An e-commerce marketing method and system based on time sequence promotion
By constructing a standardized temporal dataset and combining it with a temporal convolutional network and attention mechanism, e-commerce marketing strategies are generated and dynamically adjusted, solving the problem of insufficient temporal feature extraction in existing technologies and achieving precise and dynamic e-commerce marketing results.
Patent Information
- Application Number
- CN202610329844.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-18
- Publication Date
- 2026-07-07
AI Technical Summary
Existing e-commerce marketing methods lack in-depth analysis of the temporal characteristics of user behavior, product popularity, and marketing effectiveness, resulting in wasted promotional resources and poor marketing results. Existing systems are unable to capture temporal changes in real time and cannot quickly adjust promotional strategies.
By collecting data from e-commerce platforms, a standardized time-series dataset is constructed. Time-series features of user behavior, product popularity, and marketing effectiveness are extracted. By combining temporal convolutional networks and attention mechanisms, correlations are discovered, an initial promotion strategy is generated, and dynamic adjustments are made through iterative training to form a closed-loop optimization.
It has enabled precise and dynamic e-commerce marketing and promotion, improved marketing accuracy and efficiency, reduced resource waste, and enhanced marketing effectiveness and flexibility.
Smart Images
Figure CN122347448A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of e-commerce technology, and more specifically, to an e-commerce marketing method and system based on time-series promotion. Background Technology
[0002] With the rapid development of the e-commerce industry and increasingly fierce market competition, marketing and promotion have become core means for e-commerce companies to increase sales and accumulate users. Currently, most e-commerce marketing and promotion methods adopt a "one-size-fits-all" approach, that is, using the same promotion strategy for all users and all time periods, which often lacks precise control over the time dimension and in-depth analysis of the temporal characteristics of users and products.
[0003] While some existing e-commerce marketing methods incorporate the concept of time sequence, most simply divide promotion phases based on time nodes, failing to fully integrate the temporal patterns of user behavior, the temporal changes in product popularity, and the temporal feedback of marketing effectiveness for dynamic optimization. This leads to wasted promotional resources and poor marketing results. For example, some methods only conduct concentrated promotions on fixed holidays, without considering the temporal differences in user behavior at different lifecycle stages, nor adjusting promotional strategies based on changes in product popularity at different lifecycle stages. This results in a mismatch between promotional content and user needs, and inaccurate timing of promotions.
[0004] Meanwhile, existing e-commerce marketing systems often lack robust mechanisms for extracting temporal features and closed-loop optimization capabilities. This makes it difficult to capture real-time temporal changes during the promotion process and to quickly adjust promotional strategies, further impacting the accuracy and efficiency of marketing campaigns. With the deep penetration of AI technology into the e-commerce field, how to utilize temporal analysis techniques and deep learning models to uncover the temporal patterns of users, products, and marketing, and achieve precise and dynamic temporal promotion, has become a pressing technical challenge in the current e-commerce marketing landscape. Summary of the Invention
[0005] To address the problems existing in the prior art, this disclosure proposes an e-commerce marketing method and system based on time-series promotion to solve at least one of the aforementioned technical problems. The technical solution adopted in this disclosure is as follows: In a first aspect, this disclosure provides an e-commerce marketing method based on time-series promotion, the method comprising: Step S100: Collect user behavior data, product attribute data, marketing history data, and time series time node data from the e-commerce platform. After preprocessing the collected data, a standardized time series dataset is obtained. Step S200: Based on the standardized time series dataset, extract the time series features of user behavior, product popularity, and marketing effectiveness, and mine the correlation between different time series nodes. Construct a time series feature matrix using the time series features of user behavior, product popularity, marketing effectiveness, and correlation. Step S300: Based on the time-series feature matrix, combined with the user lifecycle stage, product lifecycle stage and preset promotion goals, determine the core promotion content, target audience and promotion channels for each promotion cycle through the preset time-series prediction model, and generate an initial time-series promotion strategy. Step S400: Perform promotion operations at the corresponding time nodes according to the initial time-series promotion strategy, collect user feedback data and promotion effect data in real time during the promotion process, and input them into the time-series prediction model for iterative training, dynamically adjust the promotion content, channel placement ratio, and time-series node arrangement, etc., to form a closed-loop optimized time-series promotion strategy. Step S500: After the promotion period ends, evaluate the effectiveness of the time-series promotion strategy after closed-loop optimization based on preset evaluation indicators, and generate an evaluation report.
