Product marketing service platform system and method thereof

Abstract

The invention relates to the technical field of e-commerce marketing, in particular to a product marketing service platform system and method, and the system comprises a user behavior tracking module, a consumption intention analysis module, a promotion time optimization module, an inventory monitoring and discount adjustment module, and a precision marketing push module. According to the method, the commodity access frequency and cross browsing data are extracted and analyzed in real time, the consumption trend of the user in a short period is rapidly identified, consumption tendency parameters are generated, potential purchase behaviors are accurately predicted by using browsing behaviors and an interaction mode, the recommendation accuracy is improved, the real-time inventory data are combined, a discount strategy and inventory management are dynamically adjusted, and the recommendation efficiency is improved. According to the method, the promotion activity is ensured to be closely matched with the inventory state, the resource utilization efficiency is improved, precise marketing pushing is performed according to the user equipment preference, the user interaction experience is enhanced, the sales growth is promoted, and the market response speed and the marketing effect are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of e-commerce marketing technology, and in particular to a product marketing service platform system and method thereof. Background Art

[0002] The field of e-commerce marketing technology includes Internet-based product promotion, sales strategy formulation, consumer behavior analysis, personalized recommendations and online transaction management. The core content of this technology field includes product information release, market data collection, user interaction analysis, advertising optimization and intelligent recommendation algorithms. E-commerce marketing systems usually rely on large-scale data processing capabilities to form precise marketing plans based on factors such as user historical behavior, geographic location, and social network data. This technology field covers application scenarios such as online retail, brand promotion, social e-commerce, search engine optimization, and advertising bidding ranking, and widely uses data mining, natural language processing, and distributed computing technologies to achieve precise marketing.

[0003] Among them, the product marketing service platform system refers to a service platform based on the e-commerce environment that integrates product display, user behavior analysis, sales data statistics and marketing strategy formulation. The system focuses on technical matters such as information release, customer portrait construction, personalized recommendation and promotion management in product marketing. Through data capture, user feature extraction, label classification, transaction record analysis and other methods, it summarizes user consumption preferences and generates matching marketing plans based on different consumer groups. The system relies on rule matching, text mining, product association analysis and other means to establish a user interest model, and combines automated pricing calculation and advertising distribution strategies to achieve refined product promotion.

[0004] Existing technologies have a certain lag in user consumption behavior analysis and marketing strategy adjustment. User behavior analysis mostly relies on historical data and lacks sensitive capture of short-term consumer interests, resulting in a mismatch between recommended products and users' current needs, reducing the effectiveness of recommendations. There is a lack of dynamic linkage between inventory management and promotion strategies, and inventory adjustments rely more on static sales data, failing to respond to market demand fluctuations in a timely manner, which can easily lead to inventory backlogs or out-of-stocks. In terms of marketing push, existing technologies do not fully consider the differences in user behavior on different devices, resulting in a decrease in the accuracy and reach of push information. In addition, the analysis dimensions of user interaction methods and behavior patterns are limited, making it difficult to fully reflect users' potential purchasing intentions, affecting the conversion effect of marketing activities. These deficiencies limit the platform's ability to respond quickly to market changes and weaken the flexibility and pertinence of marketing activities. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a product marketing service platform system and method thereof.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: A product marketing service platform system comprises: The user behavior tracking module obtains marketing activity touchpoint data, extracts device identification, page dwell time, click frequency, calculates device switching intervals, filters continuous access operation sequences, and generates marketing behavior tracking parameters; The consumption intention analysis module extracts the product visit frequency, the number of cross-browsed products, calculates the category visit weight, and generates short-term consumption tendency parameters based on the marketing behavior tracking parameters; The promotion time optimization module extracts browsing frequency, stay time, and interaction mode based on the short-term consumption tendency parameters, calculates the increase in the number of visits, determines whether the visit frequency threshold is exceeded, screens eligible products, calculates the purchase behavior offset value, and generates a discount time window matching value; The inventory monitoring and discount adjustment module extracts the marketing activity inventory data based on the discount time window matching value, calculates the inventory sales trend rate, determines the fluctuation range of the inventory saleable days, calculates the sales volume change rate ratio, and generates the inventory adjustment result; Based on the inventory adjustment results, the precision marketing push module extracts the user's main marketing interaction terminals, screens high-weight devices, calculates the user's promotion preference factors, and generates precision marketing push results.

[0007] As a further solution of the present invention, the marketing behavior tracking parameters include terminal device category, user residence time ratio, and access path continuity; the short-term consumption tendency parameters include changes in product browsing frequency, category selection preference, and cross-category correlation; the discount time window matching value includes optimal discount duration, access frequency weight, and purchase behavior fluctuation range; the inventory adjustment results include inventory sales rate, remaining inventory assessment, and replenishment demand level; the precision marketing push results include promotion method preference index, device-side interaction weight, and personalized recommendation matching degree.

[0008] As a further solution of the present invention, the user behavior tracking module includes a device data extraction submodule, a behavior feature calculation submodule, a continuous operation screening submodule, and a marketing parameter generation submodule; The device data extraction submodule obtains marketing campaign touchpoint data, extracts device identification, calculates page dwell time based on the timestamp information of the touchpoint data, obtains click frequency by counting the number of clicks of users on each page, matches and integrates page dwell time and click frequency in combination with device identification to generate a device behavior data set; The behavior feature calculation submodule calls the device behavior data set, calculates the switching time interval between the same device and the differentiated pages according to the device identifier, calculates the average stay time of the device according to the page stay time and the click frequency, iteratively determines the click density through the ratio of the click frequency to the number of page visits, and obtains the device behavior feature parameters; The continuous operation screening submodule screens the operation sequence whose device switching time interval is lower than the set threshold according to the device behavior characteristic parameters, and screens the continuous access records within the range of the operation time interval and the behavior characteristic change by judging the change of the residence time and click density of the continuous page access, and generates the continuous access operation sequence data; The marketing parameter generation submodule extracts the key behavioral features of the device in continuous access based on the continuous access operation sequence data, calculates the behavioral conversion rate and interaction frequency of the device in the marketing activities, determines the residence time distribution through the ratio of the page residence time to the access path length, and generates marketing behavior tracking parameters.

[0009] As a further solution of the present invention, the consumer intention analysis module includes a visit frequency extraction submodule, a cross-browsing analysis submodule, a category weight calculation submodule, and a consumer tendency generation submodule; The access frequency extraction submodule obtains the marketing behavior tracking parameters, screens the user's access records on each product page, counts the number of visits to each product by device ID, determines the frequency distribution based on the access time sorting, classifies the user access data to generate the product access frequency, and integrates the product access frequency to form the product access frequency data; The cross-browsing analysis submodule calls the commodity access frequency data, screens the differentiated commodities visited by the user in the same session, counts the number of commodities cross-browsed by the user by commodity category, determines the diversity of the browsed commodities by comparing the differences of commodity categories, calculates the cross-browsing ratio in combination with the user's browsing path length, and obtains the user's cross-browsing quantity data; The category weight calculation submodule calculates the proportion of differentiated product categories in the total number of visits based on the user cross-browsing quantity data and the product visit frequency data, determines the weight by calculating the ratio of the category visit frequency to the total number of visits, adjusts the weight distribution according to the frequency difference of multiple categories in user visits, and generates category visit weight parameters; The consumption tendency generation submodule screens the user's recent visit frequency to high-weight categories according to the category access weight parameters, calculates the cross-browsing ratio value of high-weight categories, analyzes the consumption concentration, and extracts short-term consumption characteristics in combination with the time interval and path sequence of the user's browsing behavior to obtain short-term consumption tendency parameters.

