SAAS platform and method based on mobile digital marketing

Abstract

The present invention relates to the field of intelligent marketing technology, specifically to a SAAS platform based on mobile digital marketing and its method, wherein the platform includes a behavior sequence module, an identity matching module, a conversion willingness evaluation module, a marketing link optimization module, and a reach strategy adjustment module. In the present invention, through the timestamp sorting and behavior pattern analysis of access data, it is possible to accurately identify the access stability between devices, effectively eliminate random interference, ensure the accuracy of identity recognition, classify unpaid behaviors, optimize marketing strategies, screen user interest preferences, match suitable marketing reach strategies in an automated manner, realize personalized push, improve user interaction rate, adjust push content and frequency, make marketing information more targeted, reduce user interference, improve click-through rate and stay time, improve user conversion rate, and optimize marketing links through a closed data loop to form a more efficient marketing strategy execution system.

Description

Technical Field

[0001] The present invention relates to the field of intelligent marketing technology, and in particular to a SAAS platform and method based on mobile digital marketing. Background Art

[0002] The SAAS platform for mobile digital marketing refers to a software solution provided through cloud services, which is specifically used to design, deploy and monitor marketing activities on mobile devices.

[0003] Existing technologies rely on general data analysis and behavior tracking methods, which make it difficult to accurately portray users' cross-device access behaviors, resulting in identity identification being easily interfered with by external factors, affecting the pertinence of marketing strategies. Due to the failure to deeply analyze mouse tracks, dwell time, and product interaction features, users' purchasing intentions are difficult to accurately predict, and invalid push notifications are prone to occur. During the implementation of marketing plans, users are usually reached by fixed rules, and personalized adjustments are ignored, causing some users to feel disgusted by excessive push notifications, affecting the brand image. Marketing feedback is mostly processed through batch analysis, which cannot achieve real-time optimization, and push frequency and content adjustments lag behind, reducing the effectiveness of marketing information. Due to the failure to form a data-driven marketing closed loop, strategy adjustments rely on manual intervention, with a slow response speed, making it difficult to adapt to rapidly changing market demands, resulting in poor user experience and limited conversion effects. Summary of the invention

[0004] In order to solve the problem that it is difficult to accurately portray the cross-device access behavior of users in the prior art, which makes identity identification susceptible to interference from external factors and affects the pertinence of marketing strategies. Due to the failure to deeply analyze the mouse trajectory, dwell time and product interaction characteristics, it is difficult to accurately predict the user's purchasing intention, and invalid push is prone to occur. In the implementation process of the marketing plan, fixed rules are usually used to reach users, and personalized adjustments are ignored, which causes some users to feel disgusted due to excessive pushes, affecting the brand image. The processing method of marketing feedback is mainly based on batch analysis, and real-time optimization cannot be achieved. The push frequency and content adjustment are delayed, which reduces the effectiveness of marketing information. Due to the failure to form a data-driven marketing closed loop, strategy adjustments rely on manual intervention, the response speed is slow, and it is difficult to adapt to rapidly changing market demands, resulting in poor user experience and limited conversion effects. Technical problems, the embodiment of the present invention provides a SAAS platform based on mobile digital marketing and its method. The technical solution is as follows: On the one hand, it provides a SAAS platform based on mobile digital marketing, which includes: The behavior sequence module obtains SAAS user access data, sorts the page access order, click frequency and stay time by timestamp, determines whether the device switching time interval meets the continuous access characteristics, analyzes the device access mode and access frequency, and obtains the stability result of the device behavior sequence; The identity matching module extracts the device interaction features of the SAAS user access data according to the device behavior sequence stabilization result, calculates the device switching interval time, and obtains the identity matching confidence value; The conversion intention evaluation module obtains the mouse hover time and click path in the SAAS user access data based on the identity matching confidence value, determines whether the same product is visited multiple times and added to the shopping cart without payment, and obtains the conversion intention classification result; The marketing link optimization module uses the conversion intention classification results to trigger limited-time discounts and inventory reminders for users with high conversion intentions, and push product reviews and user comments to users with low interest, thereby obtaining an optimized marketing reach path; The reach strategy adjustment module detects the user response in the SAAS push data based on the optimized marketing reach path, calculates the click rate and dwell time, adjusts the push frequency and content, and obtains the marketing reach adjustment result.

