Digital marketing system based on big data

Abstract

The invention discloses a digital marketing system based on big data. The digital marketing system comprises a data acquisition module, a data analysis module, a data planning module, a customer relationship management module and an effect evaluation module. The data acquisition uses advanced technology to collect multi-source data, the analysis module deeply mines values, the planning module customizes personalized strategies, the customer relationship management operates efficiently, and the effect evaluation module monitors in real time and adjusts the strategies according to results. All the modules work cooperatively, precision and intelligence of digital marketing are achieved, the marketing effect and efficiency are improved, and a corresponding innovative marketing method and related technologies are applied to all the steps to ensure efficient operation of the system; according to the digital marketing system, comprehensive and accurate data collection and integration, deep intelligent analysis and mining of data values, personalized marketing planning execution, efficient customer relationship management and real-time dynamic effect evaluation and optimization are realized, the marketing effect and efficiency are improved, resource waste is reduced, customer loyalty is enhanced, and innovative development of enterprise marketing is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of digital marketing systems, and more specifically, to a digital marketing system based on big data. Background Art

[0002] With the rapid development of information technology and the widespread popularity of the Internet, digital marketing has become increasingly prominent in the business field. However, traditional digital marketing methods face many challenges and bottlenecks.

[0003] In terms of data collection, previous methods can only obtain limited, superficial data, such as simple website visit statistics or basic customer transaction records. It is difficult to fully collect and deeply mine rich social media data, such as users' emotional expressions and interactive behavior details. In addition, the data formats of different data sources vary greatly, making integration difficult, resulting in the inability to form a complete, multi-dimensional customer portrait and market insight foundation.

[0004] In the data analysis phase, traditional analysis methods mainly rely on simple statistical tools and fixed analysis models, which are difficult to handle massive, complex and diverse big data. The ability to discover hidden patterns, potential connections and dynamic trends in the data is limited, and it is impossible to accurately grasp the personalized needs of customers and the ever-changing market. For example, it is impossible to accurately adjust marketing strategies based on customers' real-time behavior and interest changes.

[0005] In the field of marketing planning, traditional methods are usually based on experience and subjective judgment to design advertising and promotion activities, lacking scientific basis and precise positioning. It is difficult to formulate personalized marketing plans based on the unique characteristics and behavior patterns of different customer groups, resulting in poor marketing results and serious waste of resources. For example, advertisements may be delivered to people who are not interested, or promotional activities may fail to attract target customers.

[0006] In terms of customer relationship management, traditional methods mainly rely on manual recording and follow-up of customer information, which is inefficient and prone to errors. Customer feedback and complaints are not handled in a timely manner, and real-time maintenance and in-depth cultivation of customer relationships cannot be achieved. It is difficult to establish long-term and stable customer relationships, which affects customer loyalty and corporate reputation.

[0007] In the effect evaluation stage, traditional evaluation indicators are single and lagging, and can only perform simple data analysis after the marketing activities are completed, which is unable to monitor in real time and adjust strategies in time. The evaluation of the real effect and return on investment of marketing activities is inaccurate, making it difficult to provide effective guidance for subsequent marketing decisions.

[0008] To sum up, the existing digital marketing technology can no longer meet the needs of enterprises for precise marketing, efficient operation and sustainable development in the big data era. An innovative digital marketing system based on big data is urgently needed to solve these problems. Summary of the invention

[0009] The present invention proposes a digital marketing system based on big data, which is used to solve the problems mentioned in the above-mentioned prior art.

[0010] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a digital marketing system based on big data, including a data collection module, a data analysis module, a marketing planning module, a customer relationship management module and an effect evaluation module;

[0011] The data collection module is used to collect a wide range of marketing-related data, including social media interaction data, such as the comment sentiment tendency value S, obtained through the natural language processing sentiment analysis algorithm, S∈[-1,1], -1 represents extremely negative sentiment, 1 represents extremely positive sentiment, website user behavior trajectory data, transaction flow data, etc. It uses advanced distributed crawler technology and intelligent data interfaces to ensure comprehensive, accurate and efficient data collection. Its data collection efficiency evaluation indicators Where E collect represents the data collection efficiency, D i It represents the amount of data of different types collected per unit time, n represents the total number of data types, T collect Indicates the collection time.

[0012] The data analysis module deeply explores the value of data. It uses innovative fusion analysis algorithms and combines deep learning neural networks, such as the improved Transformer architecture network for processing text data feature extraction and reinforcement learning decision models, to dynamically adjust analysis strategies and conduct in-depth analysis of data. It divides customers into multiple precise customer groups based on their behavior patterns and preferences. j , where j represents the customer group number, and builds a customer behavior prediction model P(x), where x represents the customer behavior feature vector, for example, by analyzing historical data to predict the probability of a customer's future purchase intention. Its data analysis accuracy evaluation index is Among them A analysis Indicates the accuracy of data analysis, N data_corectly_analyzed Indicates the amount of data analyzed correctly, N data_k represents the size of the kth data sample, and m represents the total number of data samples.