[0010] As a further solution of the present invention, the promotion time optimization module includes a browsing behavior extraction submodule, a visit increase calculation submodule, a commodity screening submodule, and a discount time matching submodule; The browsing behavior extraction submodule obtains the short-term consumption tendency parameters, extracts the browsing frequency of users on the product page, counts the page dwell time, records the interaction methods of differentiated devices, integrates the data relationship between browsing frequency and dwell time, classifies the diversity of interaction methods, and generates browsing behavior feature data; The access increase calculation submodule calls the browsing behavior feature data, counts the number of visits to the product page in time series, calculates the increase in the number of visits in a time period, compares the access frequency with a preset access frequency threshold, determines whether the product visits exceed the threshold, and obtains the access increase data; The product screening submodule screens products with a visit increase exceeding a threshold value based on the visit number increase data, calculates the purchase behavior deviation value of the product in combination with the stay time and interaction mode in the browsing behavior feature data, and generates the purchase behavior deviation value by analyzing the difference between the visit behavior and the purchase behavior; The discount time matching submodule analyzes the optimal promotion time of the product according to the purchase behavior offset value, analyzes the change trend of the purchase offset value to match the discount time window, and combines the combined data of the visit increase and the purchase offset to generate a discount time window matching value.

[0011] As a further solution of the present invention, the purchase behavior offset value calculation formula is specifically: ; in, Represents the purchase behavior deviation value of the product. Represents the length of time the user stays on the page during the i-th visit, Represents the average length of time that the product page stays during all visits. represents the number of interactions of the user in the i-th visit, represents the average number of interactions of the product page in all visits, and n represents the total number of visits to the product page. Represents the sum of the product of the duration of each visit and the number of interactions. represents the sum of squares of the deviations of the duration of stay in all visits from its mean value, Represents the sum of squared deviations of the number of interactions from its mean across all visits.

[0012] As a further solution of the present invention, the inventory monitoring and discount adjustment module includes an inventory data extraction submodule, a sales trend calculation submodule, an inventory fluctuation judgment submodule, and an inventory adjustment generation submodule; The inventory data extraction submodule obtains the discount time window matching value, extracts the inventory data of each commodity in the current marketing activity, sorts the inventory quantity by commodity category, records the inventory changes at each time node, integrates the corresponding relationship between the inventory data and the promotion time, and generates a marketing inventory data set; The sales trend calculation submodule calls the marketing inventory data set, counts the sales volume of the products in the differentiated time periods, calculates the sales volume change rate over time, and obtains the inventory sales trend rate by comparing the sales data in the time periods before and after the promotion; The inventory fluctuation judgment submodule calculates the inventory saleable days of each commodity based on the inventory sales trend rate, analyzes the fluctuation range of the inventory saleable days within the differentiated promotion period, judges whether the inventory fluctuation exceeds the set threshold, and generates the inventory saleable days fluctuation value; The inventory adjustment generation submodule calculates the sales volume change rate value according to the inventory saleable days fluctuation value, analyzes the inventory consumption rate ratio, screens the commodities for which the discount strategy needs to be adjusted, and generates the inventory adjustment result by combining the inventory change rate and the sales trend rate.

[0013] As a further solution of the present invention, the change rate value calculation formula is specifically: ; in, Represents the sales volume change rate value, represents the actual sales volume in the kth time period, Represents the average sales volume for all time periods. a time label representing the kth time period, the time period being expressed in hours or days, Represents the average value of time labels in all time periods, Represents the time difference between adjacent time periods. Represents the sum of the absolute values ​​of the sales volume differences in consecutive time periods, represents the sum of squares of sales deviations from the mean, Represents the sum of squares of time labels' deviations from the mean.

[0014] As a further solution of the present invention, the precision marketing push module includes a user terminal extraction submodule, a high-weight device screening submodule, a promotion preference calculation submodule, and a push result generation submodule; The user terminal extraction submodule obtains the inventory adjustment results, extracts the key interactive terminals of the user in each marketing activity, counts the usage frequency of each device, records the user's interactive behavior on differentiated devices, combines the device usage time and frequency, classifies and integrates the data by terminal type, and generates user interactive terminal data; The high-weight device screening submodule calls the user interaction terminal data, calculates the device weight according to the device usage frequency and interaction intensity, screens the devices with device weight higher than the set threshold, and screens the most active devices based on the time distribution of device usage and interaction frequency to generate a high-weight device list; The promotion preference calculation submodule analyzes the user's interactive behavior in the promotion activities based on the high-weight device list, calculates the user's click-through rate and access frequency for differentiated promotion activities, and calculates the user's promotion preference coefficient based on the device interaction data and the promotion activity response intensity to obtain the user's promotion preference factor; The push result generation submodule matches the corresponding promotion content according to the user promotion preference factor, combines the user's device usage habits and promotion response characteristics, generates personalized push content, integrates the push data of each device, and generates marketing push results.

[0015] A product marketing service platform method, comprising the following steps: S1: Obtain the marketing activity touch point data, extract the device identification, page dwell time, click frequency, calculate the device switching interval, filter the continuous access operation sequence, and generate marketing behavior tracking parameters; S2: Based on the marketing behavior tracking parameters, extract the product access frequency, calculate the category access weight, and generate short-term consumption tendency parameters; S3: Based on the short-term consumption tendency parameters, the browsing frequency, the stay time, and the interaction mode are extracted, the increase in the number of visits is calculated, qualified commodities are screened, the purchase behavior deviation value is calculated, and the discount time window matching value is generated; S4: Based on the discount time window matching value, extract the marketing activity inventory data, calculate the inventory sales trend rate, calculate the sales volume change rate ratio, and generate the inventory adjustment result; S5: Based on the inventory adjustment results, extract the user's main marketing interaction terminals, filter high-weight devices, and generate precision marketing push results.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by extracting and analyzing the frequency of commodity visits and cross-browsing data in real time, the user's short-term consumption trend is quickly identified, consumption tendency parameters are generated, and browsing behavior and interaction methods are used to accurately predict potential purchase behavior, thereby improving the accuracy of recommendations. In combination with real-time inventory data, discount strategies and inventory management are dynamically adjusted to ensure that promotional activities are closely coordinated with inventory status, improve resource utilization efficiency, and perform precise marketing push based on user device preferences, thereby enhancing user interactive experience, promoting sales growth, and significantly improving market response speed and marketing effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 is a system flow chart of the present invention; Figure 2 It is a submodule flow chart of the present invention; Figure 3 This is a flow chart of the user behavior tracking module of the present invention; Figure 4 This is a flow chart of the consumption intention analysis module of the present invention; Figure 5 It is a flow chart of the promotion time optimization module of the present invention; Figure 6 It is a flow chart of the inventory monitoring and discount adjustment module of the present invention; Figure 7 This is a flow chart of the precision marketing push module of the present invention; Figure 8 The figure is a flow chart of the steps of the method of the present invention. DETAILED DESCRIPTION

