Intelligent customer obtaining method and system based on multi-modal data

Abstract

The invention discloses an intelligent customer obtaining method and system based on multi-modal data, and relates to the technical field of multi-modal data. Corresponding customer demand information is determined according to detection of a demand space; according to the content of the customer demand information, the information weight and the multi-modal data, multiple pieces of target data in the multi-modal data are determined, and the final customer obtaining path is determined based on the multiple pieces of target data, the enterprise database and the customer demand information, so that the accuracy of the customer obtaining path is improved. Therefore, the corresponding customer information is determined according to the detection of the customer obtaining node of the customer obtaining path, and the corresponding customer obtaining mode is matched based on the customer information and the customer demand information; and determining a corresponding customer obtaining number according to the execution of the customer obtaining mode, and triggering autonomous optimization of the customer obtaining mode according to the customer obtaining number and the execution condition of the customer obtaining mode until the customer obtaining number is greater than a preset customer obtaining number threshold value, so as to improve the intelligent customer obtaining effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-modal data, and particularly relates to an intelligent customer acquisition method and system based on multi-modal data. BACKGROUND

[0002] With the development of science and technology, enterprise databases are gradually applied in various enterprises and store corresponding enterprise information, which covers internal information and external interaction information of the enterprise; in the prior art, customer demand information is collected, and corresponding matching content is determined according to matching of the customer demand information and the enterprise information, and a corresponding propagation mode is triggered along the matching content, however, the propagation mode does not consider the customer acquisition path, which affects the accuracy of the customer acquisition mode and cannot realize autonomous optimization of the customer acquisition mode. SUMMARY

[0003] The present application aims to overcome the shortcomings of the prior art, and provides an intelligent customer acquisition method and system based on multi-modal data.

[0004] The present application provides an intelligent customer acquisition method based on multi-modal data, which comprises the following steps: According to the traversal of the enterprise database, industry information, talent information and enterprise information are determined, and corresponding multi-modal data is constructed based on the industry information, talent information and enterprise information; According to the traversal of the enterprise database, a corresponding demand space is determined, according to the detection of the demand space, corresponding customer demand information is determined, and the information weight of the customer demand information is marked; According to the content, information weight and multi-modal data of the customer demand information, a plurality of target data in the multi-modal data is determined, and based on the plurality of target data, the enterprise database and the customer demand information, a final customer acquisition path is determined; According to the traversal of the customer acquisition path, a corresponding customer acquisition node is determined, according to the detection of the customer acquisition node, corresponding customer information is determined, and a corresponding customer acquisition mode is matched based on each customer information and the customer demand information; According to the execution of the customer acquisition mode, a corresponding customer acquisition quantity is determined, and according to the customer acquisition quantity and the execution of the customer acquisition mode, autonomous optimization of the customer acquisition mode is triggered until the customer acquisition quantity is greater than a preset customer acquisition quantity threshold.

[0005] The present application provides an intelligent customer acquisition system based on multi-modal data, which is applied to the intelligent customer acquisition method based on multi-modal data described above, and comprises: A multi-modal data module is configured to determine industry information, talent information and enterprise information according to the traversal of the enterprise database, and construct corresponding multi-modal data based on the industry information, talent information and enterprise information. a customer demand information module configured to determine a corresponding demand space based on traversal of the enterprise database, determine corresponding customer demand information based on detection of the demand space, and mark information weights of the customer demand information; a customer acquisition path module configured to determine a plurality of target data in the multi-modal data based on contents of the customer demand information, the information weights, and the multi-modal data, and determine a final customer acquisition path based on the plurality of target data, the enterprise database, and the customer demand information; a customer acquisition method module configured to determine a corresponding customer acquisition node based on traversal of the customer acquisition path, determine corresponding customer information based on detection of the customer acquisition node, and match a corresponding customer acquisition method based on each of the customer information and the customer demand information; an autonomous optimization module configured to determine a corresponding customer acquisition quantity based on execution of the customer acquisition method, trigger autonomous optimization of the customer acquisition method based on the customer acquisition quantity and execution of the customer acquisition method, until the customer acquisition quantity is greater than a preset customer acquisition quantity threshold.

[0006] Compared with the prior art, the present application has the following advantages: In the embodiment of the present application, the industry information, talent information, and enterprise information are determined based on traversal of the enterprise database, and the corresponding multi-modal data is constructed based on the industry information, talent information, and enterprise information; the corresponding demand space is determined based on traversal of the enterprise database, the corresponding customer demand information is determined based on detection of the demand space, and the information weights of the customer demand information are marked; the plurality of target data in the multi-modal data is determined based on contents of the customer demand information, the information weights, and the multi-modal data, and the final customer acquisition path is determined based on the plurality of target data, the enterprise database, and the customer demand information, which introduces the multi-modal data, considers the plurality of target data, the enterprise database, and the customer demand information as a whole, and improves the precision of the customer acquisition path.

[0007] Therefore, the corresponding customer acquisition node is determined based on traversal of the customer acquisition path, the corresponding customer information is determined based on detection of the customer acquisition node, and the corresponding customer acquisition method is matched based on each of the customer information and the customer demand information; the corresponding customer acquisition quantity is determined based on execution of the customer acquisition method, the autonomous optimization of the customer acquisition method is triggered based on the customer acquisition quantity and execution of the customer acquisition method, until the customer acquisition quantity is greater than the preset customer acquisition quantity threshold, which improves the precision of the customer acquisition method and the autonomous optimization of the customer acquisition method, thereby improving the effect of intelligent customer acquisition. BRIEF DESCRIPTION OF DRAWINGS

[0008] Figure 1 is a flowchart of the intelligent customer acquisition method based on multi-modal data in the embodiment of the present application; Figure 2It is a structural composition schematic diagram of an intelligent customer acquisition system based on multi-modal data in an embodiment of the present application. DETAILED DESCRIPTION

[0009] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0010] Please refer to Figure 1 and Figure 2 An intelligent customer acquisition method based on multi-modal data is applied to an intelligent customer acquisition scene based on multi-modal data. The intelligent customer acquisition method based on multi-modal data comprises the following steps. Step S11: determining industry information, talent information and enterprise information according to traversal of an enterprise database, and constructing corresponding multi-modal data based on the industry information, the talent information and the enterprise information; Step S12: determining a corresponding demand space based on traversal of the enterprise database, determining corresponding customer demand information according to detection of the demand space, and marking information weights of the customer demand information; Step S13: determining a plurality of target data in the multi-modal data according to contents of the customer demand information, the information weights and the multi-modal data, and determining a final customer acquisition path based on the plurality of target data, the enterprise database and the customer demand information; Step S14: determining corresponding customer acquisition nodes based on traversal of the customer acquisition path, determining corresponding customer information according to detection of the customer acquisition nodes, and matching corresponding customer acquisition modes based on each customer information and the customer demand information; Step S15: determining a corresponding customer acquisition quantity according to execution of the customer acquisition mode, triggering autonomous optimization of the customer acquisition mode according to the customer acquisition quantity and execution of the customer acquisition mode, until the customer acquisition quantity is greater than a preset customer acquisition quantity threshold; In step S11, the industry information, the talent information and the enterprise information are determined according to traversal of the enterprise database, and the corresponding multi-modal data is constructed based on the industry information, the talent information and the enterprise information; In the specific implementation process of the present application, the specific steps are as follows: S111: collecting an enterprise database, determining an information space based on detection of the enterprise database, and determining industry information, talent information and enterprise information according to detection of the information space; S112: determining a first information interaction combination based on the industry information and the industry information, and determining a second information interaction combination based on the industry information and the enterprise information; S113: constructing corresponding multi-modal data in the enterprise database according to the first information interaction combination, the second information interaction combination and a multi-modal data mapping relationship.

