Precision marketing method and device, equipment, storage medium and program product
By constructing user profiles and decomposing marketing strategies into atomic capabilities, and combining multiple recommendation algorithms for precision marketing, the problems of incomplete user profiles and limited recommendation accuracy are solved, achieving personalized and precise marketing results.
Patent Information
- Application Number
- CN202510262918.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-06
- Publication Date
- 2025-12-16
AI Technical Summary
Existing precision marketing methods suffer from incomplete user profiles and limited recommendation accuracy.
By acquiring user data from multiple marketing platforms, user profiles are constructed, multiple marketing strategies are generated, and the strategies are decomposed into atomic capabilities and strategy groups. Logical operations are used to combine them into strategy groups, which are then combined with filtering algorithms, matrix factorization algorithms, and content recommendation algorithms for precision marketing.
It improved the completeness of user profiles and the accuracy of recommendations, enabling personalized recommendations and precise targeting of target users, thereby enhancing marketing effectiveness.
Smart Images
Figure CN121146852A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of precision marketing technology, and in particular to a precision marketing method, apparatus, equipment, storage medium, and program product. Background Technology
[0002] To meet users' personalized needs and improve conversion rates and engagement, marketing systems typically need to incorporate precision marketing. Current precision marketing solutions are applicable to single marketing platforms, relying on the platform's massive user data and employing traditional modeling and algorithm optimization to provide personalized marketing content. However, this approach suffers from issues such as incomplete user profiles and limited recommendation accuracy. Summary of the Invention
[0003] This application provides a precision marketing method, device, equipment, storage medium, and program product to address the shortcomings of existing precision marketing methods, such as incomplete user profiles and limited recommendation accuracy.
[0004] This application provides a precision marketing method, including: Acquire user data from multiple marketing platforms and construct user profiles based on the user data; the user data includes behavioral data and transaction data; Based on the user profile and the business objectives of each marketing platform, multiple marketing strategies are generated. Each of the marketing strategies is decomposed into atomic capabilities, and the multiple marketing strategies are combined into a strategy group through logical operations; the atomic capabilities include user screening and product recommendation. Based on the atomic capabilities and the strategy groups, targeted marketing is conducted to the target users of each of the marketing platforms.
[0005] In one embodiment, the step of conducting targeted marketing to target users of each of the marketing platforms based on the atomic capabilities and the strategy group includes: The atomic capabilities and the strategy groups are each encapsulated into independent interface services; the interface services are used to provide a unified calling interface and parameter specifications. Based on the business needs of each marketing platform, the calling process of the interface service is configured; the calling process includes the dependencies and calling order between the interface services; the interface service is called by each marketing platform according to the calling process to obtain the initial recommendation results for the target users of the marketing platform. The initial recommendation results are filtered and selected based on a preset recommendation algorithm in order to conduct precise marketing to the target users.
[0006] In one embodiment, the recommendation algorithm includes a filtering algorithm; the filtering of the initial recommendation results based on a preset recommendation algorithm to conduct precise marketing to the target users includes: Based on a preset recommendation algorithm, the user similarity between users of each marketing platform and the product similarity between products of each marketing platform are calculated. A neighbor list is constructed for each user based on the user similarity, and a list of similar products is constructed for each product based on the product similarity. Obtain the rated products of the target users of the marketing platform, as well as the target users' list of target neighbors; The initial recommendation results are filtered and selected based on the first product list and the second product list to obtain a list of products to be recommended; the first product list is a list of similar products to the rated products, and the second product list is a list of products ordered by each user in the target neighbor list; Based on the list of users to be recommended, targeted marketing is conducted to the target users.
[0007] In one embodiment, the recommendation algorithm includes a matrix factorization algorithm; the filtering and screening of the initial recommendation results based on a preset recommendation algorithm to conduct precise marketing to the target users includes: Based on a preset recommendation algorithm, a rating matrix between users and products of each of the aforementioned marketing platforms is generated; The scoring matrix is decomposed into a user feature matrix and a business feature matrix; Based on the user feature matrix and the business feature matrix, the initial recommendation results are filtered and selected to obtain a list of recommendations to be made; Based on the list of users to be recommended, targeted marketing is conducted to the target users.