[0019] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0020] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0021] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0022] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0023] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0024] See also Figure 1 , a product marketing service platform system includes: The user behavior tracking module obtains marketing activity touchpoint data, extracts device identification, page dwell time, click frequency, calculates device switching intervals, filters continuous access operation sequences, and generates marketing behavior tracking parameters; The consumption intention analysis module extracts the frequency of product visits and the number of cross-browsed products based on marketing behavior tracking parameters, calculates the category visit weight, and generates short-term consumption tendency parameters; The promotion time optimization module extracts browsing frequency, stay time, and interaction mode based on short-term consumption tendency parameters, calculates the increase in the number of visits, determines whether the visit frequency threshold is exceeded, screens eligible products, calculates the purchase behavior offset value, and generates the discount time window matching value; The inventory monitoring and discount adjustment module extracts marketing activity inventory data based on the discount time window matching value, calculates the inventory sales trend rate, determines the fluctuation range of the inventory saleable days, calculates the sales volume change rate ratio, and generates inventory adjustment results; The precision marketing push module extracts the user's main marketing interaction terminals based on inventory adjustment results, screens high-weight devices, calculates the user's promotion preference factors, and generates precision marketing push results.

[0025] Marketing behavior tracking parameters include terminal device category, user residence time ratio, and access path continuity. Short-term consumption tendency parameters include changes in product browsing frequency, category selection preference, and cross-category correlation. Discount time window matching values ​​include optimal discount duration, access frequency weight, and purchase behavior fluctuation range. Inventory adjustment results include inventory sales rate, remaining inventory assessment, and replenishment demand level. Precision marketing push results include promotion method preference index, device-side interaction weight, and personalized recommendation matching degree.

[0026] See also Figure 3 and Figure 2 ,The user behavior tracking module includes a device data extraction submodule, a behavior feature calculation submodule, a continuous operation screening submodule, and a marketing parameter generation submodule; The device data extraction submodule obtains marketing campaign touchpoint data, extracts device identification, calculates page dwell time based on the timestamp information of the touchpoint data, obtains click frequency by counting the number of clicks of users on each page, matches and integrates page dwell time and click frequency in combination with device identification to generate a device behavior data set; First, it is necessary to extract the unique identifier of the user device from each touch record in the marketing campaign, such as the device ID or MAC address. Then, based on the timestamp information in the touch data, calculate the user's stay time on a specific page. The specific steps are as follows: first, extract the entry and exit time of each touch, and calculate the difference between the two to obtain the stay time. For example, user A's entry time on page X is 10:00:00 and the exit time is 10:05:00, then the page stay time is 300 seconds. Then, the click frequency is obtained by counting the number of clicks of the user on each page. Assuming that user A clicks the button or link on the page 5 times within the above 5 minutes, the click frequency is 5 times / 300 seconds. Then, the page stay time and click frequency are matched and integrated in combination with the device identifier to form a device behavior record table. For example, a device with a device ID of 12345 stays on page X for 300 seconds and clicks 5 times, and stays on page Y for 200 seconds and clicks 2 times. After integrating these data, a complete device behavior data set is generated.

[0027] The behavior feature calculation submodule calls the device behavior data set, calculates the switching time interval between the same device and the differentiated pages according to the device identifier, calculates the average stay time of the device according to the page stay time and click frequency, and iteratively determines the click density through the ratio of click frequency to page visit number to obtain the device behavior feature parameters; First, filter all page access records of the same device by device ID, and then calculate the switching time interval between the differentiated pages of the device. The specific operation is: extract the exit and entry time of adjacent page visits, and calculate the time difference between the two. For example, the user with device ID 12345 leaves page X at 10:05:00 and enters page Y at 10:06:00, then the page switching time interval is 60 seconds. Then, calculate the average stay time of the device based on the page stay time and click frequency. The formula is: average stay time = (total page stay time) Sum) / (number of visited pages). Assuming that a user with device ID 12345 visited 3 pages, and stayed for 300 seconds, 200 seconds, and 400 seconds respectively, the average stay time is (300+200+400) / 3=300 seconds. The click density is further determined by the ratio of click frequency to the number of page visits. Click density = total number of clicks / number of visited pages. If the total number of clicks is 12 times, then click density = 12 / 3=4 times / page. Finally, the device behavior characteristic parameters are obtained, including page switching time interval, average stay time, click density, etc.

[0028] The continuous operation screening submodule screens the operation sequences whose device switching time interval is lower than the set threshold according to the device behavior characteristic parameters, and screens the continuous access records within the range of the operation time interval and behavior characteristic changes by judging the changes in the residence time and click density of continuous page visits, and generates continuous access operation sequence data; First, a page switching time interval threshold is set, such as 60 seconds, and page access records with a switching time of less than 60 seconds are screened out. If the device ID is 12345 and the switching time between page X and page Y is 45 seconds, this access record is screened out. Then, the changes in the dwell time and click density of consecutive page accesses are judged, and the dwell time difference and click density difference of consecutive pages are calculated. If the difference is within the set range of variation, such as the dwell time change does not exceed ±20 seconds, and the click density change does not exceed ±1 time / page, it is identified as a continuous access operation, and continuous access operation sequence data is generated. Assume that the device ID is 12345 and continuously accesses pages X, Y, and Z, with dwell times of 300 seconds, 280 seconds, and 290 seconds, respectively, and click densities of 4 times / page, 3 times / page, and 4 times / page, respectively. Since the range of variation meets the set conditions, continuous access operation sequence data is finally generated.

[0029] The marketing parameter generation submodule extracts the key behavioral features of the device in continuous access based on the continuous access operation sequence data, calculates the behavioral conversion rate and interaction frequency of the device in the marketing activities, determines the residence time distribution through the ratio of the page residence time to the access path length, and generates marketing behavior tracking parameters; First, identify the key behaviors in the continuous access sequence, such as clicking the purchase button, filling out the form, etc., record the frequency and time points of these key behaviors, and then calculate the behavior conversion rate of the device in the marketing activity. The formula is: Conversion rate = (number of key behaviors) / (total number of visits). For example, if the device ID is 12345 and there are 2 key behaviors in continuous visits and the total number of visits is 5, then the conversion rate = 2 / 5 = 40%. At the same time, calculate the interaction frequency of the device, that is, the number of clicks per unit time. The interaction frequency = total number of clicks / total dwell time. Assuming the total number of clicks is 15 times and the total dwell time is 900 seconds, the interaction frequency = 15 / 900 = 0.0167 times / second. Finally, the dwell time distribution is determined by the ratio of the page dwell time to the access path length. The formula is: Dwell time distribution = total dwell time / number of visited pages. If the total dwell time is 900 seconds and the access path length is 5 pages, then the dwell time distribution = 900 / 5 = 180 seconds / page, and finally generate the marketing behavior tracking parameters.