[0011] In the embodiments of the present application, the type of enterprise database that needs to be collected is determined, which includes internal databases (such as ERP systems, CRM systems, human resource systems, etc.) and external databases (such as industry reports, market research data, public data sources, etc.); according to the type of data source, select the appropriate data collection method; for example, for internal databases, data needs to be obtained through API interface, database query or data export tool; for external databases, data needs to be obtained through subscription service, web crawler or third-party data provider; during the process of collecting data, preliminary preprocessing of data is performed, such as de-duplication, format conversion, data cleaning, etc., to ensure the accuracy and consistency of the data.

[0012] The collected data is analyzed using data visualization tools (such as Tableau, Power BI) and statistical analysis software (such as SPSS, R language) to identify key information areas in the data, i.e. information space; according to the results of exploratory analysis, define the information space, including the main types of data, structural characteristics, potential value points, etc.; at the same time, through exploratory analysis, it is found that the employee archives contain information about the skills, experience, educational background of employees; project records contain information about the type, size, progress of projects; sales data contains information about customer purchase history, preferences, feedback, etc., which constitutes the information space and provides a basis for subsequent analysis.

[0013] According to the detection of the information space, industry information, talent information and enterprise information are determined; at this time, in the information space, information related to industry trends, market dynamics, policies and regulations is identified, and industry information is extracted; in the information space, information related to employee skills, experience, educational background, career development is identified, and talent information is extracted; in the information space, information related to enterprise size, business status, competitive advantage, market position is identified, and enterprise information is extracted.

[0014] Further, based on the industry information and the industry information, a first information interaction combination is determined, and based on the industry information and the enterprise information, a second information interaction combination is determined, which is compatible with the overall consideration of the industry information and the industry information, and ensures the accuracy of the first information interaction combination.

[0015] At this time, in-depth analysis of industry information is conducted to identify the relevance and mutual influence between different enterprises, products, services, technologies or market trends within the industry, including competition, cooperation, complementarity, substitution, etc.; based on the identified internal relationships of the industry, an information interaction network is constructed, which is a directed graph or an undirected graph, and the nodes represent different elements within the industry (such as enterprises, products, etc.), and the edges represent the interaction between these elements; in the information interaction network, identify key information interaction nodes and paths, which represent the most important information interaction combinations within the industry, which involve cooperation between multiple enterprises, changes in competitive situation, introduction of new technologies, etc.

[0016] At the same time, combined with industry information and enterprise information, the status of a specific enterprise in the industry, competitive advantage, challenges and potential growth opportunities are analyzed, including market share, technical strength, brand influence, user feedback, etc. of the enterprise; based on the relationship analysis between the enterprise and the industry, identify the information interaction mode between the enterprise and other elements within the industry, including the interaction between the enterprise and suppliers, customers, competitors, partners; in the information interaction mode, identify the most important information interaction combinations for a specific enterprise, which involve cooperation between the enterprise and key suppliers, communication with important customers, countermeasures against competitors, etc.

[0017] Therefore, in the enterprise database, the corresponding multi-modal data is constructed according to the first information interaction combination, the second information interaction combination and the multi-modal data mapping relationship, which is compatible with the overall consideration of the first information interaction combination, the second information interaction combination and the multi-modal data mapping relationship, ensuring the accuracy of the corresponding multi-modal data.

[0018] At this time, ensure that the data in the enterprise database is up-to-date and has been cleaned and pre-processed for subsequent analysis, including structured data (such as financial statements, sales records) and unstructured data (such as employee reports, market research documents); identify and collect multi-modal data related to the first information interaction combination and the second information interaction combination, including text data (such as news reports, social media posts), image data (such as product pictures, advertising posters), audio data (such as customer feedback recordings, product demonstration videos), etc.

[0019] Specifically, assume that we are building multi-modal data for a factory intermediary company. The relevant information of the factory sale projects involved by the company has been sorted out, including the enterprise database of the factory sale records, financial statements, market research reports, etc. In addition, news reports, social media posts related to factory sales are collected as text data, pictures of the appearance of the factory, interior facilities, and advertising posters are collected as image data, and recordings of customers consulting factory sale matters, factory introduction videos are collected as audio data.

[0020] The elements in the first information interaction combination and the second information interaction combination are analyzed to determine their association with the multi-modal data sources, which involves associating elements such as companies, products, and services within an industry with specific text, image, or audio data; based on the association analysis, multi-modal data mapping rules are developed, which define how to correlate data from different modalities to enable comprehensive utilization of this information in subsequent analysis; at the same time, the first information interaction combination (such as the competitive relationship between different factory intermediary companies) and the second information interaction combination (such as the cooperation between the factory intermediary company and key suppliers) are analyzed; the elements in these combinations (such as the names of factory intermediary companies and supplier names) are associated with text data (such as the dynamics of these companies mentioned in news reports), image data (such as advertising posters of these companies), and audio data (such as customer feedback recordings for these companies); then, mapping rules are developed, such as "if the text data contains the name of a factory intermediary company, then associate it with the corresponding sales record".

[0021] According to the mapping rules, data from different modalities is integrated into a unified dataset, which involves data format conversion, data synchronization, and data fusion operations; the integrated multi-modal dataset is verified to ensure data accuracy and consistency, which includes data integrity checks, data consistency verification, and data quality assessment; at the same time, according to the mapping rules, text data, image data, and audio data are integrated with sales records, financial statements, and other data in the enterprise database into a multi-modal dataset; data integrity checks are performed to ensure that each piece of data contains necessary elements such as factory intermediary company names, sales data, and news report content; data consistency verification is also performed to ensure that data from different modalities is consistent when describing the same element; finally, data quality is assessed to ensure data accuracy and reliability.

[0022] In some embodiments of the present application, an information interaction element matching table is collected, which is shown in Table 1: Match ID First information interaction element Second information interaction element Text data ID Image data ID Audio data ID 1 Factory intermediary company A Supplier B T001 I001 A001 2 Factory intermediary company A Marketing campaign C T002 I002 - 3 Factory intermediary company B Supplier B T003 - A002 ... ... ... ... ... ... In this example, the information interaction element matching table records the association between factory intermediary companies (first information interaction elements), suppliers or marketing activities (second information interaction elements), and text data, image data, and audio data; for example, the interaction combination of factory intermediary company A and supplier B is associated with text data T001, image data I001, and audio data A001.

[0023] Assuming the following weight distribution: text data weight: 0.5; image data weight: 0.3; audio data weight: 0.2; for the combination matching ID 1 (interaction between factory intermediary company A and supplier B), the score is calculated as follows: text data score: score of T001 (assuming 80 points) * 0.5 = 40 points; image data score: score of I001 (assuming 90 points) * 0.3 = 27 points; audio data score: score of A001 (assuming 70 points) * 0.2 = 14 points; total score = 40 points + 27 points + 14 points = 81 points; therefore, the combination matching ID 1 scores 81 points, reflecting the importance of this combination in the overall information space.

[0024] Based on the above matching table and score calculation, a multi-modal data set is constructed, which contains the unique identification of each matching combination, the first information interaction element, the second information interaction element, and the score of different modal data; the score matching table is shown in Table 2: Table 2 Score matching table In this example, the multi-modal data set records the unique identification of each matching combination, the interaction element and the score of different modal data, which are used for subsequent intelligent lead analysis to help better understand the complex relationships within the industry and the position and opportunities of specific enterprises in the industry.

[0025] In step S12, the corresponding demand space is determined based on the traversal of the enterprise database, the corresponding customer demand information is determined according to the detection of the demand space, and the information weight of the customer demand information is marked; In the specific implementation process of the present application, the specific steps are as follows: S121: Collecting enterprise database, determining corresponding demand space according to traversal of enterprise database, and triggering real-time detection of demand space; S122: Determining demand information sequence list according to real-time detection of demand space, and determining corresponding customer demand information according to detection of demand information sequence list; S123: Determining a plurality of demand characteristics based on analysis of customer demand information, and determining information weight of customer demand information according to a plurality of demand characteristics and corresponding weight mapping relationship.