[0008] In one embodiment, the recommendation algorithm includes a content recommendation algorithm; the filtering and selection of the initial recommendation results based on a preset recommendation algorithm for targeted marketing to the target users includes: Based on a preset recommendation algorithm, the product description text information of each marketing platform is encoded to obtain a product feature vector, and the historical behavior data of users of each marketing platform is encoded to obtain a user preference vector. Calculate the similarity between the target user's user preference vector and the product feature vector, and determine the target user's list of products of interest based on the similarity. The initial recommendation results are filtered and selected based on the list of products of interest to obtain a list of products to be recommended. Based on the list of users to be recommended, targeted marketing is conducted to the target users.
[0009] In one embodiment, after combining the atomic capabilities and the strategy to conduct targeted marketing to the target users of each of the marketing platforms, the method further includes: Monitor the feedback information of the interface service; the feedback information includes the execution status and execution result of the interface service. The interface service is optimized based on the feedback information.
[0010] This application also provides a precision marketing device, including the following modules: The user profile building module is used to acquire user data from multiple marketing platforms and build user profiles based on the user data; the user data includes behavioral data and transaction data. The marketing strategy generation module is used to generate multiple marketing strategies based on the user profile and the business objectives of each marketing platform. The marketing strategy processing module is used to decompose each of the marketing strategies into atomic capabilities, and combine the multiple marketing strategies into a strategy group through logical operations; the atomic capabilities include user screening and product recommendation. The precision marketing execution module is used to conduct precision marketing to target users of each of the marketing platforms based on the atomic capabilities and the strategy groups.
[0011] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the precision marketing method as described above.
[0012] This application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the precision marketing method as described above.
[0013] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the precision marketing method as described above.
[0014] The precision marketing method, apparatus, equipment, storage medium, and program products provided in this application construct user profiles by acquiring user data from multiple marketing platforms, thereby improving the completeness of the user profiles. Based on the constructed user profiles and the business objectives of the marketing platforms, multiple marketing strategies are generated. The marketing strategies are decomposed into atomic capabilities and combined into strategy groups through logical operations. These groups are used for precision marketing to target users of the marketing platforms, achieving personalized recommendations and precise reach to target users, and improving the effectiveness of precision marketing. Attached Figure Description
[0015] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 This is a flowchart illustrating the precision marketing method provided in the embodiments of this application.
[0017] Figure 2 This is a schematic diagram of the interface service call process provided in the embodiments of this application.
[0018] Figure 3 This is a schematic diagram of the precision marketing device provided by the present invention.
[0019] Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0021] This application provides a precision marketing method that improves the accuracy and effectiveness of marketing campaigns by integrating marketing data from multiple marketing platforms and combining a two-level strategy data structure. Specifically, Figure 1 This is a flowchart illustrating the precision marketing method provided in the embodiments of this application, such as... Figure 1 As shown, the method includes the following steps: Step 100: Obtain user data from multiple marketing platforms and construct user profiles based on the user data; the user data includes behavioral data and transaction data; Step 200: Based on the user profile and the business objectives of each marketing platform, generate multiple marketing strategies; Step 300: Decompose each of the marketing strategies into atomic capabilities, and combine the multiple marketing strategies into a strategy combination through logical operations; the atomic capabilities include user screening and product recommendation; Step 400: Based on the atomic capabilities and the strategy combination, conduct precision marketing to the target users of each of the marketing platforms.
[0022] Acquire user data from multiple marketing platforms and construct user profiles based on this data. The acquired user data includes behavioral and transaction data. User behavioral data includes search history, reviews, and selected service packages, while transaction data includes information on products ordered historically. Optionally, the constructed user profiles may include user interests, purchasing preferences, and spending power.