[0013] The marketing planning module customizes marketing strategies based on the analysis results. This includes generating personalized advertising ideas, automatically generating advertising copy and creative materials based on customer group characteristics and prediction models, planning precise promotional activities, and formulating preferential strategies that are suitable for different customer groups, such as optimizing discounts for different levels. For example, launching exclusive customized products and services for high-value customer groups.

[0014] The customer relationship management module efficiently manages customer relationships. It includes the construction of detailed customer information files, recording basic customer information, historical interaction records, and preference tag set L = {l1, l2, ..., l n}, etc. Intelligent customer communication assistance automatically provides communication suggestions and speech templates based on the customer's historical communication style and current problems, intelligently processes customer feedback, and uses intelligent customer service robots and manual customer service to work together. The robots handle common problems and transfer complex problems to manual customer service.

[0015] The effect evaluation module monitors the marketing effect in real time. By setting key performance indicators (KPIs) such as click-through rate (C click , conversion rate C convert , Customer retention growth rate R retain Etc. to conduct quantitative evaluation. Construct a dynamic effect evaluation model M(E, t), where E represents the evaluation indicator set and t represents time. The evaluation strategy and indicator weights are adjusted in real time according to data changes. Its effect evaluation timeliness indicator Where T evaluation Indicates the timeliness of effect evaluation, T total Represents the total time of marketing activities, T delay T represents the time interval from the start of the marketing campaign to the acquisition of initial evaluation data. adapt Indicates the time required for the system to adjust the evaluation strategy based on data changes.

[0016] Preferably, the distributed crawler technology in the data collection module supports multi-threaded concurrent collection and dynamic data source adaptation, and can automatically adjust the collection strategy according to changes in the data source structure.

[0017] Preferably, the fusion analysis algorithm in the data analysis module adopts a technology based on the fusion of knowledge graph and deep learning to construct a data relationship network and mine deep associations and potential patterns.

[0018] Preferably, the advertising creative generation in the marketing planning module adopts a creative generation model based on a generative adversarial network (GAN) to automatically generate attractive advertising content.

[0019] Preferably, the intelligent communication assistance in the customer relationship management module adopts an intelligent assistant based on sentiment analysis and dialogue management technology to provide personalized communication services.

[0020] Preferably, the effect evaluation dynamic model in the effect evaluation module supports automatic learning and optimization, and continuously adjusts the evaluation strategy and indicator weights according to historical marketing data and real-time feedback.

[0021] A digital marketing method based on big data, comprising the following steps:

[0022] S1. Data collection: Use the data collection module to collect marketing-related data from multiple channels, and use data preprocessing techniques (such as data cleaning, denoising, format conversion, etc.) to preliminarily organize the data.

[0023] S2. Data analysis: Transfer the preprocessed data to the data analysis module, use innovative analysis algorithms and models to conduct in-depth analysis, and explore customer needs and market trends.

[0024] S3. Marketing planning: Based on the results of data analysis, use the marketing planning module to develop personalized marketing plans, including advertising strategies, promotional activities plans, etc.

[0025] S4. Marketing execution: Use automated marketing tools (such as marketing automation software, social media promotion tools, etc.) to put marketing plans into practice.

[0026] S5. Effect evaluation: Use the effect evaluation module to monitor and evaluate the effects of marketing activities in real time, and adjust marketing strategies and plans in a timely manner based on the evaluation results.

[0027] Preferably, in the data collection step, blockchain technology is used to ensure the secure transmission and storage of data, and to achieve data traceability and non-tamperability.

[0028] Preferably, in the data analysis step, virtual reality (VR) and augmented reality (AR) technologies are used for data visualization to provide an immersive data analysis experience.

[0029] Preferably, in the marketing execution step, artificial intelligence algorithms are used to optimize the advertising and promotion execution process to improve marketing effectiveness and efficiency.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. Comprehensive and accurate data collection and integration:

[0032] Through advanced technical means, it is possible to collect a wide range of data from multiple channels, including social media, websites, and transactions, and efficiently integrate data in different formats. For example, using distributed crawler technology and intelligent data interfaces, it is possible to obtain multi-dimensional data such as the sentiment tendency values ​​of users on social media and the detailed behavioral trajectories of website users in real time, providing a solid data foundation for building comprehensive and accurate customer portraits and market insights.

[0033] 2. Deep intelligent data analysis and mining:

[0034] The use of innovative fusion analysis algorithms and advanced models can deeply explore the potential value and association patterns in the data. For example, by building a data relationship network based on the fusion of knowledge graphs and deep learning, it is possible to accurately predict customer behavior and needs and provide a scientific basis for precision marketing. At the same time, the analysis strategy is adjusted in real time according to data changes to ensure the timeliness and accuracy of the analysis results.