[0026] In the embodiments of the present application, the enterprise database is collected, the corresponding demand space is determined according to the traversal of the enterprise database, and the real-time detection of the demand space is triggered, which is compatible with the overall consideration of the traversal of the enterprise database, and ensures the accuracy of the corresponding demand space.

[0027] At this time, determine the type of enterprise database that needs to be collected, which includes customer relationship management systems (CRM), enterprise resource planning systems (ERP), product databases, sales records, market research reports, etc.; use appropriate technical means (such as API interface calls, database queries, data scraping, etc.) to extract data from the above data sources; ensure the completeness, accuracy and timeliness of the data; clean, convert and standardize the extracted data to eliminate redundancy, correct errors and unify the data format for subsequent analysis.

[0028] Use data analysis tools or algorithms (such as data mining, clustering analysis, association rule mining, etc.) to conduct in-depth analysis on the preprocessed data; based on the analysis results, identify the customer's demand set in a specific field or scenario, which involves product type, functional characteristics, price range, purchase frequency, etc. multiple dimensions; refine and define the identified demand to clarify the specific content and characteristics of each demand.

[0029] Establish a real-time monitoring mechanism for the identified demand space, which involves setting data update frequency, event trigger conditions (such as new product release, market trend change) or active data mining tasks; use real-time data analysis techniques (such as stream processing, online learning, etc.) to analyze the monitored data in real time to capture changes and trends in the demand space; when significant changes in the demand space are detected, the system automatically triggers an alarm and generates corresponding response strategies or suggestions.

[0030] Further, according to the real-time detection of the demand space, a demand information sequence list is determined, and according to the detection of the demand information sequence list, corresponding customer demand information is determined, which is compatible with the overall consideration of real-time detection of the demand space, ensuring the accuracy of the demand information sequence list.

[0031] At this time, in step S121, a real-time monitoring mechanism for the demand space has been established; in this step, in-depth analysis of these real-time monitored data is required; extract information points related to customer demand from real-time monitoring data, which involve product type, functional characteristics, price sensitivity, purchase intention, etc. multiple aspects; based on the extracted demand information, sort according to the importance, urgency or relevance of the information, generate a demand information sequence list; the basis for sorting includes the frequency of information appearing, the range of influence, the potential value, etc.

[0032] Each item in the sorted demand information sequence list is analyzed in detail to clarify the specific demand content, characteristics and requirements; map the refined demand information to specific customer groups to determine which demands are raised by which customers and how well these demands are met; integrate the mapped customer demand information to form a complete and systematic customer demand information report, providing a basis for subsequent strategy formulation and product development.

[0033] Specifically, assuming that an online factory sale platform of a factory intermediary company is monitoring the factory purchase demands of its users in real-time; by analyzing the user's browsing behavior data (such as factory type browsing volume, page dwell time, consultation interaction, etc.), the platform extracts the user's high demand information for specific factory types (such as single-story factory, multi-story factory), specific factory types (such as single-story factory, multi-story factory), and specific preferences for factory area, rent or sale price range, and leasing or purchasing services; then, the platform sorts the demand information according to the importance (such as longer page dwell time for factory type indicating higher interest) and urgency (such as popular factories for rent or sale having higher attention) of this information, generating a demand information sequence table.

[0034] The online factory sale platform of the factory intermediary company conducts a detailed analysis of each item in the demand information sequence table; for example, for the high demand information of single-story factory, the platform further analyzes the user's preferences for the structural characteristics of the factory (such as steel structure, concrete structure), the configuration requirements (such as whether to be equipped with large loading and unloading equipment, whether to have good ventilation and lighting conditions), and the expectations for leasing or purchasing services (such as rent payment method, purchase loan scheme, lease term flexibility), etc.; then, the platform maps these refined demand information with specific user groups to determine which user groups have higher interest and purchase or lease intentions for these single-story factories. For example, it is found that small manufacturing enterprises prefer single-story factories with moderate area and reasonable rent, and have higher requirements for the loading and unloading equipment configuration of the factory; finally, the platform integrates these user demand information to form a detailed customer demand information report. This report provides strong data support for subsequent factory resource allocation (such as adjusting the supply proportion of different types of factories according to demand), marketing strategy formulation (such as launching customized factory recommendation schemes for specific user groups), and financial service optimization (such as designing loan or rent payment schemes that better meet user needs).

[0035] Therefore, based on the analysis of customer demand information, multiple demand characteristics are determined, the information weight of customer demand information is determined according to the multiple demand characteristics and the corresponding weight mapping relationship, the overall consideration of multiple demand characteristics and the corresponding weight mapping relationship is compatible, and the accuracy of the information weight of customer demand information is guaranteed.

[0036] At this time, detailed customer demand information has been obtained in step S122; in this step, these information needs to be deeply analyzed to extract key demand features; feature recognition is the key step to determine which information points or attributes can represent customer demand; these features involve product function, performance, price, appearance, brand, service, etc.; refine the identified features, and clarify the specific meaning, measurement standard and value range of each feature, which helps subsequent quantification and weight allocation of features.

[0037] Weight mapping relationship refers to the importance or influence of each demand feature in customer demand, which is determined by expert scoring, market research, historical data analysis, etc.; based on the weight mapping relationship, a specific weight value is assigned to each demand feature; the size of the weight value reflects the importance of the feature in customer demand; according to the weight value of each demand feature and the specific demand information of the customer, the overall weight of the customer demand information is calculated, which helps enterprises understand which demand is the most important and urgent, and thus develop appropriate strategies to meet these demands.

[0038] Suppose a factory intermediary company is conducting demand analysis on its potential customers; through step S122, the factory intermediary company has obtained detailed demand information about the customer's demand for the factory, including the customer's preferences for factory type (such as single-story factory, multi-story factory), area, geographical location, price, facility configuration, rental or purchase method, etc.; in this step, the factory intermediary company has deeply analyzed these information and identified key demand features such as factory type, area demand, geographical location, price range, facility configuration, rental or purchase method, etc.; then, the factory intermediary company has refined these features and clarified the measurement standard and value range of each feature; Based on these weight values and the specific demand information of the customer (for example, a customer has very strict requirements for factory area, but is more flexible in price), the factory intermediary company calculates the overall weight of the customer demand information. This helps the factory intermediary company understand the customer's current most concerned demand points, and thus develop appropriate product development and marketing strategies to meet these demands; for example, if a customer's demand weight for factory area is 0.3, for geographical location is 0.25, for price is 0.2, and for facility configuration and rental method is 0.15 and 0.1 respectively, the factory intermediary company can prioritize recommending factories that meet the area and geographical location requirements, while providing some flexibility in price to meet the core needs of the customer; in this way, the factory intermediary company can more accurately meet the customer's demand for factory, improve customer satisfaction and transaction success rate.