[0023] In one embodiment, the acquired user data includes data from multiple users, including behavioral and transaction data of each user across multiple marketing platforms. Based on the behavioral and transaction data of the same user across multiple marketing platforms, a user profile is constructed to improve the completeness of the user profile.
[0024] Based on the constructed user profiles and the business objectives of each marketing platform, multiple marketing strategies are generated. Optionally, one marketing platform may correspond to one or more marketing strategies, and different marketing platforms may correspond to the same marketing strategy or different marketing strategies. Each marketing strategy has clearly defined objectives, inputs, outputs, and expected results.
[0025] Each marketing strategy is broken down into atomic capabilities, and multiple marketing strategies are combined into strategy groups through logical operations. The atomic capabilities include user filtering and product recommendation, and each atomic capability is configured with corresponding execution processes and parameters. The logical operations used to combine marketing strategies include intersection, union, and complement operations.
[0026] A two-level strategy data structure of atomic capabilities and strategy groups is adopted to store intelligent scheduling methods. The first-level strategy data stores atomic recommendation strategy data, while the second-level strategy data is strategy group data composed of multiple marketing strategies, which stores scheduling and processing data. This enables the scheduling of capabilities of multiple marketing platforms, as well as intelligent data filtering and integration between multiple marketing platforms.
[0027] Based on atomic capabilities and strategy groups, targeted marketing is conducted for target users across various marketing platforms. The target user is any one of multiple users across multiple marketing platforms. For any given user, a two-tiered strategy structure using atomic capabilities and strategy groups is employed to schedule strategies for that user, generating accurate and comprehensive recommendation results, thus achieving targeted marketing.
[0028] In this embodiment, user profiles are constructed by acquiring user data from multiple marketing platforms, which improves the completeness of the user profiles. Based on the constructed user profiles and the business objectives of the marketing platforms, multiple marketing strategies are generated. The marketing strategies are decomposed into atomic capabilities and combined into strategy groups through logical operations. These groups are used to conduct precise marketing to the target users of the marketing platforms, thereby achieving personalized recommendations and precise reach to the target users and improving the effectiveness of precision marketing.
[0029] In one embodiment, user data from multiple marketing platforms is standardized. This standardization process includes data cleaning, format conversion, and field mapping to ensure data consistency and availability.
[0030] Optionally, step 400 may also include: Step 410: Encapsulate the atomic capabilities and the policy groups into independent interface services; the interface services are used to provide a unified calling interface and parameter specifications. Step 420: Configure the calling process of the interface service according to the business needs of each marketing platform; the calling process includes the dependencies and calling order between the interface services; the interface service is called by each marketing platform according to the calling process to obtain the initial recommendation results for the target users of the marketing platform; Step 430: Filter the initial recommendation results based on a preset recommendation algorithm to conduct precise marketing to the target users.
[0031] Each atomic capability and strategy group is encapsulated as an independent interface service. This interface service provides a unified calling interface and parameter specification. Based on the business needs of each marketing platform, the calling flow of the interface service is configured. This calling flow includes the dependencies and calling order between interface services. The interface service is invoked by the marketing platform according to the configured calling flow to obtain initial recommendation results for the target users of the marketing platform.
[0032] Furthermore, based on a pre-defined recommendation algorithm, the initial recommendation results are filtered and selected to achieve precise marketing to target users. In one embodiment, when the corresponding precise marketing method is applied to different marketing platforms, communication between platforms is achieved through standardized data interfaces to promote data sharing and business collaboration. However, cross-platform architectures are limited by insufficient scalability and flexibility, making it difficult to quickly adapt to rapid changes in market demands, and require continuous technology updates and architectural restructuring.
[0033] In this embodiment, by encapsulating interface services and providing a unified calling interface and parameter specification, precision marketing methods can be applied to different marketing platforms. This facilitates the flexibility of cross-platform expansion of precision marketing, enabling rapid adaptation to quickly changing market demands. It not only achieves intelligent scheduling of capabilities across multiple marketing platforms but also improves the accuracy and effectiveness of marketing campaigns through intelligent data filtering and integration.