[0006] Preferably, in step S100, the preprocessing includes cleaning, deduplication, completion, and normalization.
[0007] Preferably, in step S200, a temporal convolutional network combined with an attention mechanism is used to extract temporal features of user behavior, temporal features of product popularity, and temporal features of marketing effectiveness.
[0008] Preferably, in step S200, the correlation between different time-series nodes is mined using a depth-first search algorithm; Preferably, in step S300, the user lifecycle stages include the new user period, the active user period, the dormant user period, and the churned user period.
[0009] Preferably, in step S300, the product life cycle stages include the introduction stage, growth stage, maturity stage, and decline stage.
[0010] Preferably, in step S300, different promotional resources and time priorities are allocated for different combinations of user lifecycle stages and product lifecycle stages.
[0011] Preferably, in step S400, the iterative training of the time-series prediction model can adopt an incremental learning approach, updating only the model parameters corresponding to the new data collected in real time, in order to improve the adjustment efficiency.
[0012] Preferably, in step S500, the preset evaluation indicators include user conversion rate, promotion ROI, product repurchase rate, channel placement efficiency, and time sequence node matching degree; the evaluation report includes the promotion effect of each time sequence node, strategy adjustment records, and optimization suggestions.
[0013] Preferably, the method further includes: Step S600: After generating the evaluation report, archive the time-series data, strategy parameters and evaluation results of this promotion to provide data support for the generation of subsequent time-series promotion strategies.
[0014] A second aspect of this disclosure provides an e-commerce marketing system based on time-series promotion, the system comprising: The data acquisition module 100 is used to collect user behavior data, product attribute data, marketing history data and time series time node data from the e-commerce platform. After preprocessing the collected data, a standardized time series dataset is obtained. The time-series feature module 200 is used to extract user behavior time-series features, product popularity time-series features, and marketing effect time-series features based on the standardized time-series dataset, mine the correlation between different time-series nodes, and construct a time-series feature matrix using the user behavior time-series features, product popularity time-series features, marketing effect time-series features, and correlations. The promotion strategy module 300 is used to determine the core promotion content, target audience and promotion channels for each promotion cycle based on the time-series feature matrix, combined with the user life cycle stage, product life cycle stage and preset promotion goals, and generate an initial time-series promotion strategy through a preset time-series prediction model. The strategy execution module 400 is used to execute promotion operations at the corresponding time nodes according to the initial time-series promotion strategy, collect user feedback data and promotion effect data in real time during the promotion process and input them into the time-series prediction model for iterative training, dynamically adjust the promotion content, channel placement ratio, and time-series node arrangement, etc., to form a closed-loop optimized time-series promotion strategy. The evaluation report module 500 is used to evaluate the effectiveness of the time-series promotion strategy after the promotion period ends, based on preset evaluation indicators, and generate an evaluation report.
[0015] Preferably, the system further includes: The archiving support module 600 is used to archive the time-series data, strategy parameters and evaluation results of this promotion after the evaluation report is generated, so as to provide data support for the generation of subsequent time-series promotion strategies.
[0016] In a third aspect, this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the e-commerce marketing method based on time-series promotion as described above.
[0017] In a fourth aspect, this disclosure provides an electronic device including a processor and a memory, the processor being configured to execute a computer program stored in the memory to implement the time-based promotion e-commerce marketing method described above.
[0018] The beneficial effects of this disclosure are as follows: This disclosure provides an e-commerce marketing method and system based on time-series promotion. By preprocessing the collected data to obtain a standardized time-series dataset, and using this dataset to obtain time-series features of user behavior, product popularity, marketing effectiveness, and related relationships, a time-series feature matrix is constructed. An initial time-series promotion strategy is generated based on this matrix, and promotion operations are then executed at the corresponding time-series nodes according to the initial time-series promotion strategy. This achieves a complete time-series feature extraction mechanism and closed-loop optimization capability, enabling real-time capture of time-series changes during the promotion process and rapid adjustment of the promotion strategy, thus improving the accuracy and efficiency of marketing promotion. After the promotion cycle ends, this disclosure can further evaluate the effectiveness of the closed-loop optimized time-series promotion strategy based on preset evaluation indicators, generating an evaluation report. This allows for timely adjustment of the model and parameters based on marketing results, thereby adjusting the time-series promotion strategy. Furthermore, after generating the evaluation report, this disclosure can archive the time-series data, strategy parameters, and evaluation results of this promotion, providing data support for the generation of subsequent time-series promotion strategies and further supporting the realization of precise and dynamic time-series promotion.