[0005] As a further solution of the present invention, the device behavior sequence stabilization results include page access order, page access frequency, and multi-device access coordination; the identity matching confidence value includes device recognition degree, user interaction consistency, and identity authentication strength; the conversion intention classification results include product attention, shopping cart behavior analysis, and purchase potential rating; the optimized marketing reach path includes limited-time discount notifications, inventory shortage reminders, product evaluation displays, and user evaluation push; the marketing reach adjustment results include advertising delivery efficiency, content matching degree, and response time optimization results.

[0006] As a further solution of the present invention, the behavior sequence module includes: The access record screening submodule obtains SAAS user access data, sorts the page access sequence by timestamp, calculates the page click frequency and stay time, screens the continuous access records of the same device, and obtains the page access behavior sequence; The device switching identification submodule calculates the device identification changes of adjacent access records based on the page access behavior sequence, extracts the device switching time interval, determines whether it is lower than the device continuous access threshold, screens the device switching records that meet the continuous access characteristics, calculates the continuous device switching ratio, analyzes the access interval distribution of differentiated device switching, and establishes a device continuous access identification; The access pattern analysis submodule calculates the device access frequency through the device continuous access identification, counts the access time distribution of differentiated devices, screens the device behavior matching the stable characteristics, and generates a stable result of the device behavior sequence.

[0007] As a further solution of the present invention, the identity matching module includes: The interaction frequency extraction submodule utilizes the device behavior sequence stabilization result and extracts the device interaction frequency, access sequence and switching features between devices according to the SAAS user access data to obtain the device interaction feature data; The device switching time calculation submodule calculates the adjacent access time intervals in the device switching records based on the device interaction feature data, counts the time interval distribution of device switching, calculates the concentration of access time between devices, calculates the mean and coefficient of variation of device switching, and generates device switching interval parameters; The matching degree evaluation submodule uses the device switching interval parameter to calculate the matching degree of the device access mode, and obtains the identity matching confidence value by combining the device interaction characteristics and time stability.

[0008] As a further solution of the present invention, the formula for calculating the coefficient of variation of the device switching is: ; Where CV represents the coefficient of variation of device switching, represents the access time interval of the ith device switching, represents the access time interval of the i-1th device switching, represents the mean of the access time interval of device switching, N represents the total number of device switching, Represents the absolute difference between two adjacent device switching time intervals.

[0009] As a further solution of the present invention, the conversion willingness evaluation module includes: The user behavior trajectory extraction submodule uses the identity matching confidence value and SAAS user access data to extract the mouse hover time, click path and access degree, screen the interaction records, and obtain user interaction trajectory data; The interactive feature analysis submodule calculates the number and interval of multiple visits of users to the same product based on the user interaction trajectory data, counts the proportion of actions of adding to the shopping cart but not paying, analyzes the distribution of mouse hovering time, calculates the product page stay rate and shopping path completeness, extracts core interactive features, and obtains key conversion behavior parameters; The willingness classification submodule calls the key conversion behavior parameters, matches the preset classification standards according to the access frequency, interaction degree and payment behavior pattern, and generates a conversion willingness classification result.

[0010] As a further solution of the present invention, the marketing link optimization module includes: The marketing matching submodule screens user groups with differentiated intentions based on the conversion intention classification results, calculates the interaction difference values ​​of the groups to differentiated marketing methods, screens the contact methods with interaction deviations, matches the optimal marketing contact methods of the groups, and obtains the user marketing contact method matching results; The promotion push optimization submodule matches the results of the user marketing contact method, screens users with high conversion intention, calculates the limited-time discount strength and inventory reminder interval, calculates the recommended content weight and promotion push period matched by interested users, calculates the optimal push method for evaluation, review and case of low-interest users, and obtains user promotion push records; The marketing content response monitoring submodule utilizes the user promotion push records to obtain user interaction data after marketing reach, calculates the marketing response deviation value, screens user groups with reach effect deviation, adjusts the push method, and obtains an optimized marketing reach path.