[0035] 3. Personalized marketing planning and execution:

[0036] Based on the results of data analysis, personalized marketing plans can be tailored for different customer groups. For example, the creative generation model based on the generative adversarial network can be used to automatically generate attractive advertising content, formulate optimized promotional activity strategies for different customer groups, improve marketing effectiveness and customer satisfaction, reduce resource waste, and achieve precise delivery of marketing resources.

[0037] 4. Efficient customer relationship management:

[0038] Through the construction of detailed customer information files and intelligent communication assistance tools, efficient management and maintenance of customer relationships can be achieved. For example, intelligent customer service robots and human customer service work together to handle customer feedback in a timely manner, provide personalized communication services based on customer history and preferences, and improve customer loyalty and corporate reputation.

[0039] 5. Real-time dynamic effect evaluation and optimization:

[0040] With rich key performance indicators and dynamic evaluation models, the effectiveness of marketing activities can be monitored and evaluated in real time. For example, marketing strategies can be adjusted in real time based on indicators such as click-through rate, conversion rate, and customer retention growth rate. By automatically learning and optimizing evaluation strategies and indicator weights, the scientificity and effectiveness of marketing decisions can be ensured, and the return on investment of marketing activities can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a schematic block diagram of a digital marketing system based on big data proposed by the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0044] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, and it can be the internal connection of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0045] Reference Figure 1 : A digital marketing system based on big data, including a data collection module, a data analysis module, a marketing planning module, a customer relationship management module and an effect evaluation module;

[0046] The data collection module is used to collect a wide range of marketing-related data, including social media interaction data, such as the sentiment tendency value S of comments, obtained through the natural language processing sentiment analysis algorithm, S∈[-1, 1], -1 represents extremely negative sentiment, 1 represents extremely positive sentiment, website user behavior trajectory data, transaction flow data, etc. It uses advanced distributed crawler technology and intelligent data interfaces to ensure comprehensive, accurate and efficient data collection. Its data collection efficiency evaluation indicators Where E collect represents the data collection efficiency, D i It represents the amount of data of different types collected per unit time, n represents the total number of data types, T collect Indicates the collection time.

[0047] The data analysis module deeply explores the value of data. It uses innovative fusion analysis algorithms and combines deep learning neural networks, such as the improved Transformer architecture network for processing text data feature extraction and reinforcement learning decision models, to dynamically adjust analysis strategies and conduct in-depth analysis of data. It divides customers into multiple precise customer groups based on their behavior patterns and preferences. j , where j represents the customer group number, and builds a customer behavior prediction model P(x), where x represents the customer behavior feature vector, for example, by analyzing historical data to predict the probability of a customer's future purchase intention. Its data analysis accuracy evaluation index is Among them A analysis Indicates the accuracy of data analysis, N data_correctly_analyzed Indicates the amount of data analyzed correctly, N data_k represents the size of the kth data sample, and m represents the total number of data samples.

[0048] The marketing planning module customizes marketing strategies based on the analysis results. This includes generating personalized advertising ideas, automatically generating advertising copy and creative materials based on customer group characteristics and prediction models, planning precise promotional activities, and formulating preferential strategies that are suitable for different customer groups, such as optimizing discounts for different levels. For example, launching exclusive customized products and services for high-value customer groups.

[0049] The customer relationship management module efficiently manages customer relationships. It includes the construction of detailed customer information files, recording basic customer information, historical interaction records, and preference tag set L = {l1, l2, ..., l n}, etc. Intelligent customer communication assistance automatically provides communication suggestions and speech templates based on the customer's historical communication style and current problems, intelligently processes customer feedback, and uses intelligent customer service robots and manual customer service to work together. The robots handle common problems and transfer complex problems to manual customer service.

[0050] The effect evaluation module monitors the marketing effect in real time. By setting key performance indicators (KPIs) such as click-through rate (C click , conversion rate C convert , Customer retention growth rate R retain Etc. to conduct quantitative evaluation. Construct a dynamic effect evaluation model M(E, t), where E represents the evaluation indicator set and t represents time. The evaluation strategy and indicator weights are adjusted in real time according to data changes. Its effect evaluation timeliness indicator Where T evaluation Indicates the timeliness of effect evaluation, T total Represents the total time of marketing activities, T delay T represents the time interval from the start of the marketing campaign to the acquisition of initial evaluation data. adapt Indicates the time required for the system to adjust the evaluation strategy based on data changes.

[0051] In the present invention, the distributed crawler technology in the data acquisition module supports multi-threaded concurrent acquisition and dynamic data source adaptation, and can automatically adjust the acquisition strategy according to changes in the data source structure.