[0019] Compared with the prior art, this disclosure has the following beneficial effects: (1). This disclosure achieves accurate extraction of time-series features by deeply mining the temporal characteristics of user behavior, product popularity and marketing effectiveness, combined with temporal convolutional networks and attention mechanisms. At the same time, it provides accurate data support for the generation of promotion strategies by mining the correlation between different time-series nodes through depth-first search algorithm, thus solving the problem of lack of accurate temporal control in existing e-commerce marketing.
[0020] (2). This disclosure combines the user lifecycle stage and the product lifecycle stage to generate promotion strategies that are adapted to different time nodes. At the same time, it constructs a closed-loop optimization mechanism, collects promotion feedback data in real time, iteratively trains the time-series prediction model, and dynamically adjusts the promotion strategy. This achieves precise and dynamic adaptation of the promotion strategy, effectively improves marketing results, and reduces the waste of promotion resources.
[0021] (3) The system structure of this disclosure is complete, and the modules work together to realize the full-process automation of data collection, preprocessing, time series feature extraction, strategy generation, execution adjustment, effect evaluation and archiving, which improves the efficiency of e-commerce marketing. At the same time, the accumulated time series promotion data can be reused, providing strong support for subsequent marketing promotion.
[0022] (4). This disclosure introduces an incremental learning approach, which only updates the model parameters corresponding to the new data collected in real time, without retraining the entire model. This effectively improves the efficiency of strategy adjustment, ensures that the promotion strategy can quickly adapt to changes in time, and further enhances the flexibility and accuracy of marketing promotion. Attached Figure Description
[0023] The accompanying drawings, which form part of this application, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.
[0024] Figure 1 This is a flowchart of an e-commerce marketing method based on time-series promotion, as shown in Embodiment 1 of this disclosure.
[0025] Figure 2 This is an architecture diagram of an e-commerce marketing system based on time-series promotion, as shown in Embodiment 2 of this disclosure. Detailed Implementation
[0026] The present disclosure will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the present application can be combined with each other.
[0027] The following detailed descriptions are exemplary and intended to provide further detailed explanation of this disclosure. Unless otherwise specified, all technical terms used in this disclosure have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this disclosure.
[0028] Example 1: like Figure 1 As shown, Embodiment 1 of this disclosure provides an e-commerce marketing method based on time-series promotion, the method including steps S100 to S600.
[0029] Step S100: Collect user behavior data, product attribute data, marketing history data, and time series time node data from the e-commerce platform. After preprocessing the collected data, a standardized time series dataset is obtained.
[0030] Furthermore, the user behavior data may include timestamps and behavior details of user browsing, favorites, adding to cart, placing orders, and repeat purchases. The time-series time node data may include preset promotion cycle division nodes, user lifecycle nodes, and product lifecycle nodes.
[0031] Furthermore, in step S100, the preprocessing includes cleaning, deduplication, completion, and normalization.
[0032] Furthermore, in step S100, the normalization process adopts the min-max normalization method to map data of different dimensions to the [0,1] interval, so as to eliminate the influence of dimensions.
[0033] Furthermore, in step S100, the promotion cycle division nodes may include a preheating period, a peak period, a continued sales period, and a closing period, wherein the duration of each cycle is dynamically adjusted according to the product attributes and marketing objectives.