[0011] As a further solution of the present invention, the reach strategy adjustment module includes: The user response monitoring submodule detects the user response in the SAAS push data through the optimized marketing reach path, extracts push click records, stay duration and interaction times, and obtains the user push response results; The reach effect analysis submodule calculates the click-through rate of differentiated push content based on the user push response results, counts the average stay time of users on the push page, analyzes the impact of push frequency on user response, calculates the push effect difference value, and generates marketing reach effect parameters; The push strategy optimization submodule uses the marketing reach effect parameters to adjust the push frequency and content, optimize the push time window and information presentation method, match user interaction characteristics, and generate marketing reach adjustment results.

[0012] As a further solution of the present invention, the push frequency and content are adjusted using the formula: ; in, Represents push optimization parameters, Represents the weight of the content pushed for the oth time, represents the average weight of the pushed content, n represents the total number of pushed content, represents the jth user interaction intensity, represents the j-th user interaction time interval, and m represents the total number of user interactions.

[0013] On the other hand, the SAAS method based on mobile digital marketing is performed based on the SAAS platform based on mobile digital marketing, and includes the following steps: S1: Obtain SAAS user access data, collect page access sequence, click frequency and stay time, calculate device switching interval, analyze access patterns, and obtain stable results of device behavior sequence; S2: Based on the stable result of the device behavior sequence, extract the device interaction features, calculate the device switching interval, count the access frequency of differentiated devices, normalize the interval time, and obtain the identity matching confidence value; S3: According to the identity matching confidence value, the mouse hover time and click path are obtained, users who repeatedly visit the product are screened, the visit repetition rate is calculated, and the conversion intention classification result is obtained; S4: Based on the conversion intention classification results, screen users with differentiated interest levels, match digital marketing reach methods, set inventory reminders and product recommendations for high-interest users, and obtain an optimized marketing reach path; S5: According to the optimized marketing reach path, detect the user response in the push data, calculate the click rate and dwell time, adjust the push frequency and content, and obtain the marketing reach adjustment result.

[0014] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: Through the timestamp sorting and behavior pattern analysis of access data, it is possible to accurately identify the access stability between devices, effectively eliminate random interference, and ensure the accuracy of identity recognition. By extracting device interaction features and calculating the switching interval time, user identity matching is made more refined, and the reliability of data analysis is improved. Combined with user behavior data such as mouse hover time and click path, it is possible to deeply analyze access intentions, accurately identify high-conversion users, and classify non-payment behaviors to optimize marketing strategies. After screening user interest preferences, the appropriate marketing contact strategy is matched in an automated manner to achieve personalized push and increase user interaction rate. With the help of user feedback data, the push content and frequency are adjusted to make marketing information more targeted, reduce user interference, and increase click-through rate and stay time. Optimize the processing method of user behavior data, enhance identity matching accuracy, improve user conversion rate, and optimize the marketing chain through data closed loop to form a more efficient marketing strategy execution system. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. 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 the drawings without paying creative work.

[0016] Figure 1 is a schematic diagram of a SAAS platform based on mobile digital marketing provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the platform framework of the present invention; Figure 3 This is a flow chart for obtaining the behavior sequence module in the present invention; Figure 4 This is a flowchart of obtaining the identity matching module in the present invention; Figure 5 This is a flowchart for obtaining the conversion willingness evaluation module in the present invention; Figure 6 This is a flowchart for obtaining the marketing link optimization module in the present invention; Figure 7 This is a flowchart for obtaining the reach strategy adjustment module in the present invention; Figure 8 It is a flow chart for obtaining a SAAS method based on mobile digital marketing provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0018] 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.