[0034] When filtering the initial recommendation results, different recommendation algorithms can be used. Optionally, the recommendation algorithm used may include a filtering algorithm. Based on this, step 430 may include: Step 501: Based on a preset recommendation algorithm, calculate the user similarity between users of each marketing platform and the product similarity between products of each marketing platform. Step 502: Construct a neighbor list for each user based on the user similarity, and construct a similar product list for each product based on the product similarity; Step 503: Obtain the rated products of the target users of the marketing platform, and the target neighbor list of the target users; Step 504: Filter the initial recommendation results according to the first product list and the second product list to obtain a list to be recommended; the first product list is a list of similar products to the rated products, and the second product list is a list of products ordered by each user in the target neighbor list; Step 505: Conduct targeted marketing to the target users based on the list of users to be recommended.
[0035] Based on the filtering algorithm, the similarity between users on each marketing platform and the similarity between products on each marketing platform are calculated. A neighbor list for each user is constructed based on the user similarity, and a list of similar products for each product is constructed based on the product similarity.
[0036] Furthermore, the system obtains the rated products of the target users on the marketing platform, as well as the target users' target neighbor list, which includes multiple neighboring users of the target users. Based on the first and second product lists, the initial recommendation results are filtered to obtain a list of products to be recommended. Targeted marketing is then conducted on the target users based on this list of products to be recommended.
[0037] The first product list is a list of similar products to the rated products, and the second product list is a list of products ordered by each neighboring user in the target neighbor list.
[0038] Optionally, the initial recommendation results contain multiple recommendation items. Based on the products already ordered by similar users of the target user, as well as the products rated by the target user and their similar products, the products to be recommended corresponding to the multiple recommendation items in the initial recommendation results are sorted, and a preset number or preset proportion of products to be recommended are filtered out according to the sorting order to obtain a list of products to be recommended, thereby achieving the filtering of the initial recommendation results.
[0039] Methods for calculating user and product similarity include, but are not limited to, cosine similarity and Pearson correlation coefficient. First, the similarity between all user pairs is calculated. Using cosine similarity and Pearson correlation coefficient, similarity values for user pairs are calculated based on user data, thereby constructing a neighbor list for each user. Each user's neighbor list includes other users most similar to them. In the product recommendation phase, products that the target user might subscribe to are predicted based on the products already ordered by similar users in the neighbor list, thus enabling personalized recommendations for the target user.
[0040] For a list of similar products, first calculate the similarity for all product pairs, then create a list of similar products for each product. When product recommendations are needed, retrieve the products that the user has rated, and make recommendations based on the list of similar products for these products.
[0041] In this context, a user pair consists of any two users, and a product pair consists of any two products.
[0042] The recommendation algorithm used to filter and select the initial recommendation results also includes a matrix factorization algorithm. Based on this, step 430 further includes: Step 601: Based on a preset recommendation algorithm, generate a rating matrix between users and products of each of the marketing platforms; Step 602: Decompose the scoring matrix into a user feature matrix and a business feature matrix; Step 603: Based on the user feature matrix and the business feature matrix, filter and select the initial recommendation results to obtain a list of recommendations to be made; Step 604: Conduct targeted marketing to the target users based on the list of users to be recommended.
[0043] Based on the matrix factorization algorithm, a rating matrix between users and products of each marketing platform is generated. This rating matrix is then decomposed into a user feature matrix and a business feature matrix. The rating matrix is the product of the user feature matrix and the business feature matrix.
[0044] Based on the user feature matrix and business feature matrix, the initial recommendation results are filtered to obtain a list of recommendations to be made, and then targeted marketing is carried out on the target users based on this list.
[0045] Matrix factorization, including but not limited to Singular Value Decomposition (SVD) and Alternating Least Squares (ALS), is used to handle large-scale datasets. Matrix factorization algorithms mine latent feature vectors by decomposing the user-product rating matrix. These feature vectors characterize the properties of users and products in the latent factor space. For example, using SVD, the rating matrix can be decomposed into the product of a user feature matrix and a set of product feature matrices, which can effectively handle sparse data and improve the accuracy and efficiency of recommendations.