[0030] See also Figure 4 and Figure 2,The consumption intention analysis module includes the access frequency extraction submodule, the ,cross-browsing analysis submodule, the category weight calculation submodule, and the consumption ,property generation submodule; The access frequency extraction submodule obtains marketing behavior tracking parameters, filters the user's access records on each product page, counts the number of visits to each product by device ID, determines the frequency distribution based on the access time, classifies the user access data to generate the product access frequency, and integrates the product access frequency to form the product access frequency data; First, filter the user's access records on each product page. The specific steps are as follows: the system extracts each user's access records from the server log or front-end embedded data. The extracted data fields include device ID, page ID, access timestamp and other information. For example, a user with device ID ABC123 visited product pages P001 and P002 on January 1, 2024. The corresponding timestamps are 10:00 and 10:05. Then, the number of visits to each product is counted by device ID. The system groups the access records with the same device ID and counts the visit frequency of each product page. Assume that device ABC123 visited product P001 three times and P002 twice in one day. times, then the number of visits to P001 is 3, and the number of visits to P002 is 2. These access records are then sorted based on the access time to determine the time distribution of the access frequency. The system arranges the access timestamps in ascending order and calculates the time interval between each visit. For example, the three visits to P001 are at 10:00, 12:00, and 14:00, with an interval of 2 hours and 2 hours respectively. The regularity of user visits is determined, and the user access data is classified to generate the product access frequency. The number of visits corresponding to each device identifier is summarized to form an access frequency table for each product. Finally, the product access frequencies of all users are integrated to form overall product access frequency data for subsequent analysis.

[0031] The cross-browsing analysis submodule calls the product access frequency data, screens the differentiated products visited by users in the same session, counts the number of products cross-browsed by users by product category, determines the diversity of browsed products by comparing the differences in product categories, calculates the cross-browsing ratio based on the user's browsing path length, and obtains the user's cross-browsing quantity data; First, filter the differentiated products that users visit in the same session. The system defines a time window for the same session (for example, if there is no operation within 30 minutes, the session is considered to be over). Extract all access records within this time range. For example, if device ABC123 visits products P001, P002, and P003 between 10:00 and 10:25, it is considered as one session. Then count the number of cross-browsed products by product category. The system classifies the visited products according to the product category labels. For example, P001 and P002 belong to electronic products, and P003 belongs to household products. Count the number of visits to products of different categories. In this example, the cross-browsed product category is 2. Then, the diversity of browsed products is determined by comparing the differences in product categories. The system calculates the proportion of different categories visited by users to the total categories. For example, if the total number of categories is 5 and the user visits 2 categories, the diversity ratio is 2 / 5=40%. The cross-browsing ratio is calculated based on the length of the user's browsing path. The path length refers to the total number of products visited by the user in a session. The cross-browsing ratio = (number of products in different categories) / (total number of visited products). Assuming that the user visits 5 products in a session, 3 of which belong to different categories, the cross-browsing ratio = 3 / 5=60%. Finally, the user's cross-browsing quantity data is obtained, which reflects the user's diversified browsing behavior.

[0032] The category weight calculation submodule is based on the user cross-browsing data and the product access frequency data, and counts the proportion of differentiated product categories in the total access volume. The weight is determined by calculating the ratio of category access frequency to total access volume, and the weight distribution is adjusted according to the frequency difference of multiple categories in user access to generate category access weight parameters. First, the proportion of differentiated product categories in the total number of visits is counted. The system classifies all users' visit records by category and calculates the proportion of visits to each category in the total number of visits. For example, the total number of visits is 1,000, the number of visits to electronic products is 400, and the number of visits to household products is 200. Then the proportion of electronic products = 400 / 1,000 = 40%, and the proportion of household products = 200 / 1,000 = 20%. Then the weight is determined by calculating the ratio of category visit frequency to total visits. Category weight = (category visit frequency) / (total visit frequency). As mentioned above, the weight of electronic products is 40%, and that of household products is 20%. Subsequently, the weight distribution is adjusted based on the frequency differences of multiple categories in user visits. The system adjusts categories with significant differences in visit frequency, defining "significant difference" as a difference in visit frequency exceeding 10% of the total visits. For example, the difference between electronic products and household products is 20%, which exceeds the threshold. The weight of electronic products is adjusted to 35%, and that of household products is increased to 25%. Finally, the category visit weight parameters are generated for subsequent consumption tendency analysis.

[0033] The consumption tendency generation submodule screens the user's recent visit frequency to high-weight categories based on the category access weight parameters, calculates the cross-browsing ratio value of high-weight categories, analyzes the consumption concentration, and extracts short-term consumption characteristics in combination with the time interval and path sequence of user browsing behavior to obtain short-term consumption tendency parameters; First, filter the user's recent visit frequency in high-weight categories. The system defines "high weight" as a category weight greater than 30%. Filter the user's visit records in these categories. For example, the weight of electronic products is 35%. Filter the records of device ABC123 visiting the electronic product page in the past 7 days, count the visit frequency, and then calculate the cross-browsing ratio value of high-weight categories. The system extracts the number of different products visited by users in high-weight categories and calculates the cross-browsing ratio = (number of different products) / (total number of visited products). Assuming that the user visited 5 electronic products, 3 of which are of different models, the cross-browsing ratio = 3 / 5 = 60%. Then analyze the consumption concentration, which = 1-cross-browsing ratio, reflecting whether the user is concentrated on certain specific products. In this example, the concentration is 40%. Combined with the time interval and path sequence of the user's browsing behavior, extract short-term consumption characteristics. The system calculates the time interval between continuous visits. If the interval is less than 10 minutes, it is regarded as a continuous behavior. At the same time, the order and frequency of the access path are recorded, and finally the short-term consumption tendency parameters are obtained, reflecting the user's recent consumption intention in high-weight categories.

[0034] See also Figure 5 and Figure 2 ,The promotion time optimization module includes a browsing behavior extraction submodule, a visit increase calculation submodule, a product screening submodule, and a discount time matching submodule; The browsing behavior extraction submodule obtains short-term consumption tendency parameters, extracts the browsing frequency of users on the product page, counts the page dwell time, records the interaction methods of differentiated devices, integrates the data relationship between browsing frequency and dwell time, classifies the diversity of interaction methods, and generates browsing behavior feature data; First, extract the browsing frequency of users on the product page from the database or user behavior tracking system, including the number of visits to each product page by each user. For example, if the database shows that user A visits product page X 7 times in a week, then the module counts the page stay time for each visit, calculated by the entry and exit timestamps in the log file. For example, the average stay time of user A on product page X is calculated as: , if user A stays for 2 minutes, 3 minutes and 4 minutes in three visits respectively, then the average stay time is (2+3+4) / 3=3 minutes. Next, record the interaction methods of differentiated devices, including clicks, slides and other behaviors, and classify and count these interaction behaviors. The system will analyze the user's interaction habits according to the type and frequency of interaction behaviors. For example, if user A clicks 10 times and slides 5 times during browsing, the data relationship between browsing frequency and stay time is integrated, and normalization is used to convert the data into a unified standard. For example, the diversity of interaction methods can be classified. The calculation method may involve a diversity index, such as: , if user A has 3 different interaction methods and a total of 20 interactions, the diversity index is 3 / 20=0.15, and the browsing behavior feature data is finally generated. This data integrates the user's attention to the product and the depth of interaction, providing a basis for the next step of consumption tendency analysis.