[0052] In the present invention, the fusion analysis algorithm in the data analysis module innovatively adopts advanced technology based on the fusion of knowledge graph and deep learning to achieve in-depth analysis and value mining of data.

[0053] In terms of knowledge graph construction, the first step is to comprehensively sort out and classify all types of collected data. For structured data, such as product information, customer information, and sales time in transaction records, they will be sorted according to preset rules and relationships to build a preliminary entity and relationship framework. For unstructured data, such as text comments on social media, pictures and videos posted by users, etc., natural language processing technology and image recognition technology will be used for analysis. For example, natural language processing technology is used to extract key entities and semantic relationships in the text, and information such as product names and brand reviews mentioned by users are converted into nodes and edges in the knowledge graph. For image and video data, image recognition technology is used to identify key elements such as objects and scenes, and associate them with related products or user behaviors.

[0054] In the application of deep learning technology, a variety of deep learning models are used to extract features and recognize patterns from data. For example, using convolutional neural networks to extract features from image and video data can automatically learn features such as texture and shape in images, as well as dynamic patterns and scene changes in videos. For text data, using recurrent neural networks or language models with Transformer architectures can deeply understand the semantics and contextual relationships of text. At the same time, through unsupervised learning, deep learning models can automatically discover potential patterns and structures in data.

[0055] When integrating knowledge graphs and deep learning, the prior knowledge in the knowledge graph is used to guide the training of the deep learning model. For example, when analyzing customer purchasing behavior, the relationship network between customers and products that has been constructed in the knowledge graph can be used as the initial input of the deep learning model to help the model understand the associations in the data more quickly. At the same time, the new associations and patterns discovered by the deep learning model during the training process can also be fed back to the knowledge graph to continuously enrich and improve the content of the knowledge graph.

[0056] The data relationship network constructed by this fusion technology can cross different types of data and fields to discover deep connections and potential patterns. For example, when analyzing customer behavior, not only can the direct connection of customers purchasing a certain product be discovered, but also the indirect connection between customers' social media behavior, browsing history, etc. and purchasing behavior can be discovered through the fusion analysis of knowledge graph and deep learning. For example, if a customer frequently follows topics in a specific field on social media and also frequently browses related product information on the website, through the analysis of the data relationship network, it can be inferred that this customer has a high potential purchase intention for products in this field.

[0057] This fusion analysis algorithm can also achieve dynamic update and real-time analysis of data. With the continuous influx of new data, the knowledge graph can update entities and relationships in real time, and the deep learning model can also continuously adjust parameters and optimize the model structure to adapt to data changes. For example, when a new product is launched on the market, the system can quickly incorporate it into the knowledge graph and analyze its potential association with existing products and customers through the deep learning model, providing timely support for marketing decisions.

[0058] In the present invention, the advertising creative generation in the marketing planning module adopts a creative generation model based on a generative adversarial network (GAN) to automatically generate attractive advertising content.

[0059] In the present invention, the intelligent communication assistance in the customer relationship management module makes full use of the intelligent assistant based on sentiment analysis and dialogue management technology to provide customers with highly personalized communication services.

[0060] In terms of sentiment analysis, the smart assistant will first conduct a comprehensive text analysis of the customer's historical communication records. It uses the sentiment analysis algorithm in natural language processing technology to make detailed emotional judgments on the sentences expressed by customers in previous communications with the company. For example, when customers used positive words and tones in past consultations, such as "very satisfied" and "I like this product very much", the smart assistant will mark this information as positive emotions and record it. On the contrary, if the customer uses words of complaint or dissatisfaction, such as "the quality is too poor" and "the service is terrible", the smart assistant will identify it as negative emotions.

[0061] By analyzing the sentiment of a large number of historical communication records, the intelligent assistant is able to build an emotional profile of each customer. In subsequent communications, when the customer contacts the company again, the intelligent assistant can quickly understand the customer's emotional state based on the customer's emotional profile. For example, if a customer has had many positive communication records before, but showed some negative emotions in the most recent communication, the intelligent assistant will immediately detect this change and use it as a signal to focus on.

[0062] Based on the results of sentiment analysis, the intelligent assistant will adopt corresponding strategies in dialogue management. At the beginning of the conversation, for customers with positive emotions, the intelligent assistant will use warm and friendly opening remarks, such as "It's nice to communicate with you again. How did you use the product you mentioned last time?" to further enhance customer favorability and loyalty. For customers with negative emotions, the intelligent assistant will communicate with a caring and understanding tone, such as "I'm very sorry to hear that you are dissatisfied. We will definitely solve your problem as soon as possible."