[0034] Specifically, during data collection, a multi-source data fusion approach can be adopted, covering internal data from the e-commerce platform and external related data to ensure data comprehensiveness. During data cleaning, outliers (such as order amounts exceeding reasonable ranges, abnormal timestamps), missing values, and duplicate data are removed. Missing data is filled using interpolation. Normalization employs a min-max normalization method, mapping data from different dimensions (such as user browsing time, product sales, and promotion costs) to the [0,1] interval to eliminate the influence of unit dimensions, facilitating subsequent temporal feature extraction and model training. Promotion cycle nodes are dynamically adjusted based on product attributes (such as FMCG and durable goods) and marketing goals (such as sales growth and brand exposure), typically divided into a pre-sale period, a peak period, a continuation period, and a closing period. For example, the peak period for FMCG can be set to 3-7 days, and for durable goods, it can be set to 7-15 days. User lifecycle nodes are divided based on user behavior temporal characteristics, while product lifecycle nodes are divided based on changes in product popularity over time. Step S200: Based on the standardized time series dataset, extract the time series features of user behavior, product popularity, and marketing effectiveness, and mine the correlation between different time series nodes. Construct a time series feature matrix using the time series features of user behavior, product popularity, marketing effectiveness, and correlation.
[0035] Furthermore, in step S200, a temporal convolutional network combined with an attention mechanism is used to extract temporal features of user behavior, temporal features of product popularity, and temporal features of marketing effectiveness.
[0036] Furthermore, in step S200, the correlation between different time-series nodes is mined using a depth-first search algorithm.
[0037] Furthermore, in step S200, extracting the temporal features of user behavior may specifically include: The standardized temporal dataset is processed by two inception layers of a temporal convolutional network. One inception layer uses a tanh activation function for feature filtering, while the other uses a sigmoid activation function to control the feature propagation. The outputs of the two inception layers are concatenated and then stacked with convolutional layers to obtain the final temporal feature matrix. It is understood that temporal convolutional networks (TCNs) have excellent temporal feature capture capabilities and can effectively extract feature information at different time scales. This disclosure relies on a temporal convolutional network combined with an attention mechanism to highlight the feature weights of key temporal nodes, thereby improving the accuracy of feature extraction.
[0038] Furthermore, the relationships include time-series relationships of user behavior, time-series relationships of product popularity, and relationships between marketing effectiveness and time-series nodes.
[0039] Step S300: Based on the time-series feature matrix, combined with the user lifecycle stage, product lifecycle stage and preset promotion goals, determine the core promotion content, target audience and promotion channels for each promotion cycle through the preset time-series prediction model, and generate an initial time-series promotion strategy.
[0040] Understandably, the user lifecycle stages can be divided based on the temporal characteristics of user behavior, specifically including the new user stage, active user stage, dormant user stage, and churned user stage. During the new user stage, the focus is on pushing new user benefits and popular products to attract users to complete their first order. During the active user stage, the focus is on pushing personalized products and membership benefits to increase user repurchase rates. During the dormant user stage, the focus is on pushing reactivation benefits and exclusive offers to guide users back to activity. During the churned user stage, the focus is on pushing recall activities and limited-time offers to attempt to win back users.
[0041] A product's lifecycle can be divided into stages based on its popularity and timing, specifically including the introduction, growth, maturity, and decline phases. The introduction phase focuses on brand exposure and product introduction to increase brand awareness. The growth phase focuses on increasing promotional efforts and expanding channels to boost sales. The maturity phase focuses on optimizing promotional efficiency, maintaining user loyalty, and stabilizing sales. The decline phase focuses on clearing inventory, introducing alternatives, and reducing wasted promotional resources.
[0042] Preset promotion goals can include sales growth goals, user growth goals, and promotion ROI goals. The time-series prediction model predicts the effects of different promotion strategies at each time-series node based on the time-series feature matrix. Combined with the preset promotion goals, the optimal initial time-series promotion strategy is selected, and the core promotion content (such as product selling points and promotional activities), target audience (such as users of specific age groups and consumption levels), and promotion channels (such as short video platforms, social platforms, and e-commerce site recommendations) for each promotion cycle are clarified.
[0043] Furthermore, the time-series prediction model can be constructed by stacking convolutional layers, and the temporal dimension can be optimized by combining tanh activation function filtering and sigmoid activation function weighting.
[0044] Furthermore, in step S300, the user lifecycle stages include the new user period, the active user period, the dormant user period, and the churned user period.
[0045] Furthermore, in step S300, the product life cycle stages include the introduction stage, growth stage, maturity stage, and decline stage.
