comprehensive user data set; S12 feature extraction and labeling According to the basic information and behavior data of relatives, valuable features are extracted and labeled; S13 Portrait Construction and Analysis Build user portraits based on the tag information of users and their relatives, and find out the user's behavior patterns and preferences by analyzing the behavior data of users and their relatives; S14 Personalized recommendation and marketing Provide users with personalized product recommendations and services based on user portraits and behavior analysis results, and develop personalized marketing plans for different user groups and their relationships; S15 dynamic update and optimization By real-time monitoring of the behavioral data and information changes of users and their relatives, user portraits are updated in a timely manner. At the same time, based on user feedback and data analysis results, the user portrait construction model and algorithm are continuously optimized to improve the accuracy and effectiveness of the portrait.

5. The marketing method based on user behavior preference according to claim 1, characterized in that: In the in-depth analysis of user motivation, programming is used to implement the data analysis algorithm, through Python's data analysis library, and Pandas for data preprocessing and cleaning, and Numpy for numerical calculations. For association rule mining, the Apriori algorithm in the mlxtend library can be used, and the classification algorithm is implemented through the Scikit-learn library, including decision trees and logistic regression. In the programming process, the parameters of the algorithm need to be adjusted and optimized to improve the accuracy of the analysis results.

6. The marketing method based on user behavior preference according to claim 1, characterized in that: The personalized recommendation optimization obtains image, video and audio content resources from the content management system, and based on the results of in-depth analysis, sends a request to the CMS to obtain content that meets the user's interests. At the same time, it interacts with the recommendation algorithm module, inputs user behavior data and content features into the recommendation algorithm, obtains recommendation results and reasons for recommendation, and then feeds this information back to the CMS for display to the user.

7. The marketing method based on user behavior preference according to claim 6, characterized in that: The content management system interacts with the user interface software to present personalized recommended content and reasons to the user, and receives user feedback on the UI including clicks and browsing time, and passes this feedback information to the recommendation algorithm module for optimizing subsequent recommendations.

8. The marketing method based on user behavior preference according to claim 1, characterized in that: In terms of VR and AR programming, the expanded and innovative marketing channels use Unity3D and UnrealEngine for programming, creating virtual models of products, designing interactive scenes and realizing interactive functions between users and products through code writing.

9. The marketing method based on user behavior preference according to claim 1, characterized in that: The user loyalty management includes the following steps: S51 Data Collection and Integration First, collect user data from multiple channels, including users’ browsing history, purchasing behavior, participation in activities, and social media interactions on the platform, and integrate them into a comprehensive user behavior data set; S52 Feature Extraction and Engineering Extract key loyalty-related features from the dataset, including purchase frequency, purchase amount, last purchase time, and user activity, and perform data transformation and engineering processing to build new features; S53 model selection and training Divide the processed user data into training set, validation set and test set for model training and optimization; S54 Loyalty Prediction and Evaluation The behavioral data of new or old users is input into the trained model to obtain the loyalty prediction results. At the same time, the model performance is evaluated using indicators such as accuracy, recall, F1-score and area under the ROC curve, and the model is optimized based on the feedback.

10. The marketing method based on user behavior preference according to claim 1, characterized in that: An incentive mechanism is designed in the operation of the user community to encourage all users to participate in community activities, including holding lucky draws for ordinary users and setting up a user growth system to increase the participation enthusiasm of ordinary users, and establishing clear community rules, strengthening supervision of the community, and promptly handling violations.

Images