[0063] In the application of dialogue management technology, the intelligent assistant has a strong ability to control the dialogue process. It will intelligently guide the direction of the conversation according to the type of customer's question and emotional state. For example, when a customer asks about how to use a product, the intelligent assistant will gradually guide the customer to understand and master the correct usage steps based on the preset knowledge base and process. If the customer's question is more complicated, the intelligent assistant will automatically determine whether it needs to be transferred to a human customer service, and provide detailed customer information and problem background during the transfer process so that the human customer service can better understand the customer's needs.

[0064] At the same time, the intelligent assistant can also provide personalized responses based on the customer's conversation habits and preferences. For example, for customers who prefer concise and clear answers, the intelligent assistant will provide short and accurate answers; for customers who prefer detailed explanations, the intelligent assistant will provide more abundant information and cases. In addition, the intelligent assistant can also adjust its expression according to the customer's language style, making its communication with the customer more natural and smooth.

[0065] Throughout the communication process, the intelligent assistant will continue to learn and optimize. It will adjust its sentiment analysis model and dialogue management strategy based on the results of each conversation and customer feedback. For example, if a customer is dissatisfied with an answer, the intelligent assistant will record it and improve the answer the next time it encounters a similar question. Through this continuous learning and optimization, the intelligent assistant can continuously improve its communication ability and provide customers with better personalized communication services.

[0066] In the present invention, the effect evaluation dynamic model of the effect evaluation module demonstrates extremely excellent automatic learning and optimization capabilities, plays a key role in the effectiveness evaluation of digital marketing, and can continuously improve evaluation strategies and accurately adjust indicator weights based on rich historical marketing data and real-time feedback.

[0067] In the implementation path of automatic learning, the model first established a huge data storage and analysis system. It will comprehensively collect detailed data of various marketing activities in the past. These data cover the effect data generated by different advertising channels, such as the number of ad impressions, clicks, user stay time and final conversion data on various social media platforms; it also includes sales data under different promotional activities, such as product sales changes, sales growth rate and customer participation corresponding to promotional activities with different discount strengths; it also includes data on customer participation over time, such as the frequency of customer response to various marketing activities in different time periods, and the trend of interactive behavior.

[0068] Specifically, the model uses advanced machine learning algorithms to conduct in-depth mining and analysis of these historical data. It will classify and organize the data and divide it according to different marketing scenarios, target customer groups, and time dimensions. By building complex mathematical models and algorithms, the model can automatically discover hidden patterns and laws in the data. For example, it may find that during a specific holiday period, when a certain style of advertising creativity and distribution channel combination is used for a customer group of a certain age group with specific consumption preferences, a higher conversion rate can be achieved. Based on these findings, the model will continuously update and improve its own knowledge system so that it can better understand and predict the effects of different marketing initiatives in future evaluations.

[0069] When it comes to optimizing based on real-time feedback, the dynamic model of effect evaluation is closely connected and interactive with all aspects of the entire marketing system. Through the real-time monitoring system, it can quickly receive feedback information from various channels. For example, during the advertising process, the model will use advanced monitoring technology to track users' various interactive behaviors on the advertisement in real time. This includes not only common indicators such as clicks and browsing time, but also in-depth analysis of a series of subsequent behaviors of users after watching the advertisement, such as whether to search for related products, whether to discuss on social media, etc.

[0070] If the click-through rate of an ad suddenly drops, the model will immediately start the intelligent analysis mechanism. It will comprehensively consider the current market environment changes, such as whether competitors in the same industry have launched similar ads or new market trends have caused user interest to shift; it will also analyze competitors' dynamics, including their advertising strategy adjustments, promotion activities, etc.; at the same time, it will also pay close attention to the real-time trends of user behavior, such as changes in user search keywords and changes in browsing habits.

[0071] Based on these comprehensive analysis results, the model will automatically adjust the evaluation strategy. For example, if it is found that users are tired of the visual style of the advertisement, the model will suggest adjusting the creative design of the advertisement, adding novel elements or changing the presentation method; if the time of advertisement delivery does not match the user's active time, the model will automatically adjust the delivery schedule. In addition, for some advertisements with more negative user feedback, the model will increase the monitoring frequency so as to detect and solve problems in a timely manner.

[0072] In terms of adjusting the weights of evaluation indicators, the model will use intelligent algorithms to dynamically optimize based on the comprehensive situation of historical data and real-time feedback. Specifically, at a certain marketing stage, if data analysis finds that customer retention rate plays a vital role in the long-term development of the company, and historical data also shows that there is a certain correlation between customer retention rate and other indicators, then the model will automatically use the weight adjustment algorithm to increase the weight of customer retention rate in the evaluation system.

[0073] For example, if it is found that the improvement of customer retention rate can significantly drive subsequent repeat purchases and word-of-mouth communication, thus having a significant impact on the company's long-term profit growth, the model may increase the weight of the customer retention rate indicator accordingly, making it occupy a more important position in the evaluation. At the same time, for some relatively unimportant indicators or those that are relatively stable at the current stage, the model will reduce their weight.