[0039] Specifically, assume that a user has explicitly expressed the following preferences: commodity type: factory building (score 1, as it meets the user's preferences); price range: low price (score 0.8, as the user prefers low prices but accepts a certain range of price fluctuations); brand preference: no preference for developers (score 0.5, as the user has no special preference); commodity evaluation: factory facility integrity rate higher than 90% (score 1, as it fully meets the user's requirements); logistics speed: expect fast house viewing (score 1, as it fully meets the user's requirements); based on the above weights and scores, the overall weight of the user's demand information is calculated: overall weight = commodity type score x commodity type weight + price range score x price range weight + brand preference score x brand preference weight + commodity evaluation score x commodity evaluation weight + logistics speed score x logistics speed weight; At this time, the commodity type score is 1 (factory building); the price range score is 0.8 (low price); the brand preference score is 0.5 (no preference for developers); the commodity evaluation score is 1 (factory facility integrity rate higher than 90%); the logistics speed score is 1 (expect fast house viewing); assume the weight distribution is as follows: commodity type weight: 0.2; price range weight: 0.3; brand preference weight: 0.1; commodity evaluation weight: 0.25; logistics speed weight: 0.15; the overall weight is calculated as follows: Overall weight = 1 x 0.2 + 0.8 x 0.3 + 0.5 x 0.1 + 1 x 0.25 + 1 x 0.15 = 0.2 + 0.24 + 0.05 + 0.25 + 0.15 = 0.89; this overall weight value (0.89) reflects the importance of the user's demand information in the online factory sale platform recommendation system of the factory intermediary company. According to this weight value, the online factory sale platform of the factory intermediary company adjusts its recommendation algorithm to better meet the user's needs. For example: recommend low-price factory buildings that meet the user's preferences: prioritize recommending factory buildings within the user's acceptable price range. Recommend facility-complete factory buildings: ensure that the recommended factory buildings have a facility integrity rate higher than 90%. Respond quickly to house viewing needs: ensure that the house viewing process is fast and convenient to meet the user's expectation of fast house viewing. However, consider the developer brand more flexibly: since the user has no special preference for the developer brand, the platform can more flexibly recommend factory buildings of different developers. In this way, the online factory sale platform of the factory intermediary company can more accurately meet the user's needs, improve user satisfaction and platform operation efficiency.

[0040] In step S13, a plurality of target data in the multi-modal data are determined according to the content of the customer demand information, the information weight and the multi-modal data, and a final customer acquisition path is determined based on the plurality of target data, the enterprise database and the customer demand information; In the specific implementation process of the present application, the specific steps are: S131: Collect customer demand information, and mark the content and information weight of the customer demand information, determine the first target data combination according to the content of the customer demand information and the multi-modal data; S132: Determine the second target data combination according to the information weight of the customer demand information and the multi-modal data; determine a plurality of target data in the multi-modal data based on the matching of the first target data combination and the second target data combination; S133: Determine a plurality of pre-selected paths based on the matching of the plurality of target data and the enterprise database, determine the corresponding path matching coefficient according to the matching of the plurality of pre-selected paths and the customer demand information, and determine the final customer acquisition path according to the comparison of the path matching coefficients of the plurality of pre-selected paths.

[0041] In the embodiment of the present application, the customer demand information is collected, the content and information weight of the customer demand information are marked, and the first target data combination is determined according to the content of the customer demand information and the multi-modal data, which is compatible with the overall consideration of the content of the customer demand information and the multi-modal data, and ensures the accuracy of the first target data combination.

[0042] At this time, the customer demand information is comprehensively collected through questionnaire survey, customer interview, social media analysis, online behavior tracking and other ways; ensure that the collected information covers multiple aspects of the customer, including factory function demand, price sensitivity, purchase or rental intention, use scenario, preference, etc.; ensure that the collected data is accurate and reliable, and avoid misleading subsequent analysis; optionally, further control in combination with the customer's demand, and can be connected with the corresponding database of the enterprise, design the content of multiple enterprise modules, at the same time, support telephone sales, email, short message and other ways, so as to further screen the customers with intention.

[0043] The collected customer demand information is classified and marked, such as function demand, price demand, location demand, etc.; according to the importance, urgency or influence degree on customer purchase decision of the information, a weight value is allocated to each information; the weight value is represented by a numerical value (such as a floating point number between 0 and 1) or a grade (such as high, medium and low); ensure the objectivity and consistency of weight allocation, and avoid the influence of subjective bias; at this time, set the weight for each question in the questionnaire, such as the function demand question weight is 0.4, the price demand question weight is 0.3, the location demand question weight is 0.2, and the other questions are 0.1; in the customer interview, the importance of demand is judged according to the customer's speech and behavior, and the corresponding weight is allocated; the weight of social media and online behavior data is allocated, and the training and optimization are carried out based on historical data and customer feedback; Multi-modal data includes text, images, videos, audio, and other formats of data, which can provide more comprehensive and comprehensive customer information; according to the content of customer demand information, the data combination directly related to it is screened out from the multi-modal data, which should be able to fully reflect the customer's needs and expectations; the screened data is cleaned, sorted and standardized to ensure the accuracy and effectiveness of subsequent analysis; optionally, assuming that the customer demand information indicates that the customer has very high functional requirements for the product (weight 0.4) and a certain sensitivity to price (weight 0.3); from the multi-modal data, text descriptions, image displays, video demonstrations, and other data related to product functions, as well as price lists, promotional information, user reviews, and other data related to price are screened out; these data are combined to form the first target data combination, which will be used for subsequent customer demand analysis and product optimization; Specifically, assume that a factory intermediary company is developing a new factory recommendation system to better meet the needs of customers. The company collects customer demand information for factories through questionnaires, customer interviews, on-site viewing experiences, and online behavior tracking. Among them, customers have very high demand for the geographical location of the factory (such as proximity to transportation hubs, industrial parks, or city centers) (weight 0.4), and have a certain sensitivity to price (weight 0.3). At the same time, the company also collects user evaluations of the completeness of the facilities of the factory, the reputation of the developer, and after-sales service from social media, industry forums, and other channels.

[0044] Based on this information, the company screens out text descriptions, image displays, video demonstrations, and on-site viewing feedback related to geographical location, as well as price lists, promotional activities, user rental or purchase evaluations related to price from multi-modal data. These data are carefully combined to form the first target data combination; Further, the second target data combination is determined based on the information weight of the customer demand information and the multi-modal data; based on the matching of the first target data combination and the second target data combination, a plurality of target data in the multi-modal data is determined, which takes into account the matching of the first target data combination and the second target data combination, ensuring the accuracy of the plurality of target data in the multi-modal data.

[0045] At this time, based on the content of customer demand information and information weight that has been marked, further use these weights to filter multi-modal data; focus on those multi-modal data that are highly related to customer demand information and have high weight, which can more directly reflect the focus of customer demand; when filtering data, also ensure the diversity of data to comprehensively cover multiple dimensions of customer demand; at this time, preliminarily filter multi-modal data, exclude those data that are not related to customer demand information or have very low weight; according to the information weight, sort the remaining data and preferentially select data with higher weight; ensure that the selected data has certain diversity in type (such as text, image, video) and content to provide a more comprehensive perspective of customer information.

[0046] Match the first target data combination (data filtered based on customer demand information content) with the second target data combination (data filtered based on customer demand information weight), find out their common points and complementarity; based on the matching result, determine multiple target data in multi-modal data, which should be able to comprehensively and accurately reflect the customer's demand and expectation, and at the same time have high information value and practicality; verify the determined target data to ensure its accuracy and reliability, and avoid misleading subsequent analysis. At this time, cross-compare the first target data combination and the second target data combination to find out their overlapping parts and unique information points; considering the integrity, accuracy and practicality of data, select multiple target data from the matching result; use internal verification (such as data consistency check) and external verification (such as customer feedback) to verify the determined target data.

[0047] Therefore, based on the matching of multiple target data and enterprise database, multiple pre-selected paths are determined, according to the matching of multiple pre-selected paths and customer demand information, corresponding path matching coefficients are determined, and finally the customer acquisition path is determined according to the comparison of path matching coefficients of multiple pre-selected paths, which takes into account the overall comparison of path matching coefficients of multiple pre-selected paths, ensures the accuracy of the final customer acquisition path, at the same time, introduces multi-modal data, which takes into account the overall consideration of multiple target data, enterprise database and customer demand information, and improves the accuracy of customer acquisition path.

[0048] At this time, the previously determined multiple target data reflecting the needs and expectations of the customers are matched with the historical customer data, product data, marketing strategy data, etc. in the enterprise database; based on the matching results, the customer acquisition paths are explored, which involve different marketing channels, promotion methods, product combinations, price strategies, etc.; from the multiple paths explored, the paths with potential feasibility and attractiveness are selected as pre-selected paths. Optionally, the target data is cross-compared with the data in the enterprise database using data analysis tools; according to the comparison results, the marketing channels and promotion methods that successfully attracted similar customers in history are identified; combined with the current market trends, competitive environment and enterprise resources, multiple pre-selected paths with feasibility are screened.