[0019] 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.

[0020] 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.

[0021] 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.

[0022] The embodiment of the present invention provides a SAAS platform based on mobile digital marketing, such as Figure 1-2 The schematic diagram of the SAAS platform based on mobile digital marketing is shown, and the platform includes: The behavior sequence module obtains SAAS user access data, sorts the page access order, click frequency and stay time by timestamp, determines whether the device switching time interval meets the continuous access characteristics, analyzes the device access mode and access frequency, and obtains the stability result of the device behavior sequence; The identity matching module extracts the device interaction features of the SAAS user access data based on the device behavior sequence stabilization results, calculates the behavior fit, determines the access consistency, calculates the interval time of device switching, and obtains the identity matching confidence value; The conversion intention evaluation module obtains the mouse hover time and click path in the SAAS user access data based on the identity matching confidence value, determines whether the same product is visited multiple times and added to the shopping cart without payment, and obtains the conversion intention classification result; The marketing link optimization module uses the conversion intention classification results to match the marketing reach methods of differentiated types of users, triggers promotion push of limited-time discounts and inventory reminders for users with high conversion intentions, optimizes recommended content for interested users and pushes promotion information at specific times, reduces direct promotions for low-interest users, pushes product reviews, user comments and use cases, and monitors users' responsiveness to marketing content to obtain optimized marketing reach paths; The reach strategy adjustment module detects user responses in SAAS push data based on the optimized marketing reach path, calculates click-through rate and dwell time, adjusts push frequency and content, and obtains marketing reach adjustment results; The stable results of device behavior sequence include page access order, page access frequency, and multi-device access coordination. The identity matching confidence value includes device recognition degree, user interaction consistency, and identity authentication strength. The conversion intention classification results include product attention, shopping cart behavior analysis, and purchase potential rating. The optimized marketing reach path includes limited-time discount notifications, inventory shortage reminders, product evaluation displays, and user evaluation push. The marketing reach adjustment results include advertising delivery efficiency, content matching degree, and response time optimization results.

[0023] Specifically, if Figure 2 , 3 As shown, the behavior sequence module includes: The access record screening submodule obtains SAAS user access data, sorts the page access sequence by timestamp, calculates the page click frequency and stay time, screens the continuous access records of the same device, and obtains the page access behavior sequence; Get SAAS user access data, extract all users' page access records for user access data, including user ID, device ID, page ID, access timestamp, session ID and other information, sort the access records of the same user in ascending order according to timestamp to obtain the time sequence of page access. In this process, use the double pointer method to traverse the user's access record array to ensure that the timestamp is sorted in ascending order, calculate the page click frequency, count the number of clicks on a certain page by the same user within a specific time window (such as 1 hour or 1 day), use a dictionary structure to store the mapping relationship between page ID and its access times, and map it to the page ID for each access. The count value of the page ID should be +1, and the residence time is calculated at the same time. The residence time is calculated for each page access record, that is, the difference between the timestamp of the current access record and the timestamp of the next access record. If the current access record is the last one, the default residence time is set to a specific timeout time (such as 30 minutes). The continuous access records of the same device are filtered, the access records are grouped based on the device ID, and each group of access records is traversed. After sorting by timestamp, the time difference of adjacent records is calculated. If the time difference is less than the set user offline threshold (such as 10 minutes), it is determined to be a continuous access, otherwise it is regarded as a new session, forming a page access behavior sequence.