[0074] For example, during a promotion, real-time feedback shows that customers are highly sensitive to price discounts, and the model will respond quickly. It will increase the weight of price-related indicators (such as the impact of discount intensity on sales) accordingly through a comprehensive analysis of sales data, customer purchasing behavior, and market competition. In this way, when evaluating the effectiveness of promotions, the influence of price-related indicators will be enhanced, so as to more accurately measure the actual effectiveness of promotions.

[0075] Through the above series of advanced technical means and intelligent algorithms, the dynamic model of effect evaluation can ensure that the evaluation results can always accurately reflect the true performance of marketing activities and provide the most valuable reference for the company's marketing decisions.

[0076] In terms of automatic learning, the model will continuously conduct in-depth analysis and mining of historical marketing data. It will collect and organize detailed data on various marketing activities in the past, including performance data of different advertising channels, sales data under different promotional activities, data on customer engagement over time, and so on. Specifically, for advertising performance data, the model will record in detail the time, location, audience characteristics, and corresponding key indicators such as click-through rate and conversion rate of each ad display. By learning from these large amounts of historical data, the model can automatically discover patterns and rules. For example, it may find that ads targeting a specific age group in a specific time period have a relatively high conversion rate on a specific platform, or that a certain form of promotional activity is more effective in a specific season. Based on these findings, the model can gradually establish a cognitive system for the effects of different marketing scenarios and strategies, providing knowledge reserves for subsequent evaluation and optimization.

[0077] When it comes to optimizing based on real-time feedback, the dynamic model for effect evaluation is closely integrated with all aspects of the marketing system, receiving feedback information from various channels in real time. Specifically, during the advertising process, the model will monitor users' interactive behaviors on ads in real time, such as clicks, browsing time, sharing, etc. If it is found that the click-through rate of an ad suddenly drops, the model will immediately start the analysis mechanism. It will comprehensively consider factors such as changes in the current market environment, the dynamics of competitors, and real-time trends in user behavior. For example, it may be because competitors have launched similar ads to attract user attention, or new hot topics have appeared in the market, causing user interest to shift. Based on these analysis results, the model will automatically adjust the evaluation strategy, such as increasing the monitoring frequency of ads with negative user feedback, or changing the focus of key indicators for evaluating the effectiveness of the ad.

[0078] In terms of adjusting the weights of evaluation indicators, the model will make dynamic adjustments based on the comprehensive situation of historical data and real-time feedback. Specifically, at a certain stage, if it is found that customer retention rate is crucial to the long-term development of the company, and historical data also shows that there is a certain correlation between customer retention rate and other indicators, then the model will automatically increase the weight of customer retention rate in the evaluation system. At the same time, reduce the weight of some relatively unimportant indicators or indicators that are relatively stable at the current stage. For example, during a promotion, real-time feedback shows that customers are more sensitive to price discounts, then the model will correspondingly increase the weight of price-related indicators (such as indicators of the impact of discount intensity on sales) in order to more accurately evaluate the effectiveness of the promotion. In this way, it is ensured that the evaluation results can always accurately reflect the true performance of marketing activities and provide the most valuable reference for decision-making.

[0079] A digital marketing method based on big data, comprising the following steps:

[0080] S1. Data collection: Use the data collection module to collect marketing-related data from multiple channels, and use data preprocessing techniques (such as data cleaning, denoising, format conversion, etc.) to preliminarily organize the data.

[0081] S2. Data analysis: Transfer the preprocessed data to the data analysis module, use innovative analysis algorithms and models to conduct in-depth analysis, and explore customer needs and market trends.

[0082] S3. Marketing planning: Based on the results of data analysis, use the marketing planning module to develop personalized marketing plans, including advertising strategies, promotional activities plans, etc.

[0083] S4. Marketing execution: Use automated marketing tools (such as marketing automation software, social media promotion tools, etc.) to put marketing plans into practice.

[0084] S5. Effect evaluation: Use the effect evaluation module to monitor and evaluate the effects of marketing activities in real time, and adjust marketing strategies and plans in a timely manner based on the evaluation results.

[0085] In the data collection step, blockchain technology is innovatively introduced to ensure the secure transmission and storage of data in all aspects.

[0086] When data starts to be collected from various channels, the distributed ledger feature of blockchain immediately comes into play. Each data collection node is equivalent to a node in the blockchain network, and the transmission of data between these nodes adopts encrypted point-to-point communication. For example, when collecting user comment data from social media platforms, the data is divided into multiple small blocks and encrypted using advanced encryption algorithms. These encrypted data blocks are transmitted through multiple nodes in the blockchain network, and each node verifies and records the data blocks to ensure that the data is not tampered with or stolen during transmission. Even if a malicious attacker tries to intercept the data, it is difficult for the attacker to crack and tamper with the data due to the complexity of the encryption algorithm and the distributed nature of the blockchain network.