[0049] The matching degree of each pre-selected path with the customer demand information is evaluated, which includes whether the path can directly meet the customer demand, whether it can improve the customer experience, whether it can bring a higher conversion rate, etc.; based on the evaluation results, a path matching coefficient is calculated for each pre-selected path, which is a numerical value (such as a floating point number between 0 and 1) representing the matching degree of the path with the customer demand; ensure that the path matching coefficients of all pre-selected paths are compared under the same standard to avoid the influence of subjective bias; at this time, a set of evaluation indexes are designed to measure the matching degree of the pre-selected path with the customer demand information; each pre-selected path is evaluated one by one, and the path matching coefficient is calculated according to the evaluation results; the calculated path matching coefficients are standardized to ensure that they are comparable on the same scale.

[0050] The path matching coefficients of multiple pre-selected paths are compared to find the path with the highest matching degree; in the comparison process, in addition to considering the path matching coefficient, other factors such as cost-effectiveness, implementation difficulty, time period, etc. are also considered for comprehensive consideration; based on the comparison results and comprehensive consideration, the final customer acquisition path is determined; at this time, a path matching coefficient comparison table is made to visually display the matching degree of each pre-selected path; organize team meetings to discuss and analyze the comparison table to ensure the comprehensiveness and accuracy of the decision; according to the discussion results, the final customer acquisition path is determined, and a detailed implementation plan is made.

[0051] In some embodiments of the present application, a weight matching table of matching points is introduced, which is shown in Table Three: Table Three Weight Matching Table of Matching Points Pre-selected path Matching point Weight Social media targeted advertising High overlap between target customer group and social media users 0.4 High relevance between advertising content and customer needs 0.3 High historical advertising conversion rate 0.2 High cost-effectiveness ratio of advertising 0.1 Outdoor forum cooperation promotion Forum users are mostly outdoor activity enthusiasts, matching the target customer group 0.5 Forum cooperation can directly reach potential customers and improve brand exposure 0.3 Cooperation cost is relatively low 0.1 Forum rules have fewer restrictions, with high marketing flexibility 0.1 Offline outdoor activity planning High relevance between activity content and customer needs, improving customer experience 0.6 Offline activities can enhance brand awareness and customer loyalty 0.3 Activity cost is relatively high, but long-term benefits are significant 0.1 Activity organization and execution are difficult, requiring consideration of multiple factors 0.0 For each pre-selected path, the scores of all its matching points (weight multiplied by matching degree, assuming matching degree is a value between 0 and 1) are added to obtain the path matching coefficient (total score); for example, if the scores of all the matching points of the social media targeted advertising are 0.4 (high coincidence), 0.24 (high correlation), 0.04 (high conversion rate, assuming conversion rate is 20%), and 0.02 (high cost-effectiveness ratio, assuming cost-effectiveness ratio is 20%), then the path matching coefficient is 0.7; assuming that after calculation, the path matching coefficient of the social media targeted advertising is 0.7, the outdoor forum cooperation promotion is 0.55, and the offline outdoor activity planning is 0.65; in this case, the social media targeted advertising has the highest score, and thus is selected as the final customer acquisition path; the company formulates a detailed social media advertising plan based on this, including target audience positioning, advertising content design, delivery time and budget, etc.

[0052] In step S14, the corresponding customer acquisition node is determined based on the traversal of the customer acquisition path, the corresponding customer information is determined according to the detection of the customer acquisition node, and the corresponding customer acquisition mode is determined based on the matching of each customer information and customer demand information; In the specific implementation process of the present application, the specific steps are as follows: S141: obtaining the customer acquisition path, traversing the customer acquisition path, determining the corresponding customer acquisition node according to the traversal of the customer acquisition path, at this time, detecting each customer acquisition node and determining the corresponding customer information; S142: collecting each customer information, determining the corresponding customer acquisition coefficient according to each customer information and customer demand information, and determining the corresponding customer acquisition mode according to the matching of the customer acquisition coefficient and the customer acquisition mode mapping relationship, at this time, the customer acquisition mode includes telephone customer acquisition, WeChat customer acquisition, email customer acquisition and associated person customer acquisition.

[0053] In the embodiment of the present application, the customer acquisition path is obtained, the customer acquisition path is traversed, the corresponding customer acquisition node is determined according to the traversal of the customer acquisition path, at this time, each customer acquisition node is detected and the corresponding customer information is determined, which is compatible with the overall consideration of the traversal of the customer acquisition path and ensures the accuracy of the corresponding customer acquisition node.

[0054] At this time, the customer acquisition path determined from the previous steps or decision-making process needs to be obtained, which is one or more specific marketing or sales activity sequences designed to attract and convert potential customers; the customer acquisition path includes online advertising, social media promotion, offline activities, email marketing, telephone marketing and other ways, which depends on the marketing strategy and target customer group of the company.

[0055] Traversing the customer acquisition path means analyzing each step or node in the order of the path, which helps understand the entire process from the customer's first contact with the brand to the final purchase. During the traversal, the role of each node, the customer's behavior at that node, and the conversion rate between nodes need to be focused on. During the traversal of the customer acquisition path, key customer acquisition nodes need to be identified, which are usually the key points where customers interact with the brand or make decisions. Customer acquisition nodes include ad display, click-through, registration, coupon redemption, purchase, etc.

[0056] After determining the customer acquisition nodes, each node needs to be detected to collect and analyze customer behavior data and information at that node, including customer click behavior, browsing history, registration information, transaction records, etc. By analyzing this information, a deeper understanding of customer needs, preferences, and behavior patterns can be gained, providing data support for subsequent marketing and sales activities.

[0057] Specifically, the online factory sales platform designs and implements a customer acquisition strategy to attract and convert potential customers. The determined customer acquisition path is: "social media factory ad → click-through to factory detail page → register user account → get a house viewing coupon → make an appointment to view or rent / purchase a factory"; During the traversal of this customer acquisition path, it is found that the click-through rate of social media factory ads is not satisfactory, but those who click on the ads often have a high interest and intention to rent or purchase a factory (i.e. they are more likely to register an account and eventually rent / purchase a factory). Therefore, it is decided to optimize the ad content to make it more attractive to improve the click-through rate. At the same time, each customer acquisition node is carefully detected, and extensive feedback information from potential customers is collected; In the registration account step, some problems are found, such as a complex registration process, user concerns about privacy protection, and a lack of intuitive registration progress prompts, which affect the user's registration experience and subsequent rental / purchase intention; To improve the user experience, a series of measures are taken: Simplify the registration process and reduce unnecessary fill-in items; Strengthen the transparency of the privacy protection policy and clearly inform users how their data will be used and protected; Increase real-time prompts for registration progress so that users can clearly see each step of the registration process; Provide immediate feedback, such as giving clear prompts after the user completes each step to enhance the user's operation experience. By optimizing the customer acquisition path and actively collecting and analyzing customer information, the conversion rate of potential customers and the willingness to rent / purchase have been successfully improved. This not only significantly increases the sales of the online factory sales platform, but also effectively improves brand awareness and customer satisfaction.

[0058] Further, various customer information is collected, a corresponding customer acquisition coefficient is determined according to the various customer information and customer demand information, and a corresponding customer acquisition mode is determined according to matching of the customer acquisition coefficient and the customer acquisition mode mapping relationship. At this time, the customer acquisition mode includes telephone customer acquisition, WeChat customer acquisition, email customer acquisition, and associated person customer acquisition, and the overall consideration of the matching of the customer acquisition coefficient and the customer acquisition mode mapping relationship is compatible, thereby ensuring the accuracy of the corresponding customer acquisition mode.