[0035] The visit increase calculation submodule calls the browsing behavior feature data, counts the number of visits to the product page in time series, calculates the increase in the number of visits in the time period, compares the visit frequency with the preset visit frequency threshold, determines whether the product visits exceed the threshold, and obtains the visit increase data; First, this module classifies the data by time series and counts the number of visits to the product page in each time period. For example, the system counts by hour and finds that the number of visits to product page X is 20 times between 9 and 10 o'clock and 30 times between 10 and 11 o'clock. The increase in the number of visits in each time period is calculated using the following formula: , in this example, the increase from 9-10 to 10-11 is Then, the module compares the calculation result with the preset access frequency threshold. For example, if the threshold is set to 50%, it determines whether the number of visits to the product in each time period exceeds this threshold to identify abnormal or significant visit growth, and finally obtains the visit frequency increase data, which shows the changing trend of the visit frequency in each time period and provides a basis for further marketing decisions.

[0036] The product screening submodule screens products with visit increase exceeding the threshold based on the visit number increase data, calculates the purchase behavior deviation value of the products based on the stay time and interaction mode in the browsing behavior feature data, and generates the purchase behavior deviation value by analyzing the difference between the visit behavior and the purchase behavior; The specific calculation formula for the purchase behavior offset value is: ; in, Represents the purchase behavior deviation value of the product. Represents the length of time the user stays on the page during the i-th visit, Represents the average length of time that the product page stays during all visits. represents the number of interactions of the user in the i-th visit, represents the average number of interactions of the product page in all visits, and n represents the total number of visits to the product page. Represents the sum of the product of the duration of each visit and the number of interactions. represents the sum of squares of the deviations of the duration of stay in all visits from its mean value, Represents the sum of squared deviations of the number of interactions from its mean across all visits.

[0037] This formula is used to calculate the purchase behavior deviation value of a product. Indicates the degree to which a user's interaction behavior on a specific product page differs from normal purchasing behavior.

[0038] Acquisition and quantification standards of various parameters: The length of time a user stays on a page during the i-th visit, in seconds. This data is recorded through page embedding technology to record the user's entry and exit times, and the length of stay is the exit time minus the entry time.

[0039] The average duration of all visits to the product page, calculated by dividing the sum of the duration of all visits by the number of visits.

[0040] The number of interactions of the user in the i-th visit. Interaction behaviors include clicks, swipes, and add-to-cart. The number of clicks and swipes is recorded by the front-end tracking point, and add-to-cart and purchase behaviors are recorded by the transaction system.

[0041] Is the average number of interactions across all visits, calculated as the total number of interactions divided by the number of visits.

[0042] n is the total number of visits to the product page, which is calculated from the system log.

[0043] Specific numerical setting and derivation process: Assume that within the 7 days of monitoring, the page of product A is visited 5 times, and the user's stay time and number of interactions in each visit are as follows: First visit: Duration of stay = 120 seconds, number of interactions =5 Second time: Duration of stay =180 seconds, number of interactions =7 Third time: Duration of stay = 240 seconds, number of interactions =6 4th time: Duration of stay = 150 seconds, number of interactions =4 Fifth time: Duration of stay = 210 seconds, number of interactions =8 Calculate average length of stay and average number of interactions : ; ; calculate : ; ; ; ; ; ; calculate : ; Compute the sum of squares in the denominator: Calculate the final denominator: ; Substitute the formula to calculate the purchase behavior offset value : ; The results show that the purchase behavior offset value of product A is 47.38, indicating that there is a significant difference between the user's interaction behavior on the product page and the actual purchase behavior. The higher offset value means that although users have more interactions and stay, the proportion of final conversion to purchase is low, which provides a basis for subsequent analysis.

[0044] The discount time matching submodule analyzes the optimal promotion time of the product according to the purchase behavior offset value, analyzes the change trend of the purchase offset value to match the discount time window, and combines the combined data of the visit increase and the purchase offset to generate the discount time window matching value; First, analyze the optimal promotion time for the product. This analysis is based on the changing trend of the product's purchase behavior offset value in different time periods. For example, if the offset value of product X from 12 noon to 1 pm is significantly higher than other time periods, then this time period may be the optimal promotion time. Then analyze the changing trend of the purchase offset value to match the discount time window. Use the following formula to calculate the discount time window matching value: For example, if the visit increase of product X is 50% between 12:00 and 13:00, and the purchase behavior offset is 60%, then the discount time window matching value is ,This value is higher than the set benchmark value (such as 50%), indicating that this time period is the best time to promote discounts. Finally, the module generates a discount time window matching value to guide the actual implementation of discount strategies and optimize the timeliness and effectiveness of promotional activities.

[0045] See also Figure 6 and Figure 2 ,The inventory monitoring and discount adjustment module includes an inventory data extraction submodule, a sales trend calculation submodule, an inventory fluctuation judgment submodule, and an inventory adjustment generation submodule; The inventory data extraction submodule obtains the discount time window matching value, extracts the inventory data of each product in the current marketing activity, sorts the inventory quantity by product category, records the inventory changes at each time node, integrates the corresponding relationship between inventory data and promotion time, and generates a marketing inventory data set; First, the inventory data of each product in the current marketing campaign is extracted. The system reads the real-time inventory information of all products participating in the promotion from the inventory management database. The data fields include product ID, product name, current inventory quantity, etc. For example, the inventory quantity of product A is 500 pieces, and that of product B is 300 pieces. Then, the inventory quantity is sorted by product category. The system classifies and summarizes the inventory data according to the category labels of the products (such as electronic products, clothing, household items, etc.). For example, the total inventory of products under the electronic products category is 1,500 pieces, and that of the clothing category is 1,200 pieces. Then, the inventory changes at each time node are recorded, and the system sets a scheduled task (such as every hour). Record inventory data once), analyze inventory consumption by comparing inventory quantities at different time nodes. For example, the inventory of product A at 9 a.m. is 500 pieces, and it becomes 480 pieces at 10 a.m., indicating that 20 pieces were sold within this hour. Next, the corresponding relationship between inventory data and promotion time is integrated. The system matches inventory changes with the time nodes of promotion activities, such as inventory changes before and after the promotion starts. By analyzing the relationship between inventory consumption rate and promotion time, a marketing inventory data set is generated. This data set records in detail the changes in the inventory quantity of goods at different time nodes and during promotion activities, providing a basis for subsequent sales trend analysis.

[0046] The sales trend calculation submodule calls the marketing inventory data set, counts the sales volume of products in differentiated time periods, calculates the sales volume change rate over time, and obtains the inventory sales trend rate by comparing the sales data before and after the promotion. The specific calculation formula for the change rate value is: ; in, Represents the sales volume change rate value, represents the actual sales volume in the kth time period, Represents the average sales volume for all time periods. a time label representing the kth time period, the time period being expressed in hours or days, Represents the average value of time labels in all time periods, Represents the time difference between adjacent time periods. Represents the sum of the absolute values ​​of the sales volume differences in consecutive time periods, represents the sum of squares of sales deviations from the mean, Represents the sum of squares of time labels' deviations from the mean.