[0087] In terms of data storage, the tamper-proof nature of blockchain is fully reflected. Each piece of collected data will be packaged into a data block and linked to the previous data block through a cryptographic hash function to form a chain structure. When a new data block is added to the blockchain, it needs to be verified by consensus of multiple nodes in the network. Only when most nodes recognize the authenticity and integrity of the new data can the data block be successfully added to the blockchain. In this way, once the data is stored on the blockchain, it is almost impossible to tamper with it. For example, if someone tries to tamper with a transaction data, the hash value of the data block where the data is located and the previous and next data blocks associated with it will change, which will be immediately detected by other nodes in the blockchain network, thereby ensuring the originality and authenticity of the data.

[0088] At the same time, the traceability of blockchain provides a clear record of the source and flow path of data. Every data collection, transmission and storage operation will be recorded on the blockchain, forming an indelible transaction record. Through the traceability function of blockchain, you can clearly understand the source channel, collection time and every operation node in the entire data flow process of each piece of data. For example, when you need to query the source of a certain user behavior data, you can quickly trace it back to the website or application that originally collected the data, as well as every node passed through in the subsequent transmission and storage process through the transaction record on the blockchain. This not only helps to quickly locate and solve data problems when they arise, but also provides strong support for data compliance and auditing.

[0089] In the present invention, in the data analysis step, virtual reality (VR) and augmented reality (AR) technologies are used to visualize data and provide an immersive data analysis experience.

[0090] In the marketing execution step, artificial intelligence algorithms are fully utilized to achieve comprehensive optimization of the advertising and promotion execution process, thereby significantly improving marketing effectiveness and efficiency.

[0091] For advertising, artificial intelligence algorithms will first conduct in-depth analysis of the massive amount of data collected to accurately identify the characteristics and behavior patterns of the target customer group. For example, by mining multi-dimensional data such as users' browsing history on various websites, interactive behavior on social media, and consumption records, a detailed user portrait is constructed. Based on these portraits, the algorithm can accurately determine the best channels and timing for advertising. For example, for users who are often active on sports websites and have a record of purchasing sports equipment, the system will choose to place sports brand advertisements on the sports websites they visit most frequently and during the live broadcasts of related sports events.

[0092] At the same time, artificial intelligence algorithms can also monitor the effectiveness of advertising in real time. Through real-time tracking and analysis of key indicators such as click-through rate, browsing time, conversion rate, etc., advertising strategies can be adjusted in a timely manner. For example, if it is found that the click-through rate of an advertisement on a specific platform is low, the algorithm will automatically analyze the reasons, which may be poor ad placement, insufficient ad creativity, or inaccurate target audience positioning. Then, based on the analysis results, the ad placement, creative content, or target audience group can be automatically adjusted. For example, the ad position can be adjusted to a more prominent position, or the ad creativity can be redesigned according to the preferences of the target audience, such as using more attractive pictures, more vivid videos, or more targeted copy.

[0093] Artificial intelligence algorithms also play a key role in the execution of promotional activities. It can tailor personalized promotional plans for each customer based on information such as the customer's purchase history, browsing behavior, and preferences. For example, for those customers who frequently purchase electronic products, when new electronic products are launched or promotional activities are carried out, the system will automatically push relevant preferential information and personalized discount plans to them. At the same time, the algorithm can also analyze and learn from the data of historical promotional activities to predict the effects of different promotional strategies. For example, by analyzing the number of participants in similar promotional activities in the past, sales growth, and customer feedback, the best discount strength, promotion time, and promotion method for this promotion can be determined.

[0094] In addition, AI algorithms can also achieve real-time dynamic adjustments to promotional activities. During the promotional activities, the promotional strategy can be adjusted in a timely manner based on real-time sales data, customer engagement, market feedback and other information. For example, if it is found that the sales of a certain promotional product are not as good as expected, the algorithm will automatically analyze the reasons, which may be too high a price, insufficient publicity, or competitors launching more attractive products. Then, according to the analysis results, the promotional price can be adjusted in a timely manner, publicity channels can be increased, or the promotional product portfolio can be optimized. For example, the price of the product can be lowered, publicity on social media can be increased, or the product can be combined with other popular products for promotion to improve the effectiveness and efficiency of the promotional activities.