[0059] At this time, detailed information of the customer needs to be collected from various sources, including the customer's basic information (such as name, age, gender, contact information, etc.), purchase history, browsing behavior, preference settings, interaction records, etc.; the purpose of collecting customer information is to better understand the customer's needs, preferences and behavior patterns, thereby providing data support for subsequent marketing and sales activities; the collection methods include online forms, customer registration, transaction records, social media interactions, questionnaires, etc. At this time, the customer's purchase history and basic information are queried through the company's system or database; the customer's browsing behavior and page dwell time are collected through website analysis tools; the customer's interaction records and preference settings are obtained through social media platforms.

[0060] The customer acquisition coefficient is an index for measuring the potential value and purchase intention of the customer; it is derived based on the comparison and analysis of customer information and customer demand information; the customer acquisition coefficient involves multiple factors, such as the customer's purchase history, browsing behavior, interaction frequency, preference matching degree, etc.; through algorithms or models, these factors are converted into a specific numerical value (customer acquisition coefficient) for evaluating the potential value of the customer; at this time, the customer's purchase history (such as the number of purchases, purchase amount) and browsing behavior (such as page dwell time, number of clicks) are converted into the customer acquisition coefficient; the purchase history is given a higher weight because historical purchase behavior is an important indicator for predicting future purchase intention; through calculation, the customer acquisition coefficient of each customer is obtained for subsequent selection of customer acquisition modes.

[0061] The most suitable customer acquisition mode needs to be determined according to the customer acquisition coefficient and the preset customer acquisition mode mapping relationship; the customer acquisition mode mapping relationship is a table or rule set that defines which customer acquisition mode should be used for customers in different customer acquisition coefficient ranges; the customer acquisition mode includes telephone customer acquisition, WeChat customer acquisition, email customer acquisition, and associated person customer acquisition, etc.; at the same time, a customer acquisition mode mapping relationship table is collected, which divides the customer acquisition coefficient into different intervals and specifies one or more customer acquisition modes for each interval; for example, for customers with a high customer acquisition coefficient, telephone customer acquisition is preferred to quickly establish contact and promote transactions; for customers with an average customer acquisition coefficient but a certain interaction history, WeChat or email customer acquisition is used to maintain long-term communication and provide personalized recommendations; for customers introduced by associated persons or customers with special needs, associated person customer acquisition is used to utilize existing customer relationship networks and meet the special needs of customers.

[0062] Specifically, assuming that a marketing manager of an online factory sales platform is responsible for designing and implementing a customer acquisition strategy; detailed information of each customer has been collected, and the customer acquisition coefficient has been calculated according to the information; now, the most appropriate customer acquisition method needs to be determined according to the customer acquisition coefficient; a customer acquisition method mapping table is designed, as shown in Table Four: Table Four Customer Acquisition Method Mapping Table Customer acquisition coefficient range Customer acquisition method 0.9-1.0 Telephone customer acquisition 0.7-0.89 WeChat customer acquisition 0.5-0.69 Email customer acquisition 0.0-0.49 Associate customer acquisition / pending Assuming that the customer acquisition coefficient of a customer is 0.85, according to the customer acquisition method mapping table, the WeChat customer acquisition method will be adopted; personalized recommendations, preferential information or product updates are sent to the customer through WeChat to maintain long-term communication and improve the customer's purchase intention; in this way, the appropriate customer acquisition method is selected according to the customer's potential value and purchase intention, thereby improving the marketing efficiency and customer satisfaction.

[0063] In step S15, the corresponding customer acquisition quantity is determined according to the execution of the customer acquisition method, and the autonomous optimization of the customer acquisition method is triggered according to the customer acquisition quantity and the execution of the customer acquisition method until the customer acquisition quantity is greater than the preset customer acquisition quantity threshold; In the specific implementation process of the present application, the specific steps are as follows: S151: Collecting customer acquisition methods, and outputting customer demand information to customers along with the execution of the customer acquisition methods, real-time monitoring the execution of each customer acquisition method, and determining the corresponding customer acquisition quantity according to the analysis of the execution of each customer acquisition method; S152: Collecting the preset customer acquisition quantity threshold based on the detection of the enterprise database, and comparing each customer acquisition quantity with the preset customer acquisition quantity threshold; S153: If the customer acquisition quantity is lower than the preset customer acquisition quantity threshold, determining the corresponding optimization event according to the difference between the customer acquisition quantity and the preset customer acquisition quantity threshold, the customer acquisition method and the optimization measure mapping relationship, and triggering the autonomous optimization of the customer acquisition method according to the optimization event to regulate the customer acquisition quantity of the customer acquisition method until the customer acquisition quantity is greater than the preset customer acquisition quantity threshold.

[0064] In the embodiments of the present application, the customer acquisition methods are collected, and the customer demand information is outputted to the customers along with the execution of the customer acquisition methods, the execution of each customer acquisition method is real-time monitored, and the corresponding customer acquisition quantity is determined according to the analysis of the execution of each customer acquisition method, which is compatible with the overall consideration of the analysis of the execution of each customer acquisition method, and ensures the accuracy of the corresponding customer acquisition quantity.

[0065] At this time, it is necessary to collect and record all the customer acquisition methods currently used by the enterprise, including but not limited to telephone marketing, email marketing, social media advertising, search engine optimization (SEO), content marketing, offline activities, partner recommendations, etc.; Ensure that each customer acquisition method is clearly understood and recorded. For each customer acquisition method, it is necessary to design and output information closely related to customer needs, which means understanding the interests, needs, pain points, etc. of the target customers and customizing marketing information accordingly; Ensure that the marketing information is targeted and can arouse the interest and empathy of the customers.

[0066] Based on the monitoring of the performance of the customer acquisition methods, data analysis is needed to determine the number of customers brought in by each customer acquisition method, which is achieved by analyzing direct indicators such as registration volume and purchase volume after marketing activities; At the same time, some indirect indicators such as website traffic and brand awareness are also considered to evaluate the long-term effectiveness of customer acquisition methods.

[0067] Specifically, as the head of the marketing department of an online factory sales platform, the head recorded three customer acquisition methods: telephone marketing, email marketing, and social media advertising. For each customer acquisition method, marketing information closely related to the needs of potential customers for factories was designed, including factory type, area, price, location, facility configuration, promotional activities, rental or purchase options, etc. At the same time, CRM systems and email marketing software were used to track the performance of these marketing activities in real time; After a month of implementation, it was found that: telephone marketing successfully attracted 200 new customers to inquire about factory information or make online appointments to view houses; Email marketing brought 150 new customers to express their intention to buy or actually rent / purchase houses; Social media advertising prompted 100 users to click on links to learn more about factory information or directly enter the rental / purchase process.

[0068] Based on the above data, the effectiveness of each customer acquisition method was quantitatively evaluated: telephone marketing showed high conversion rates due to its directness and personalized communication; Email marketing achieved stable customer acquisition growth by accurately targeting the interests and needs of potential customers. Email content can be personalized based on customer browsing history and preferences to increase customer attention and response rates; Social media advertising, although slightly inferior in quantity, plays an important role in expanding brand exposure, attracting young entrepreneurs and small and medium-sized enterprises. Through creative advertising and multimedia content, it can effectively attract the attention of potential customers.

[0069] Next, based on these evaluation results, flexibly adjust the marketing strategy to improve the number and quality of customers: for telephone marketing, further optimize the language, increase the connection rate and conversion rate, and consider increasing the dialing period and frequency to cover more potential customers; for email marketing, analyze the opening rate, click rate, etc. Data, optimize email content to make it more attractive, and consider increasing personalized recommendations and dynamic content to improve user experience; for social media advertising, adjust the delivery strategy, such as more accurate targeting of target audience, optimize ad creative and copy, and use multimedia formats such as short videos to enhance the appeal and interactivity of the ad.

[0070] Further, based on the detection of the enterprise database, the preset customer acquisition quantity threshold is collected, and each customer acquisition quantity is compared with the preset customer acquisition quantity threshold.