[0046] Furthermore, in step S300, different promotional resources and time priorities are allocated for different combinations of user lifecycle stages and product lifecycle stages. For example, a combination of new user stage and product growth stage is allocated more promotional resources and prioritized for time-series promotion, focusing on user acquisition and product exposure; a combination of churned user stage and product decline stage is allocated fewer promotional resources and only a small amount of recall promotion is carried out.
[0047] Step S400: Perform promotion operations at the corresponding time nodes according to the initial time-series promotion strategy, collect user feedback data and promotion effect data in real time during the promotion process, and input them into the time-series prediction model for iterative training, dynamically adjust the promotion content, channel placement ratio, and time-series node arrangement, etc., to form a closed-loop optimized time-series promotion strategy.
[0048] Furthermore, promotional operations can be executed sequentially according to time nodes. For example, the pre-sale period focuses on content preparation and user acquisition; the peak period focuses on concentrated advertising and promotional offers; the follow-up period focuses on user maintenance and repeat purchase guidance; and the closing period focuses on inventory clearance and post-mortem analysis. Real-time collected user feedback data can include information such as user clicks, favorites, comments, and complaints about promotional content, while promotional performance data can include information such as user conversion rate, promotion ROI, product sales, and channel deployment efficiency.
[0049] Further, in step S400, the dynamic adjustment may specifically include: When the promotion effect data is lower than the preset threshold, the promotion channel placement ratio for the corresponding time node is adjusted, prioritizing channels with good user feedback data. Simultaneously optimize promotional content to ensure that it matches user demand and product popularity at the current time point.
[0050] Furthermore, in step S400, the iterative training of the time-series prediction model can adopt an incremental learning approach, updating only the model parameters corresponding to the newly collected data in real time, thereby improving adjustment efficiency. It is understood that by using incremental learning for iterative training, the time-series prediction model can update only the model parameters corresponding to the newly collected data in real time, without retraining the entire model, thus effectively improving adjustment efficiency. When the promotion effect data is lower than a preset threshold, the strategy execution and real-time adjustment module automatically adjusts the promotion channel placement ratio for the corresponding time-series node, prioritizing channels with good user feedback data; simultaneously, it optimizes the promotion content to match user demand and product popularity at the current time-series node, ensuring the dynamic adaptability of the promotion strategy.
[0051] Step S500: After the promotion period ends, evaluate the effectiveness of the time-series promotion strategy after closed-loop optimization based on preset evaluation indicators, and generate an evaluation report.
[0052] Understandably, pre-set evaluation metrics may include user conversion rate, promotion ROI, product repurchase rate, channel deployment efficiency, and time-series node matching. Each metric is weighted and summed to calculate a comprehensive evaluation score, clearly identifying the strengths and weaknesses of the current promotion strategy. The evaluation report may include the promotion effects at each time point, strategy adjustment records, comprehensive evaluation score, and optimization suggestions, providing a reference for optimizing future promotion strategies.
[0053] Step S600: After generating the evaluation report, archive the time-series data, strategy parameters and evaluation results of this promotion to provide data support for the generation of subsequent time-series promotion strategies.
[0054] Example 2: like Figure 2 As shown, Embodiment 2 of this disclosure provides an e-commerce marketing system based on time-series promotion, the system comprising: The data acquisition module 100 is used to collect user behavior data, product attribute data, marketing history data and time series time node data from the e-commerce platform. After preprocessing the collected data, a standardized time series dataset is obtained. The time-series feature module 200 is used to extract user behavior time-series features, product popularity time-series features, and marketing effect time-series features based on the standardized time-series dataset, mine the correlation between different time-series nodes, and construct a time-series feature matrix using the user behavior time-series features, product popularity time-series features, marketing effect time-series features, and correlations. The promotion strategy module 300 is used to determine the core promotion content, target audience and promotion channels for each promotion cycle based on the time-series feature matrix, combined with the user life cycle stage, product life cycle stage and preset promotion goals, and generate an initial time-series promotion strategy through a preset time-series prediction model. The strategy execution module 400 is used to execute promotion operations at the corresponding time nodes according to the initial time-series promotion strategy, collect user feedback data and promotion effect data in real time during the promotion process and input them into the time-series prediction model for iterative training, dynamically adjust the promotion content, channel placement ratio, and time-series node arrangement, etc., to form a closed-loop optimized time-series promotion strategy. The evaluation report module 500 is used to evaluate the effectiveness of the time-series promotion strategy after the promotion period ends, based on preset evaluation indicators, and generate an evaluation report. The archiving support module 600 is used to archive the time-series data, strategy parameters and evaluation results of this promotion after the evaluation report is generated, so as to provide data support for the generation of subsequent time-series promotion strategies.