[0047] This formula is used to calculate the sales volume change rate value. , which represents the relationship between sales volume fluctuation and time changes in different time periods, and helps measure the actual impact of promotional activities on sales volume.

[0048] Parameter acquisition and quantification process: : The actual sales volume in the kth time period. The data is obtained through the commodity sales records in the sales system. The sales volume data in hours is extracted from the retail system.

[0049] : The average of the sales, calculated as the sum of the sales in all time periods divided by the number of time periods.

[0050] : The time label of the kth time period, marked by hours, = 1 hour, = 2 hours, and so on.

[0051] : The average value of the time label, calculated as the sum of the time periods divided by the number of time periods.

[0052] : The time difference between adjacent time periods, in hours. In the case of equal intervals, this value is 1 hour.

[0053] p: represents the total number of time periods, determined from the time range of the monitored promotional activities.

[0054] Actual data collection and examples: The promotion is carried out within 8 hours. The system monitors the hourly sales of product A and records the following data: Hour 1: =50 pieces Hour 2: =70 pieces Hour 3: =65 pieces Hour 4: =80 pieces Hour 5: =90 pieces Hour 6: =85 pieces Hour 7: =75 pieces Hour 8: =60 pieces Calculate the average sales volume : ; Calculate the average of the time labels : ; Calculate the sum of the absolute values ​​of the sales volume differences between adjacent time periods :

[0055] Calculate the sum of time differences : Since monitoring is done every hour, the time difference is 1 hour: =7 hours; Calculate the sum of squared deviations in sales volume :

[0056] Calculate the sum of squares of time deviations :

[0057]

[0058] Calculate the denominator: ; Substitute the formula to calculate the sales volume change rate : ; Explanation of calculation results: The results show that the sales volume change rate of product A during the promotion period is 15.63, which indicates the volatility of sales volume during the promotion period and the combined strength of the time span. The higher value indicates that the promotion activity has caused significant fluctuations in sales volume, providing data support for subsequent inventory adjustments and promotion strategy optimization.

[0059] The inventory fluctuation judgment submodule calculates the inventory saleable days of each product based on the inventory sales trend rate, analyzes the fluctuation range of the inventory saleable days within the differentiated promotion period, determines whether the inventory fluctuation exceeds the set threshold, and generates the inventory saleable days fluctuation value; First, calculate the number of days that each product can be sold. The number of days that the inventory can be sold = current inventory / average daily sales. For example, the current inventory of product A is 500 pieces, and the average daily sales is 50 pieces. Then the number of days that the inventory can be sold = 500 / 50 = 10 days. Then analyze the fluctuation range of the number of days that the inventory can be sold during the differentiated promotion period. The system records the number of days that the inventory can be sold in each time period before and after the promotion. By calculating the fluctuation range = (maximum value - minimum value) / average value × 100%, for example, the number of days that the product A can be sold before the promotion is 10 days, and after the promotion it drops to 5 days, then the fluctuation range is (10-5) / 7.5 × 100% = 66.67%. Then determine whether the inventory fluctuation exceeds the set threshold. The system sets the inventory fluctuation threshold to 50%. If the fluctuation range exceeds this threshold, it is determined that the inventory fluctuation is significant. Finally, the fluctuation value of the number of days that the inventory can be sold is generated. This value is used to monitor inventory stability and provide a basis for inventory management and promotion adjustments.

[0060] The inventory adjustment generation submodule calculates the sales volume change rate value based on the inventory saleable days fluctuation value, analyzes the inventory consumption rate ratio, selects the products that need to adjust the discount strategy, and generates the inventory adjustment results by combining the inventory change rate and sales trend rate; First, calculate the sales volume change rate value, sales volume change rate = (sales volume during promotion - sales volume before promotion) / sales volume before promotion × 100%. For example, the average daily sales volume of product A before promotion is 50 pieces, and the average daily sales volume during promotion is 100 pieces. Then the sales volume change rate is (100-50) / 50 × 100% = 100%. Then analyze the ratio of inventory consumption rate, inventory consumption rate = sales volume change rate / inventory fluctuation value. For example, if the inventory fluctuation value is 66.67%, then the inventory consumption rate ratio is 100% / 66.67% ≈ 1.5. Then filter out the products that need to adjust the discount strategy. The system sets the inventory consumption rate ratio threshold (such as 1.2). If the ratio exceeds this threshold, it indicates that the inventory is consumed too quickly and the discount strategy needs to be adjusted. Finally, combined with the inventory change rate and sales trend rate, the system comprehensively analyzes the inventory and sales data to generate inventory adjustment results. For example, reduce the discount intensity for products that are consumed too quickly, increase promotions for products that are consumed slowly, optimize inventory turnover, and ensure inventory and sales balance.

[0061] See also Figure 7 and Figure 2 ,The precision marketing push module includes a user terminal extraction submodule, a high-weight device screening submodule, a promotion preference calculation submodule, and a push result generation submodule; The user terminal extraction submodule obtains inventory adjustment results, extracts the key interactive terminals of users in each marketing activity, counts the usage frequency of each device, records the user's interactive behavior on differentiated devices, and integrates the data by terminal type based on the device usage time and frequency to generate user interactive terminal data; First, the key interactive terminals of users in each marketing activity are extracted. The system extracts the device identification, device type (such as mobile phone, tablet, PC, etc.) and its corresponding interaction data from the user interaction log. For example, user A uses a mobile phone (device ID: M123) and a PC (device ID: P456) to browse and purchase in a promotion. Then the usage frequency of each device is counted. The system records the number of visits to the device during the activity. Assuming that user A visits 30 times with a mobile phone and 10 times with a PC, the usage frequency of the mobile phone is 3 times that of the PC. Then the user's interaction behavior on the differentiated devices is recorded, including specific operations such as clicking, browsing, adding to cart, and placing orders. The system The interactive behaviors of different devices are counted by category. For example, user A made 20 clicks and 5 add-to-cart orders on the mobile phone, and 5 clicks and 3 add-to-cart orders on the PC. Combined with the device usage time and frequency, the system calculates the device's active time period and usage intensity. For example, user A uses his mobile phone most frequently from 18:00 to 20:00 every day. Usage frequency = number of visits / length of time period. If he visits 20 times within 2 hours, the usage frequency is 10 times / hour. The system integrates all data by terminal type and ultimately generates user interaction terminal data. The data includes device type, usage frequency, interactive behavior and active time period, providing a basis for subsequent device screening.

[0062] The high-weight device screening submodule calls the user interaction terminal data, calculates the device weight according to the device usage frequency and interaction intensity, screens the devices with device weight higher than the set threshold, and screens the most active devices based on the device usage time distribution and interaction frequency to generate a high-weight device list; First, calculate the device weight based on the device usage frequency and interaction intensity. The device weight calculation formula is: Device weight = (usage frequency × interaction intensity coefficient). The interaction intensity coefficient is determined by the number of specific operations on the device, for example, click = 1, add to cart = 2, purchase = 3. Assuming that user A uses his mobile phone 30 times, the interaction intensity (20 clicks, 5 add to cart) = 20 × 1 + 5 × 2 = 30, then the device weight of the mobile phone = 30 × 30 = 900. The PC is used 10 times, the interaction intensity (5 clicks, 3 add to cart) = 5 × 1 + 3 × 2 = 11, then the device weight of the PC = 1 0×11=110, then filter out devices with weights higher than the set threshold. The system sets the device weight threshold to 500, and filters out devices with weights greater than 500. User A’s mobile phone weight is 900, which exceeds the threshold and is filtered out. The PC weight is 110 and is not filtered out. Then, combine the time distribution of device usage and interaction frequency to analyze the activity of the device. Activity = usage frequency / total usage time of the device. If the total usage time of the mobile phone is 10 hours and the activity is 3 times / hour, the devices with the highest activity are finally filtered out, and a list of high-weight devices is generated to provide input data for subsequent promotion preference analysis.