[0071] At this time, it is necessary to access the enterprise's database to find and collect the preset customer acquisition quantity threshold, which is usually set in advance according to the business goals, market conditions, historical data, etc. Factors of the enterprise, used to evaluate the performance of different customer acquisition channels or activities; The threshold exists in the form of absolute quantity (such as the number of new customers per month) or relative proportion (such as conversion rate); After collecting the preset customer acquisition quantity threshold, the actual number of customers obtained by each customer acquisition channel or activity needs to be compared with the corresponding threshold. The purpose of this step is to evaluate whether the performance of the customer acquisition activity meets the expectations, so as to decide whether to take optimization measures.

[0072] Specifically, as a manager of a factory intermediary company, I am responsible for evaluating and optimizing the company's customer acquisition strategy. In order to ensure that the company can effectively attract potential customers, I obtain the preset customer acquisition quantity threshold from the company's CRM system or data analysis platform. These thresholds are set according to the company's business goals and market analysis, and are used to measure the performance of different customer acquisition channels.

[0073] Next, I compared the actual customer acquisition quantity (i.e. the number of potential customers consulting, reserving a house or actually renting / purchasing a factory) of the telephone marketing, email marketing and social media advertising in the last quarter with these preset thresholds.

[0074] Based on these key data, consider adjusting and optimizing the strategy from the following aspects: Adjust the marketing strategy: according to the performance difference of each channel, reallocate the marketing budget, increase the investment in channels with greater potential performance; For example, if social media advertising performs well among young factory buyers, consider increasing investment in this channel and developing more creative ads targeting young consumers; Optimize customer screening criteria: Analyze existing customer data in-depth to identify common characteristics of high-conversion customer groups, such as age, occupation, and factory preferences. Adjust customer screening criteria accordingly to improve the accuracy and efficiency of marketing activities. Increase the attractiveness of marketing information: Design more attractive marketing information for different customer groups, such as emphasizing superior geographical location, complete facilities, and price discounts. Use big data and artificial intelligence technology to achieve personalized recommendations and enhance the relevance and attractiveness of information. Strengthen channel synergy: Strengthen the synergy between telephone marketing, email marketing, and social media advertising. Integrate marketing information across channels to form a unified brand image and communication strategy, and improve overall customer acquisition results. Regular evaluation and adjustment: Establish a regular evaluation mechanism to continuously track changes in the number of customers acquired through each channel and conversion rates. Identify problems and make adjustments in a timely manner. Also, pay attention to market dynamics and competitor strategies to respond flexibly to market changes. Through such comparisons and analyses, you can more accurately assess the performance of customer acquisition activities, identify existing problems and opportunities, and make targeted optimization decisions. This will help factory intermediary companies attract potential customers more effectively, improve sales performance, and enhance market competitiveness.

[0075] Therefore, if the number of customers acquired is less than the preset threshold, the corresponding optimization event is determined according to the difference between the number of customers acquired and the preset threshold, the acquisition method, and the optimization measure mapping relationship, and the autonomous optimization of the acquisition method is triggered according to the optimization event to regulate the number of customers acquired through the acquisition method, until the number of customers acquired is greater than the preset threshold. This considers the overall difference between the number of customers acquired and the preset threshold, the acquisition method, and the optimization measure mapping relationship, improves the accuracy of the acquisition method, and allows autonomous optimization of the acquisition method to improve the effectiveness of intelligent customer acquisition.

[0076] At this point, the actual number of customers acquired through each channel or activity needs to be compared with the preset threshold; if the actual number of customers acquired is less than the threshold, the optimization process needs to be entered; at this point, the preset threshold is to acquire at least 500 new customers through social media advertising per month; however, only 450 new customers were acquired through social media advertising in the last month; therefore, the optimization process needs to be entered.

[0077] After determining that the number of customers acquired is less than the preset threshold, the difference between the actual number of customers acquired and the preset threshold needs to be calculated; this difference will help understand how many new customers need to be added to achieve the expected target; at the same time, the preset threshold is 500 new customers, while the actual number of customers acquired is 450; therefore, the difference is 50 new customers.

[0078] The specific optimization event is determined according to the mapping relationship between the previously established customer acquisition method and optimization measure and the calculated customer acquisition quantity difference, and the optimization event includes increasing the marketing budget, adjusting the marketing information, optimizing the target audience positioning, improving the customer experience, etc. Meanwhile, a mapping relationship between the customer acquisition method and the optimization measure is established. For example, if the number of customers acquired through social media advertising is less than the preset threshold, consider increasing the advertising budget, optimizing the advertising creative, adjusting the target audience positioning, etc. Based on this mapping relationship and the difference of 50 new customers, it is decided to increase the budget of social media advertising and optimize the advertising creative to attract more potential customers.

[0079] After determining the optimization event, the autonomous optimization mechanism of the customer acquisition method needs to be triggered, which means that the parameters of the marketing activity need to be adjusted, the marketing content needs to be updated, the resource input needs to be increased, etc. Ensure that the optimization measure can be implemented quickly and effectively. At the same time, it is decided to increase the budget of social media advertising and optimize the advertising creative. Therefore, cooperation with the advertising platform is needed to adjust the advertising budget and design new advertising creatives. At the same time, the performance of the advertisement also needs to be monitored to ensure the effectiveness of the optimization measure.

[0080] After triggering the autonomous optimization of the customer acquisition method, the change of the number of customers acquired needs to be monitored continuously. If the optimization measure is effective, the number of customers acquired should gradually increase. Once the number of customers acquired exceeds the preset threshold, consider maintaining the current marketing strategy or further exploring other customer acquisition channels. If the optimization measure is not effective, the optimization strategy needs to be adjusted and the autonomous optimization mechanism needs to be triggered again. At this time, the budget of social media advertising is increased and the advertising creative is optimized. After a period of monitoring, it is found that the number of customers acquired gradually increases and eventually exceeds the preset threshold of 500 new customers. Therefore, it is considered to maintain the current social media advertising strategy and continue to explore other effective customer acquisition channels.

[0081] In some embodiments of the present application, a collection optimization measure matching table is collected, which is shown in Table Five: Table Five Optimization Measure Matching Table Difference range Customer acquisition method Optimization measures (optimization events) 0-50 Social media Increase advertising frequency, optimize advertising copy 51-100 Search engine Increase keyword bid, optimize landing page design 101-200 Telephone marketing Extend call duration, increase customer follow-up frequency 201 or more Offline activities Expand activity scale, increase activity appeal Suppose the preset customer acquisition quantity threshold is 500, and the actual customer acquisition quantity is 450, the difference is 50. According to the matching table, the corresponding optimization measure is "increase the frequency of social media advertising and optimize the advertising copy"; therefore, the autonomous optimization process is triggered, the advertising strategy of social media advertising is adjusted, and the change of the number of customers acquired is continuously monitored.

[0082] Please refer to Figure 2 , Figure 2 is a structural composition schematic diagram of the intelligent customer acquisition system based on multi-modal data in the embodiments of the present application; the intelligent customer acquisition system based on multi-modal data comprises: A multi-modal data module 21 is configured to determine industry information, talent information and enterprise information according to traversal of the enterprise database, and construct corresponding multi-modal data based on the industry information, the talent information and the enterprise information; A customer demand information module 22 is configured to determine a corresponding demand space based on traversal of the enterprise database, determine corresponding customer demand information according to detection of the demand space, and mark information weights of the customer demand information; A customer acquisition path module 23 is configured to determine a plurality of target data in the multi-modal data according to content, information weights and the multi-modal data of the customer demand information, and determine a final customer acquisition path based on the plurality of target data, the enterprise database and the customer demand information; A customer acquisition method module 24 is configured to determine corresponding customer acquisition nodes based on traversal of the customer acquisition path, determine corresponding customer information according to detection of the customer acquisition nodes, and match corresponding customer acquisition methods based on each customer information and the customer demand information; An autonomous optimization module 25 is configured to determine a corresponding customer acquisition quantity according to execution of the customer acquisition method, trigger autonomous optimization of the customer acquisition method according to the customer acquisition quantity and execution of the customer acquisition method, until the customer acquisition quantity is greater than a preset customer acquisition quantity threshold.