[0046] The initial recommendation results can be filtered and selected according to a ranking algorithm. Based on factors such as the predicted scores (e.g., user rating predictions) represented by the user feature matrix and business feature matrix, diversity indicators (e.g., product heterogeneity in the recommendation list), and novelty, the recommended items in the initial recommendation results can be ranked to ensure that the recommendation list is both accurate and attractive.
[0047] The recommendation algorithm used to filter the initial recommendation results may also include a content recommendation algorithm. Based on this, step 430 may also include: Step 701: Based on a preset recommendation algorithm, the product description text information of each marketing platform is encoded to obtain a product feature vector, and the historical behavior data of users of each marketing platform is encoded to obtain a user preference vector. Step 702: Calculate the similarity between the target user's user preference vector and the product feature vector, and determine the target user's list of products of interest based on the similarity. Step 703: Filter the initial recommendation results according to the list of products of interest to obtain a list of products to be recommended; Step 704: Conduct targeted marketing to the target users based on the list of users to be recommended.
[0048] Based on content recommendation algorithms, the product description text information of each marketing platform is encoded to obtain the product feature vector of each product, and the historical behavior data of users of each marketing platform is encoded to obtain the user preference vector.
[0049] The similarity between the target user's user preference vector and the product feature vectors of each product is calculated. Based on this similarity, the target user is placed in a list of products of interest. The initial recommendation results are then filtered based on the target user's list of products of interest to obtain a list of products to be recommended. Targeted marketing is then conducted on the target user based on this list of products to be recommended.
[0050] Optionally, the content recommendation algorithm is implemented based on the BERT (Bidirectional Encoder Representations from Transformers) model. BERT is a pre-trained deep learning model used for natural language processing tasks. Its core advantage lies in its ability to capture deep semantic relationships in text and understand the multiple meanings of words in different contexts.
[0051] In content-based recommendation algorithms, BERT can be used to encode product description text information, such as detailed product information and service features, converting the product description text information into a high-dimensional vector. Similarly, user preferences can be encoded by analyzing users' historical behavioral data, such as their search history, reviews, and selected service packages, to obtain user preference vectors.
[0052] Product descriptions are processed by the BERT model and transformed into a series of numerical vectors, capturing key product information and contextual semantics. User behavior is also converted into vectors by the BERT model to represent user product interests and needs.
[0053] When recommending products, the similarity between product feature vectors and user preference vectors is calculated to identify the products that best match the user's interests. Methods for calculating similarity include, but are not limited to, cosine similarity and Euclidean distance. Based on the similarity score, one or more recommendable products are obtained, and the product with the highest similarity score is selected for recommendation.
[0054] Content-based recommendation algorithms, by deeply understanding the semantic level of product content and user preferences, can provide users with more accurate and in-depth personalized recommendations. For new users and products, content-based recommendation algorithms can quickly process and recommend content without relying on the interaction data of other users.
[0055] In one embodiment, product recommendations can be based on predefined rules, which can be dynamically adjusted according to the user's real-time behavior. Optionally, the predefined rules can be configured as triggers; for example, if a user continuously exceeds their data allowance for three months, a data upgrade package can be automatically recommended. In another embodiment, more accurate personalized recommendations can be provided by utilizing the user's current contextual information, such as location, time, and device type.
[0056] Optionally, the recommendation results can be pre-generated, based on a trained model, by pre-generating some recommendation results or optimizing model parameters to reduce the computational load of online recommendations. Integrating offline computation results with real-time data provides comprehensive data support for online recommendations.
[0057] As business continues to grow, user feedback and product information can be continuously collected, and the semantic understanding capabilities of the BERT model can be leveraged to continuously optimize the accuracy of product recommendations. Based on this, after step 400, the following may also be included: Step 401: Monitor the feedback information of the interface service; the feedback information includes the execution status and execution result of the interface service; Step 402: Optimize the interface service based on the feedback information.