[0063] The promotion preference calculation submodule analyzes the user's interactive behavior in promotion activities based on the high-weight device list, calculates the user's click-through rate and access frequency for differentiated promotion activities, and calculates the user's promotion preference coefficient based on the device interaction data and the promotion activity response intensity to obtain the user's promotion preference factor; First, the user's interactive behavior in the promotion activities is analyzed. The system extracts the user's click-through rate and visit frequency during the promotion period. Click-through rate = number of clicks / number of visits. Assuming that user A visited 50 times and clicked 40 times when participating in the promotion using a mobile phone, the click-through rate = 40 / 50 = 80%. Then, the user's response to differentiated promotions is calculated. The system classifies the user's click-through rate and visit frequency in different promotions. For example, the click-through rate of activity A is 80%, and the click-through rate of activity B is 50%. Then, based on the device interaction data and the response intensity of the promotion activities, the user's promotion preference coefficient is calculated. Promotion preference coefficient = (click-through rate × interaction intensity) / device usage frequency. For example, user A's promotion preference coefficient in activity A = (80% × 30) / 50 = 0.48, and the promotion preference coefficient in activity B = (50% × 30) / 50 = 0.3. Finally, the user's promotion preference factor is obtained, which is used to identify the user's interest and participation in different promotions, providing a basis for personalized push.

[0064] The push result generation submodule matches the corresponding promotion content according to the user's promotion preference factor, combines the user's device usage habits and promotion response characteristics, generates personalized push content, integrates the push data of each device, and generates marketing push results; First, the corresponding promotional content is matched. The system compares the user's promotion preference factor with the available promotional activities and selects the activity with the highest user preference factor to match. For example, user A's preference factor for activity A is 0.48 and activity B is 0.3. The system prioritizes pushing activity A. Then, based on the user's device usage habits and promotion response characteristics, the system analyzes the active time periods and interaction patterns of the user's most commonly used devices. For example, user A uses his mobile phone most frequently between 18:00 and 20:00, and the push time is set to this time period. Personalized push content is generated, including promotional products, discount information and personalized recommendations. The system integrates these contents into the push template, and then integrates the push data of each device to ensure that the push content of different devices is consistent in form and time. Finally, a marketing push result is generated, which includes detailed information such as the pushed device ID, push content, and push time, which are used to execute personalized promotion push.

[0065] See also Figure 8 , a product marketing service platform method, comprising the following steps: S1: Obtain marketing activity touchpoint data, extract device identification, page dwell time, click frequency, calculate device switching interval, filter continuous access operation sequence, and generate marketing behavior tracking parameters; S2: Based on the marketing behavior tracking parameters, extract the product access frequency, calculate the category access weight, and generate short-term consumption tendency parameters; S3: Based on the short-term consumption tendency parameters, it extracts browsing frequency, stay time, and interaction mode, calculates the increase in the number of visits, screens eligible products, calculates the purchase behavior offset value, and generates the discount time window matching value; S4: Based on the discount time window matching value, extract the marketing activity inventory data, calculate the inventory sales trend rate, calculate the sales volume change rate ratio, and generate the inventory adjustment result; S5: Based on the inventory adjustment results, extract the user's main marketing interaction terminals, filter high-weight devices, and generate precise marketing push results.

[0066] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A product marketing service platform system, characterized by: The system comprises: The user behavior tracking module obtains marketing activity touchpoint data, extracts device identification, page dwell time, click frequency, calculates device switching intervals, filters continuous access operation sequences, and generates marketing behavior tracking parameters; The consumption intention analysis module extracts the product visit frequency, the number of cross-browsed products, calculates the category visit weight, and generates short-term consumption tendency parameters based on the marketing behavior tracking parameters; The promotion time optimization module extracts browsing frequency, stay time, and interaction mode based on the short-term consumption tendency parameters, calculates the increase in the number of visits, determines whether the visit frequency threshold is exceeded, screens eligible products, calculates the purchase behavior offset value, and generates a discount time window matching value; The inventory monitoring and discount adjustment module extracts the marketing activity inventory data based on the discount time window matching value, calculates the inventory sales trend rate, determines the fluctuation range of the inventory saleable days, calculates the sales volume change rate ratio, and generates the inventory adjustment result; Based on the inventory adjustment results, the precision marketing push module extracts the user's main marketing interaction terminals, screens high-weight devices, calculates the user's promotion preference factors, and generates precision marketing push results.

2. The product marketing service platform system according to claim 1, characterized in that: The marketing behavior tracking parameters include terminal device category, user residence time ratio, and access path continuity; the short-term consumption tendency parameters include changes in product browsing frequency, category selection preference, and cross-category correlation; the discount time window matching value includes optimal discount duration, access frequency weight, and purchase behavior fluctuation range; the inventory adjustment results include inventory sales rate, remaining inventory assessment, and replenishment demand level; the precision marketing push results include promotion method preference index, device-side interaction weight, and personalized recommendation matching degree.

3. The product marketing service platform system according to claim 1, characterized in that: The user behavior tracking module includes a device data extraction submodule, a behavior feature calculation submodule, a continuous operation screening submodule, and a marketing parameter generation submodule; The device data extraction submodule obtains marketing campaign touchpoint data, extracts device identification, calculates page dwell time based on the timestamp information of the touchpoint data, obtains click frequency by counting the number of clicks of users on each page, matches and integrates page dwell time and click frequency in combination with device identification to generate a device behavior data set; The behavior feature calculation submodule calls the device behavior data set, calculates the switching time interval between the same device and the differentiated pages according to the device identifier, calculates the average stay time of the device according to the page stay time and the click frequency, iteratively determines the click density through the ratio of the click frequency to the number of page visits, and obtains the device behavior feature parameters; The continuous operation screening submodule screens the operation sequence whose device switching time interval is lower than the set threshold according to the device behavior characteristic parameters, and screens the continuous access records within the range of the operation time interval and the behavior characteristic change by judging the change of the residence time and click density of the continuous page access, and generates the continuous access operation sequence data; The marketing parameter generation submodule extracts the key behavioral features of the device in continuous access based on the continuous access operation sequence data, calculates the behavioral conversion rate and interaction frequency of the device in the marketing activities, determines the residence time distribution through the ratio of the page residence time to the access path length, and generates marketing behavior tracking parameters.