[0083] Any combination of the technical features of the above embodiments is possible. In order to make the description concise, not all combinations of the technical features in the above embodiments are described, however, as long as the combinations of the technical features do not exist, they should be considered as the scope of the present disclosure.

Claims

1. An intelligent customer acquisition method based on multimodal data, characterized in that: include: Determine industry information, talent information, and enterprise information by traversing the enterprise database, and construct corresponding multimodal data based on the industry information, talent information, and enterprise information; Determine the corresponding demand space based on the traversal of the enterprise database, determine the corresponding customer demand information based on the detection of the demand space, and mark the information weight of the customer demand information; Determine multiple target data in the multimodal data based on the content, information weight, and multimodal data of the customer demand information, and determine the final customer acquisition path based on the multiple target data, the enterprise database, and the customer demand information; Determine the corresponding customer acquisition node based on the traversal of the customer acquisition path, determine the corresponding customer information based on the detection of the customer acquisition node, and match the corresponding customer acquisition method based on each customer information and customer demand information; The corresponding number of customers acquired is determined based on the execution of the customer acquisition method, and the autonomous optimization of the customer acquisition method is triggered based on the number of customers acquired and the execution of the customer acquisition method, until the number of customers acquired is greater than the preset customer acquisition number threshold.

2. The intelligent customer acquisition method based on multimodal data according to claim 1, characterized in that: The industry information, talent information, and enterprise information are determined based on the traversal of the enterprise database, and corresponding multimodal data are constructed based on the industry information, talent information, and enterprise information, including: Collect enterprise databases, determine information space based on the detection of enterprise databases, and determine industry information, talent information and enterprise information based on the detection of information space; Determine a first information interaction combination based on the industry information and the sector information, and determine a second information interaction combination based on the industry information and the enterprise information; In the enterprise database, corresponding multimodal data is constructed according to the first information interaction combination, the second information interaction combination, and the multimodal data mapping relationship.

3. The intelligent customer acquisition method based on multimodal data according to claim 1, characterized in that: The process of determining a corresponding demand space based on traversing the enterprise database, determining corresponding customer demand information based on detection of the demand space, and marking the information weight of the customer demand information includes: Collect enterprise databases, determine the corresponding demand space based on the traversal of the enterprise database, and trigger real-time detection of the demand space; Determine a demand information sequence list based on real-time detection of demand space, and determine corresponding customer demand information based on the detection of the demand information sequence list; Based on the analysis of the customer demand information, a plurality of demand features are determined, and the information weight of the customer demand information is determined according to the plurality of demand features and corresponding weight mapping relationships.

4. The intelligent customer acquisition method based on multimodal data according to claim 1, characterized in that: The step of determining multiple target data in the multimodal data based on the content, information weight, and multimodal data of the customer demand information, and determining a final customer acquisition path based on the multiple target data, the enterprise database, and the customer demand information, includes: Collect customer demand information, mark the content and information weight of the customer demand information, and determine a first target data combination based on the content of the customer demand information and multimodal data; A second target data combination is determined according to the information weight of the customer demand information and the multimodal data; and a plurality of target data in the multimodal data is determined based on a match between the first target data combination and the second target data combination.

5. The intelligent customer acquisition method based on multimodal data according to claim 4, characterized in that: The step of determining multiple target data in the multimodal data based on the content, information weight, and multimodal data of the customer demand information, and determining a final customer acquisition path based on the multiple target data, the enterprise database, and the customer demand information, further includes: Based on the matching of multiple target data and enterprise databases, multiple pre-selected paths are determined; based on the matching of multiple pre-selected paths and customer demand information, corresponding path matching coefficients are determined; and based on the comparison of the path matching coefficients of multiple pre-selected paths, the final customer acquisition path is determined.

6. The intelligent customer acquisition method based on multimodal data according to claim 1, characterized in that: The method of determining the corresponding customer acquisition node based on the traversal of the customer acquisition path, determining the corresponding customer information based on the detection of the customer acquisition node, and matching the corresponding customer acquisition method based on each customer information and customer demand information includes: The customer acquisition path is obtained and traversed, and the corresponding customer acquisition node is determined based on the traversal of the customer acquisition path. At this time, each customer acquisition node is detected and the corresponding customer information is determined.

7. The intelligent customer acquisition method based on multimodal data according to claim 6, characterized in that: The method further includes determining the corresponding customer acquisition node based on the traversal of the customer acquisition path, determining the corresponding customer information based on the detection of the customer acquisition node, and matching the corresponding customer acquisition method based on each customer information and customer demand information. Collect information from each customer, determine the corresponding customer acquisition coefficient based on the information and customer demand information, and determine the corresponding customer acquisition method based on the matching relationship between the customer acquisition coefficient and the customer acquisition method. At this time, the customer acquisition methods include telephone acquisition, WeChat acquisition, email acquisition, and associated person acquisition.

8. The intelligent customer acquisition method based on multimodal data according to claim 1, characterized in that: The determining of the corresponding number of customers acquired based on the execution of the customer acquisition method, and triggering autonomous optimization of the customer acquisition method based on the number of customers acquired and the execution of the customer acquisition method, until the number of customers acquired exceeds a preset customer acquisition number threshold, includes: Collect customer acquisition methods and output customer demand information to customers in a targeted manner based on the execution of the customer acquisition methods. Monitor the execution of each customer acquisition method in real time and determine the corresponding number of customers acquired based on the analysis of the execution of each customer acquisition method. The preset customer acquisition threshold is collected based on the detection of the enterprise database, and the number of each customer acquired is compared with the preset customer acquisition threshold.

9. The intelligent customer acquisition method based on multimodal data according to claim 8, characterized in that: The method further includes determining the corresponding number of acquired customers based on the execution of the customer acquisition method, triggering autonomous optimization of the customer acquisition method based on the number of acquired customers and the execution of the customer acquisition method, until the number of acquired customers exceeds a preset customer acquisition number threshold, and further includes: If the number of customers acquired is lower than the preset customer acquisition threshold, the corresponding optimization event is determined based on the difference between the number of customers acquired and the preset customer acquisition threshold, the mapping relationship between the customer acquisition method and the optimization measures, and the autonomous optimization of the customer acquisition method is triggered according to the optimization event to regulate the number of customers acquired by the customer acquisition method until the number of customers acquired is greater than the preset customer acquisition threshold.

10. An intelligent customer acquisition system based on multimodal data, characterized in that: The intelligent customer acquisition system based on multimodal data is applied to the intelligent customer acquisition method based on multimodal data according to any one of claims 1 to 9, and the intelligent customer acquisition system based on multimodal data includes: The multimodal data module is used to determine industry information, talent information and enterprise information based on the traversal of the enterprise database, and to construct corresponding multimodal data based on the industry information, talent information and enterprise information; The customer demand information module is used to determine the corresponding demand space based on the traversal of the enterprise database, determine the corresponding customer demand information based on the detection of the demand space, and mark the information weight of the customer demand information; A customer acquisition path module is used to determine multiple target data in the multimodal data based on the content, information weight and multimodal data of the customer demand information, and to determine the final customer acquisition path based on the multiple target data, the enterprise database and the customer demand information; The customer acquisition method module is used to determine the corresponding customer acquisition node based on the traversal of the customer acquisition path, determine the corresponding customer information based on the detection of the customer acquisition node, and match the corresponding customer acquisition method based on the customer information and customer demand information; The autonomous optimization module is used to determine the corresponding number of customers acquired based on the execution of the customer acquisition method, and trigger autonomous optimization of the customer acquisition method based on the number of customers acquired and the execution of the customer acquisition method until the number of customers acquired is greater than the preset customer acquisition number threshold.

Images