[0095] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A digital marketing system based on big data, characterized in that: Including data collection module, data analysis module, marketing planning module, customer relationship management module and effect evaluation module; Data collection module: used to collect marketing data, social media interaction data, website user behavior trajectory data, and transaction flow data. The social media interaction data includes the comment sentiment tendency value S, which is obtained through the natural language processing sentiment analysis algorithm, and S∈[-1, 1], -1 represents extremely negative sentiment, and 1 represents extremely positive sentiment. Distributed crawler technology and intelligent data interface are used, and the data collection efficiency evaluation index formula is: Where E collect represents the data collection efficiency, D i It represents the amount of data of different types collected per unit time, n represents the total number of data types, T collect Indicates the collection time; The data analysis module uses a fusion analysis algorithm, combined with a deep learning Transformer architecture neural network, to process text data feature extraction and reinforcement learning decision models, to dynamically adjust analysis strategies, and to segment customers into precise customer groups based on their behavior patterns and preferences. j , where j represents the customer group number, and builds a customer behavior prediction model P(x), where x represents the customer behavior feature vector. By analyzing historical data, the probability of a customer's future purchase intention is predicted. The data analysis accuracy evaluation index formula is: Among them A analysis Indicates the accuracy of data analysis, N data_correctly_analyzed Indicates the amount of data analyzed correctly, N data_k represents the kth data sample size, and m represents the total number of data samples; Marketing planning module: customizes marketing strategies based on analysis results, including personalized advertising creative generation, automatically generates advertising copy and creative materials based on customer group characteristics and prediction models, accurately plans promotional activities, and formulates preferential strategies suitable for different customer groups; Customer relationship management module: manage customer relationships, including building customer information archives, recording basic customer information, historical interaction records, and preference tag set L = {l1, l2, ..., l n Intelligent customer communication assistance automatically provides communication suggestions and speech templates based on the customer's historical communication style and current problems, intelligently processes customer feedback, and uses intelligent customer service robots and manual customer service to work together. The robot handles common problems and transfers complex problems to manual customer service; The effect evaluation module monitors the marketing effect in real time and sets the click rate C click , conversion rate C convert , Customer retention growth rate R retain Performance indicators are quantitatively evaluated and a dynamic effect evaluation model M(E, t) is constructed. t represents the evaluation indicator set and t represents time. The evaluation strategy and indicator weights are adjusted in real time according to data changes. The timeliness indicator formula for effect evaluation is: Where T evaluation Indicates the timeliness of effect evaluation, T total Represents the total time of marketing activities, T delay T represents the time interval from the start of the marketing campaign to the acquisition of initial evaluation data. adapt Indicates the time required for the system to adjust the evaluation strategy based on data changes.

2. The digital marketing system based on big data according to claim 1, characterized in that: The distributed crawler technology in the data collection module supports multi-threaded concurrent collection and dynamic data source adaptation, and can automatically adjust the collection strategy according to changes in the data source structure.

3. The digital marketing system based on big data according to claim 1, characterized in that: The fusion analysis algorithm in the data analysis module adopts the technology based on the fusion of knowledge graph and deep learning to build a data relationship network and mine associations and potential patterns.

4. The digital marketing system based on big data according to claim 1, characterized in that: The advertising creative generation in the marketing planning module adopts a creative generation model based on the generative adversarial network (GAN) to automatically generate attractive advertising content.

5. The digital marketing system based on big data according to claim 1, characterized in that: The intelligent communication assistance in the customer relationship management module adopts an intelligent assistant based on sentiment analysis and dialogue management technology.

6. The digital marketing system based on big data according to claim 1, characterized in that: The effect evaluation dynamic model in the effect evaluation module supports automatic learning and optimization, and adjusts the evaluation strategy and indicator weights according to historical marketing data and real-time feedback.

7. A digital marketing method based on big data, used to implement the digital marketing system based on big data as described in any one of claims 1 to 6, characterized in that: The following steps are involved: S1. Data collection: Use the data collection module to collect marketing-related data from channels, and use data preprocessing techniques such as data cleaning, denoising, and format conversion to preliminarily organize the data; S2. Data analysis: transfer the pre-processed data to the data analysis module, use innovative analysis algorithms and models to conduct in-depth analysis, and explore customer needs and market trends; S3. Marketing planning: Based on the data analysis results, use the marketing planning module to develop personalized marketing plans, including advertising strategies and promotional activities plans; S4. Marketing execution: Use automated marketing tools, including marketing automation software and social media promotion tools, to implement marketing plans; S5. Effect evaluation: Use the effect evaluation module to monitor and evaluate the effects of marketing activities in real time, and adjust marketing strategies and plans in a timely manner based on the evaluation results.

8. The digital marketing method based on big data according to claim 7, characterized in that: In the data collection step, blockchain technology is used to ensure the secure transmission and storage of data, and to achieve data traceability and non-tamperability.

9. The digital marketing method based on big data according to claim 7, characterized in that: In the data analysis step, virtual reality VR and augmented reality AR technologies are used to visualize data and provide an immersive data analysis experience.

10. The digital marketing method based on big data according to claim 7, characterized in that: In the marketing execution step, artificial intelligence algorithms are used to optimize the advertising and promotion execution process to improve marketing effectiveness and efficiency.

Images