[0058] Monitor feedback information from each interface service, including its execution status and results. Based on this feedback, optimize the interface services, including improvements to atomic capabilities and policy groups. By monitoring the execution status and results of interface services in real time, optimize strategies and services based on the feedback information.
[0059] In one embodiment, refer to Figure 2 The illustrated API call flow involves obtaining the corresponding API configuration information (including input and output parameters) based on the API call object. The API input parameters are then assembled to call the API address, and the API configuration rules are obtained. Further, the API configuration rules are matched against the rules of each configured API service. If a match is found, the API configuration rules exist. Following the API call flow, the configuration information for the next API is obtained. Similarly, the input parameters for the next API are assembled to call the API address, and the configuration rules for the next API are obtained. If the configuration rules for the next API match, API processing information is obtained. Based on this processing information, the processing result is fed back to the API call object, completing the API call and providing the corresponding API service.
[0060] In this embodiment, for cross-platform data isolation, when data between platforms cannot be shared, an independent data processing and scheduling mechanism is built by encapsulating interface services. This ensures that data integration, analysis and scheduling can be achieved without directly accessing the original data, thereby protecting the data security and privacy of each platform.
[0061] With flexible interface adaptation and data integration capabilities, it can seamlessly connect to the interfaces of multiple marketing platforms, enabling the extraction, transformation, and loading of data from different platforms. Simultaneously, its data integration capabilities can consolidate data from various marketing platforms and in different formats into a unified data model, providing a foundation for intelligent scheduling of subsequent product recommendations.
[0062] The dual-intelligent scheduling mechanism not only enables intelligent product recommendation but also introduces strategy scheduling, forming a dual mechanism of recommendation and scheduling strategies. This allows for flexible adjustment of recommendation and scheduling strategies based on business needs and market changes, thereby improving the accuracy and efficiency of marketing campaigns.
[0063] Furthermore, a two-level strategy data structure is adopted. The first-level strategy data stores atomic recommendation strategy data, ensuring the basic nature and flexibility of the strategy. The second-level strategy data combines multiple recommendation strategies into strategy groups, realizing collaborative work between strategies and intelligent data filtering and integration. This not only improves the scalability and maintainability of precision marketing, but also greatly improves the efficiency and accuracy of strategy execution.
[0064] By using multiple recommendation algorithms to accurately filter and sort the recommendation results, removing recommendations that do not match the user's interests, and recommending the recommendations that best meet the user's needs, the personalization and relevance of the recommendation results are greatly improved, thereby enhancing the user experience and satisfaction.
[0065] The precision marketing device provided in the embodiments of this application is described below. The precision marketing device described below can be referred to in correspondence with the precision marketing method described above.
[0066] Reference Figure 3 The precision marketing device provided in this application includes: User profile building module 10 is used to acquire user data from multiple marketing platforms and build user profiles based on the user data; the user data includes behavioral data and transaction data; The marketing strategy generation module 30 is used to generate multiple marketing strategies based on the user profile and the business objectives of each marketing platform. The marketing strategy processing module 30 is used to decompose each of the marketing strategies into atomic capabilities, and combine the multiple marketing strategies into a strategy group through logical operations; the atomic capabilities include user screening and product recommendation. The precision marketing execution module 40 is used to conduct precision marketing to target users of each of the marketing platforms based on the atomic capabilities and the strategy groups.
[0067] In one embodiment, the precision marketing execution module 40 is further configured to: The atomic capabilities and the strategy groups are each encapsulated into independent interface services; the interface services are used to provide a unified calling interface and parameter specifications. Based on the business needs of each marketing platform, the calling process of the interface service is configured; the calling process includes the dependencies and calling order between the interface services; the interface service is called by each marketing platform according to the calling process to obtain the initial recommendation results for the target users of the marketing platform. The initial recommendation results are filtered and selected based on a preset recommendation algorithm in order to conduct precise marketing to the target users.