[0055] Furthermore, the timing feature module 200 may include: The temporal convolutional unit includes two parallel Inception layers, which combine the tanh activation function and the sigmoid activation function respectively to process the standardized temporal dataset, thereby achieving feature filtering and feature transfer control. The association mining unit uses a depth-first search algorithm to mine the association relationships between various time-series features and time-series nodes, and outputs a time-series feature matrix.
[0056] Furthermore, the promotion strategy module 300 may also include: The lifecycle segmentation unit is used to divide the user lifecycle into stages based on the time sequence characteristics of user behavior and the product lifecycle into stages based on the time sequence characteristics of product popularity, and to allocate promotional resources and time sequence priorities to different stage combinations.
[0057] Furthermore, the policy execution module 400 may include: The real-time data collection unit is used to collect user feedback data and promotion effect data during the promotion process; The model iteration unit adopts an incremental learning approach, using real-time collected data to iteratively train the time series prediction model; The strategy adjustment unit is used to adjust the promotion content, channel allocation ratio, and timing node arrangement based on the model iteration results.
[0058] The data acquisition module 100, time-series feature module 200, promotion strategy module 300, strategy execution module 400, evaluation report module 500, and archiving support module 600 described in Example 2 correspond to steps S100, S200, S300, S400, S500, and S600, respectively. It is worth noting that the system described in Example 2 is merely one system implementation of the method described in Example 1, and does not imply that the method described in Example 1 must depend on the system described in Example 2.
[0059] Example 3: Embodiment 3 of this disclosure provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the e-commerce marketing method based on time-series promotion as described in Embodiment 1.
[0060] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program modules, or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.
[0061] Example 4: Embodiment 4 of this disclosure provides an electronic device, which includes a processor and a memory, wherein the processor is used to execute a computer program stored in the memory to implement the e-commerce marketing method based on time-series promotion described in Embodiment 1.
[0062] Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, and read-only memory (ROM).
[0063] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0064] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0065] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0066] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0067] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0068] In summary, the e-commerce marketing method and system based on time-series promotion provided in embodiments 1-4 of this disclosure obtains a standardized time-series dataset after preprocessing the collected data. Based on the standardized time-series dataset, user behavior time-series features, product popularity time-series features, marketing effect time-series features, and correlations are obtained to construct a time-series feature matrix. An initial time-series promotion strategy is generated based on the time-series feature matrix, and then promotion operations are executed at the corresponding time-series nodes according to the initial time-series promotion strategy. This achieves a complete time-series feature extraction mechanism and closed-loop optimization capability, enabling real-time capture of time-series changes during the promotion process and rapid adjustment of the promotion strategy, thereby improving the accuracy and efficiency of marketing promotion. After the promotion cycle ends, this disclosure can further evaluate the effect of the closed-loop optimized time-series promotion strategy based on preset evaluation indicators and generate an evaluation report. This allows for timely adjustment of the model and parameters based on marketing results, thereby adjusting the time-series promotion strategy. After generating the evaluation report, this disclosure can also archive the time-series data, strategy parameters, and evaluation results of this promotion, thereby providing data support for the generation of subsequent time-series promotion strategies and further supporting the realization of precise and dynamic time-series promotion.
[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this disclosure and not to limit them. Although this disclosure has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of this disclosure. Any modifications or equivalent substitutions that do not depart from the spirit and scope of this disclosure should be covered within the protection scope of the claims of this disclosure.