4. The product marketing service platform system according to claim 1, characterized in that: The consumption intention analysis module includes a visit frequency extraction submodule, a cross-browsing analysis submodule, a category weight calculation submodule, and a consumption tendency generation submodule; The access frequency extraction submodule obtains the marketing behavior tracking parameters, screens the user's access records on each product page, counts the number of visits to each product by device ID, determines the frequency distribution based on the access time sorting, classifies the user access data to generate the product access frequency, and integrates the product access frequency to form the product access frequency data; The cross-browsing analysis submodule calls the commodity access frequency data, screens the differentiated commodities visited by the user in the same session, counts the number of commodities cross-browsed by the user by commodity category, determines the diversity of the browsed commodities by comparing the differences of commodity categories, calculates the cross-browsing ratio in combination with the user's browsing path length, and obtains the user's cross-browsing quantity data; The category weight calculation submodule calculates the proportion of differentiated product categories in the total number of visits based on the user cross-browsing quantity data and the product visit frequency data, determines the weight by calculating the ratio of the category visit frequency to the total number of visits, adjusts the weight distribution according to the frequency difference of multiple categories in user visits, and generates category visit weight parameters; The consumption tendency generation submodule screens the user's recent visit frequency to high-weight categories according to the category access weight parameters, calculates the cross-browsing ratio value of high-weight categories, analyzes the consumption concentration, and extracts short-term consumption characteristics in combination with the time interval and path sequence of the user's browsing behavior to obtain short-term consumption tendency parameters.

5. The product marketing service platform system according to claim 1, characterized in that: The promotion time optimization module includes a browsing behavior extraction submodule, a visit increase calculation submodule, a product screening submodule, and a discount time matching submodule; The browsing behavior extraction submodule obtains the short-term consumption tendency parameters, extracts the browsing frequency of users on the product page, counts the page dwell time, records the interaction methods of differentiated devices, integrates the data relationship between browsing frequency and dwell time, classifies the diversity of interaction methods, and generates browsing behavior feature data; The access increase calculation submodule calls the browsing behavior feature data, counts the number of visits to the product page in time series, calculates the increase in the number of visits in a time period, compares the access frequency with a preset access frequency threshold, determines whether the product visits exceed the threshold, and obtains the access increase data; The product screening submodule screens products with a visit increase exceeding a threshold value based on the visit number increase data, calculates the purchase behavior deviation value of the product in combination with the stay time and interaction mode in the browsing behavior feature data, and generates the purchase behavior deviation value by analyzing the difference between the visit behavior and the purchase behavior; The discount time matching submodule analyzes the optimal promotion time of the product according to the purchase behavior offset value, analyzes the change trend of the purchase offset value to match the discount time window, and combines the combined data of the visit increase and the purchase offset to generate a discount time window matching value.

6. The product marketing service platform system according to claim 5, characterized in that: The purchase behavior deviation value calculation formula is specifically as follows: ； in, Represents the purchase behavior deviation value of the product. Represents the length of time the user stays on the page during the i-th visit, Represents the average length of time that the product page stays during all visits. represents the number of interactions of the user in the i-th visit, represents the average number of interactions of the product page in all visits, and n represents the total number of visits to the product page. Represents the sum of the product of the duration of each visit and the number of interactions. represents the sum of squares of the deviations of the duration of stay in all visits from its mean value, Represents the sum of squared deviations of the number of interactions from its mean across all visits.

7. The product marketing service platform system according to claim 1, characterized in that: The inventory monitoring and discount adjustment module includes an inventory data extraction submodule, a sales trend calculation submodule, an inventory fluctuation judgment submodule, and an inventory adjustment generation submodule; The inventory data extraction submodule obtains the discount time window matching value, extracts the inventory data of each commodity in the current marketing activity, sorts the inventory quantity by commodity category, records the inventory changes at each time node, integrates the corresponding relationship between the inventory data and the promotion time, and generates a marketing inventory data set; The sales trend calculation submodule calls the marketing inventory data set, counts the sales volume of the products in the differentiated time periods, calculates the sales volume change rate over time, and obtains the inventory sales trend rate by comparing the sales data in the time periods before and after the promotion; The inventory fluctuation judgment submodule calculates the inventory saleable days of each commodity based on the inventory sales trend rate, analyzes the fluctuation range of the inventory saleable days within the differentiated promotion period, judges whether the inventory fluctuation exceeds the set threshold, and generates the inventory saleable days fluctuation value; The inventory adjustment generation submodule calculates the sales volume change rate value according to the inventory saleable days fluctuation value, analyzes the inventory consumption rate ratio, screens the commodities for which the discount strategy needs to be adjusted, and generates the inventory adjustment result by combining the inventory change rate and the sales trend rate.

8. The product marketing service platform system according to claim 7, characterized in that: The specific calculation formula of the change rate value is: ； in, Represents the sales volume change rate value, represents the actual sales volume in the kth time period, Represents the average sales volume for all time periods. a time label representing the kth time period, the time period being expressed in hours or days, Represents the average value of time labels in all time periods, Represents the time difference between adjacent time periods. Represents the sum of the absolute values ​​of the sales volume differences in consecutive time periods, represents the sum of squares of sales deviations from the mean, Represents the sum of squares of time labels' deviations from the mean.

9. The product marketing service platform system according to claim 1, characterized in that: The precision marketing push module includes a user terminal extraction submodule, a high-weight device screening submodule, a promotion preference calculation submodule, and a push result generation submodule; The user terminal extraction submodule obtains the inventory adjustment results, extracts the key interactive terminals of the user in each marketing activity, counts the usage frequency of each device, records the user's interactive behavior on differentiated devices, combines the device usage time and frequency, classifies and integrates the data by terminal type, and generates user interactive terminal data; The high-weight device screening submodule calls the user interaction terminal data, calculates the device weight according to the device usage frequency and interaction intensity, screens the devices with device weight higher than the set threshold, and screens the most active devices based on the time distribution of device usage and interaction frequency to generate a high-weight device list; The promotion preference calculation submodule analyzes the user's interactive behavior in the promotion activities based on the high-weight device list, calculates the user's click-through rate and access frequency for differentiated promotion activities, and calculates the user's promotion preference coefficient based on the device interaction data and the promotion activity response intensity to obtain the user's promotion preference factor; The push result generation submodule matches the corresponding promotion content according to the user promotion preference factor, combines the user's device usage habits and promotion response characteristics, generates personalized push content, integrates the push data of each device, and generates marketing push results.

10. A product marketing service platform method, characterized in that: The product marketing service platform system according to any one of claims 1 to 9 is implemented, comprising the following steps: S1: Obtain the marketing activity touch point data, extract the device identification, page dwell time, click frequency, calculate the device switching interval, filter the continuous access operation sequence, and generate marketing behavior tracking parameters; S2: Based on the marketing behavior tracking parameters, extract the product access frequency, calculate the category access weight, and generate short-term consumption tendency parameters; S3: Based on the short-term consumption tendency parameters, the browsing frequency, the stay time, and the interaction mode are extracted, the increase in the number of visits is calculated, qualified commodities are screened, the purchase behavior deviation value is calculated, and the discount time window matching value is generated; S4: Based on the discount time window matching value, extract the marketing activity inventory data, calculate the inventory sales trend rate, calculate the sales volume change rate ratio, and generate the inventory adjustment result; S5: Based on the inventory adjustment results, extract the user's main marketing interaction terminals, filter high-weight devices, and generate precision marketing push results.

Images