[0068] In one embodiment, the recommendation algorithm includes a filtering algorithm; the precision marketing execution module 40 is further configured to: Based on a preset recommendation algorithm, the user similarity between users of each marketing platform and the product similarity between products of each marketing platform are calculated. A neighbor list is constructed for each user based on the user similarity, and a list of similar products is constructed for each product based on the product similarity. Obtain the rated products of the target users of the marketing platform, as well as the target users' list of target neighbors; The initial recommendation results are filtered and selected based on the first product list and the second product list to obtain a list of products to be recommended; the first product list is a list of similar products to the rated products, and the second product list is a list of products ordered by each user in the target neighbor list; Based on the list of users to be recommended, targeted marketing is conducted to the target users.
[0069] In one embodiment, the recommendation algorithm includes a matrix factorization algorithm; the precision marketing execution module 40 is further configured to: Based on a preset recommendation algorithm, a rating matrix between users and products of each of the aforementioned marketing platforms is generated; The scoring matrix is decomposed into a user feature matrix and a business feature matrix; Based on the user feature matrix and the business feature matrix, the initial recommendation results are filtered and selected to obtain a list of recommendations to be made; Based on the list of users to be recommended, targeted marketing is conducted to the target users.
[0070] In one embodiment, the recommendation algorithm includes a content recommendation algorithm; the precision marketing execution module 40 is further configured to: Based on a preset recommendation algorithm, the product description text information of each marketing platform is encoded to obtain a product feature vector, and the historical behavior data of users of each marketing platform is encoded to obtain a user preference vector. Calculate the similarity between the target user's user preference vector and the product feature vector, and determine the target user's list of products of interest based on the similarity. The initial recommendation results are filtered and selected based on the list of products of interest to obtain a list of products to be recommended. Based on the list of users to be recommended, targeted marketing is conducted to the target users.
[0071] In one embodiment, the precision marketing device further includes a marketing strategy optimization module 40, used for: Monitor the feedback information of the interface service; the feedback information includes the execution status and execution result of the interface service. The interface service is optimized based on the feedback information.
[0072] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communications interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a precision marketing method, which includes: Acquire user data from multiple marketing platforms and construct user profiles based on the user data; the user data includes behavioral data and transaction data; Based on the user profile and the business objectives of each marketing platform, multiple marketing strategies are generated. Each of the marketing strategies is decomposed into atomic capabilities, and the multiple marketing strategies are combined into a strategy group through logical operations; the atomic capabilities include user screening and product recommendation. Based on the atomic capabilities and the strategy groups, targeted marketing is conducted to the target users of each of the marketing platforms.
[0073] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0074] On the other hand, embodiments of this application also provide a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the precision marketing method provided by the above methods, which includes: Acquire user data from multiple marketing platforms and construct user profiles based on the user data; the user data includes behavioral data and transaction data; Based on the user profile and the business objectives of each marketing platform, multiple marketing strategies are generated. Each of the marketing strategies is decomposed into atomic capabilities, and the multiple marketing strategies are combined into a strategy group through logical operations; the atomic capabilities include user screening and product recommendation. Based on the atomic capabilities and the strategy groups, targeted marketing is conducted to the target users of each of the marketing platforms.
[0075] In another aspect, embodiments of this application also provide a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the precision marketing methods provided by the methods described above, the method comprising: Acquire user data from multiple marketing platforms and construct user profiles based on the user data; the user data includes behavioral data and transaction data; Based on the user profile and the business objectives of each marketing platform, multiple marketing strategies are generated. Each of the marketing strategies is decomposed into atomic capabilities, and the multiple marketing strategies are combined into a strategy group through logical operations; the atomic capabilities include user screening and product recommendation. Based on the atomic capabilities and the strategy groups, targeted marketing is conducted to the target users of each of the marketing platforms.
[0076] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0077] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A method of precision marketing, characterized by, The method comprises: acquiring user data of a plurality of marketing platforms and constructing a user portrait according to the user data; the user data comprises behavior data and transaction data; generating a plurality of marketing strategies based on the user portrait and business objectives of each of the marketing platforms; decomposing each of the marketing strategies into atomic capabilities and combining the plurality of marketing strategies into a strategy group through logical operation; the atomic capabilities comprise user screening and product recommendation; performing precision marketing on target users of each of the marketing platforms according to the atomic capabilities and the strategy group.