Claims
1. An e-commerce marketing method based on time-series promotion, characterized in that, The method includes: Step S100: Collect user behavior data, product attribute data, marketing history data, and time series time node data from the e-commerce platform. After preprocessing the collected data, a standardized time series dataset is obtained. Step S200: Based on the standardized time series dataset, extract the time series features of user behavior, product popularity, and marketing effectiveness, and mine the correlation between different time series nodes. Construct a time series feature matrix using the time series features of user behavior, product popularity, marketing effectiveness, and correlation. Step S300: Based on the time-series feature matrix, combined with the user lifecycle stage, product lifecycle stage and preset promotion goals, determine the core promotion content, target audience and promotion channels for each promotion cycle through the preset time-series prediction model, and generate an initial time-series promotion strategy. Step S400: Perform promotion operations at the corresponding time nodes according to the initial time-series promotion strategy, collect user feedback data and promotion effect data in real time during the promotion process, and input them into the time-series prediction model for iterative training, dynamically adjust the promotion content, channel placement ratio, and time-series node arrangement to form a closed-loop optimized time-series promotion strategy; Step S500: After the promotion period ends, evaluate the effectiveness of the time-series promotion strategy after closed-loop optimization based on preset evaluation indicators, and generate an evaluation report.
2. The e-commerce marketing method based on time-series promotion according to claim 1, characterized in that, In step S100, the preprocessing includes cleaning, deduplication, completion, and normalization.
3. The e-commerce marketing method based on time-series promotion according to claim 1, characterized in that, In step S200 Temporal convolutional networks combined with attention mechanisms are used to extract temporal features of user behavior, product popularity, and marketing effectiveness. The correlation between different time-series nodes was discovered using a depth-first search algorithm; The relationships include time-series correlations of user behavior, time-series correlations of product popularity, and correlations between marketing effectiveness and time-series nodes.
4. The e-commerce marketing method based on time-series promotion according to claim 1, characterized in that, In step S300, the user lifecycle stages include the new user period, the active user period, the dormant user period, and the churned user period. The product life cycle stages include the introduction stage, growth stage, maturity stage, and decline stage; Different promotional resources and time priorities are allocated for different combinations of user lifecycle stages and product lifecycle stages.
5. The e-commerce marketing method based on time-series promotion according to claim 1, characterized in that, In step S400, the dynamic adjustment specifically includes: When the promotion effect data is lower than the preset threshold, the promotion channel placement ratio for the corresponding time node is adjusted, prioritizing channels with good user feedback data. Simultaneously optimize promotional content to ensure that it matches user demand and product popularity at the current time point.
6. The e-commerce marketing method based on time-series promotion according to claim 1, characterized in that, In step S500, the preset evaluation indicators include user conversion rate, promotion ROI, product repurchase rate, channel placement efficiency, and time sequence node matching degree; the evaluation report includes the promotion effect of each time sequence node, strategy adjustment records, and optimization suggestions.
7. The e-commerce marketing method based on time-series promotion according to claim 1, characterized in that, The method further includes: Step S600: After generating the evaluation report, archive the time-series data, strategy parameters and evaluation results of this promotion to provide data support for the generation of subsequent time-series promotion strategies.
8. An e-commerce marketing system based on time-series promotion, characterized in that, The system includes: The data acquisition module (100) is used to collect user behavior data, product attribute data, marketing history data and time series time node data of e-commerce platforms. After preprocessing the collected data, a standardized time series dataset is obtained. The time-series feature module (200) is used to extract user behavior time-series features, product popularity time-series features and marketing effect time-series features based on the standardized time-series dataset, mine the correlation between different time-series nodes, and construct a time-series feature matrix based on the user behavior time-series features, product popularity time-series features, marketing effect time-series features and correlations. The promotion strategy module (300) is used to determine the core promotion content, target audience and promotion channels for each promotion cycle based on the time-series feature matrix, combined with the user life cycle stage, product life cycle stage and preset promotion goals, and generate an initial time-series promotion strategy through a preset time-series prediction model. The strategy execution module (400) is used to perform promotion operations at the corresponding time nodes according to the initial time-series promotion strategy, collect user feedback data and promotion effect data in real time during the promotion process and input them into the time-series prediction model for iterative training, dynamically adjust the promotion content, channel placement ratio and time-series node arrangement to form a closed-loop optimized time-series promotion strategy. The evaluation report module (500) is used to evaluate the effectiveness of the time-series promotion strategy after the promotion period ends, based on preset evaluation indicators, and generate an evaluation report.
9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the e-commerce marketing method based on time-series promotion as described in any one of claims 1-7.
10. An electronic device comprising a processor and a memory, characterized in that, The processor is used to execute computer programs stored in the memory to implement the e-commerce marketing method based on time-series promotion as described in any one of claims 1-7.