2. The method of precision marketing of claim 1, wherein, The precision marketing on the target users of each of the marketing platforms according to the atomic capabilities and the strategy group comprises: encapsulating the atomic capabilities and the strategy group into independent interface services respectively; the interface services are used to provide unified calling interfaces and parameter specifications; configuring a calling process of the interface services according to business demands of each of the marketing platforms; the calling process comprises a dependency relationship and a calling sequence between the interface services; the interface services are called by each of the marketing platforms according to the calling process to obtain an initial recommendation result for the target users of the marketing platform; filtering and screening the initial recommendation result based on a preset recommendation algorithm to perform precision marketing on the target users.
3. The method of precision marketing of claim 2, wherein, The recommendation algorithm comprises a filtering algorithm; the filtering and screening of the initial recommendation result based on the preset recommendation algorithm to perform precision marketing on the target users comprises: calculating user similarity between users of each of the marketing platforms and product similarity between products of the marketing platform based on the preset recommendation algorithm; constructing a neighbor list of each of the users according to the user similarity and constructing a similar product list of each product according to the product similarity; acquiring a scored product of a target user of the marketing platform and a target neighbor list of the target user; filtering and screening the initial recommendation result according to a first product list and a second product list to obtain a to-be-recommended list; the first product list is a similar product list of the scored product, and the second product list is a subscribed product list of each of the users in the target neighbor list; performing precision marketing on the target user based on the to-be-recommended list.
4. The method of precision marketing of claim 2, wherein, The recommendation algorithm comprises a matrix decomposition algorithm; the filtering and screening of the initial recommendation result based on the preset recommendation algorithm to perform precision marketing on the target users comprises: generating a score matrix between users and products of each of the marketing platforms based on the preset recommendation algorithm; decomposing the score matrix into a user feature matrix and a business feature matrix; filtering and screening the initial recommendation result according to the user feature matrix and the business feature matrix to obtain a to-be-recommended list; performing precision marketing on the target user based on the to-be-recommended list.
5. The method of precision marketing of claim 2, wherein, The recommendation algorithm comprises a content recommendation algorithm; the filtering and screening of the initial recommendation result based on the preset recommendation algorithm to perform precision marketing on the target users comprises: Based on the preset recommendation algorithm, the product description text information of each marketing platform is encoded to obtain a product feature vector, and the historical behavior data of the users of each marketing platform is encoded to obtain a user preference vector; The similarity between the user preference vector of the target user and the product feature vector is calculated, and a list of interested products of the target user is determined according to the similarity; The initial recommendation result is filtered and screened according to the list of interested products to obtain a to-be-recommended list; Based on the to-be-recommended list, the target user is accurately marketed.
6. The method of precision marketing of claim 2, wherein, In combination of the atomic ability and the strategy, after accurately marketing the target user of each marketing platform, the method further comprises: Monitoring feedback information of the interface service; the feedback information includes execution state and execution result of the interface service; According to the feedback information, the interface service is optimized.
7. A precision marketing device, characterized by, It comprises: A user portrait construction module is configured to obtain user data of a plurality of marketing platforms and construct a user portrait based on the user data; The user data includes behavior data and transaction data; A marketing strategy generation module is configured to generate a plurality of marketing strategies based on the user portrait and business objectives of each marketing platform; A marketing strategy processing module is configured to decompose each marketing strategy into atomic abilities and combine the plurality of marketing strategies into a strategy group through logical operation; the atomic ability includes user screening and product recommendation; An accurate marketing execution module is configured to accurately market target users of each marketing platform according to the atomic ability and the strategy group.
8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the accurate marketing method of any one of claims 1 to 6. 9.A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the accurate marketing method of any one of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to realize the accurate marketing method of any one of claims 1 to 6.