[0024] The device switching identification submodule calculates the device identification changes of adjacent access records based on the page access behavior sequence, extracts the device switching time interval, determines whether it is lower than the device continuous access threshold, screens the device switching records that meet the continuous access characteristics, calculates the continuous device switching ratio, analyzes the access interval distribution of differentiated device switching, and establishes the device continuous access identification; Calculate the device identification change of adjacent access records, extract the device ID of adjacent access records, and determine whether the two are the same. If they are different, record the device switching time interval, that is, the timestamp of the latter access record minus the timestamp of the previous access record. After extracting the device switching time interval, compare it with the device continuous access threshold. The threshold is obtained based on user access pattern statistics. For example, the 95th percentile of all device switching time intervals in the past 30 days is taken as the threshold. If the device switching time interval is less than the threshold, it is determined to be a continuous access device switch. Filter the device switching records that meet the continuous access characteristics and use the filter to traverse all devices. Switching records, retain records that meet the above judgment conditions, calculate the continuous device switching ratio, that is, the ratio of the number of device switching times that meet the continuous access characteristics to the total number of device switching times, analyze the access interval distribution of differentiated device switching, count the time intervals of all device switching, divide the intervals by time intervals, such as [0-1min], [1-5min], [5-15min], [more than 15min], count the number of switching times in each interval, and calculate the proportion of each interval, establish a device continuous access identification, group the access records according to the device ID, and mark whether they belong to a continuous access device, and establish a device continuous access identification.

[0025] The access pattern analysis submodule uses the device continuous access identification to calculate the device access frequency, count the access time distribution of differentiated devices, screen the device behaviors that match the stable features, and generate stable results of the device behavior sequence; Calculate the device access frequency, count the number of visits of each device within a specific time window (such as one day or one week), and store them in a hash table with the key being the device ID and the value being the number of visits within the corresponding time window. Count the access time distribution of differentiated devices, traverse all device access records, calculate the average access time of each device, that is, the sum of the stay time of all access records divided by the number of visits, and divide the access time into intervals, such as [0-30s], [30-120s], [2-10min], and [more than 10min]. Count the proportion of devices in each interval, filter out device behaviors that match stable characteristics, set stable characteristic standards, such as access frequency stability (the standard deviation of the number of daily visits in the past 7 days is less than 1) and access time stability (the coefficient of variation of the access time distribution is less than 0.3), traverse all device data, calculate and determine whether they meet the stable characteristic conditions, and generate stable results for the device behavior sequence.

[0026] Specifically, if Figure 2 , 4 As shown, the identity matching module includes: The interaction frequency extraction submodule uses the device behavior sequence stabilization results to extract the device interaction frequency, access sequence and switching features between devices based on the SAAS user access data to obtain device interaction feature data; According to the user access data of SAAS and based on the device behavior sequence data, all users' access records are traversed to extract the device interaction frequency, that is, the number of visits of each device within a fixed time window (such as 1 hour, 1 day) is calculated and stored in a hash table, where the key is the device ID and the value is the number of visits in the corresponding time window. The access records are sorted by timestamp to obtain the access order, that is, the access records are grouped by device ID, each group of data is traversed, the page access path arranged in chronological order is extracted, and stored in an array. At the same time, the switching features between devices are extracted. For the same user (based on login information or identity matching confidence value), all device access records are traversed to calculate the device switching time interval of adjacent access records, that is, the first access timestamp of the latter device minus the last access timestamp of the previous device, and stored in a list. If the interval is lower than the set threshold (such as 5 minutes), it is determined to be a fast device switch, and the device ID switching sequence is recorded, including the device interaction frequency, access order and device switching features, and stored in the database for subsequent matching analysis to obtain device interaction feature data, The device switching time calculation submodule calculates the adjacent access time intervals in the device switching records based on the device interaction feature data, counts the time interval distribution of device switching, calculates the concentration of access time between devices, calculates the mean and coefficient of variation of device switching, and generates device switching interval parameters; The formula for calculating the coefficient of variation of device switching is: ; Where CV represents the coefficient of variation of device switching, represents the access time interval of the ith device switching, represents the access time interval of the i-1th device switching, represents the mean of the access time interval of device switching, N represents the total number of device switching, Represents the absolute difference between two adjacent device switching time intervals; The formula is used to calculate the coefficient of variation CV of the equipment switching time interval and evaluate the degree of dispersion of the equipment switching time interval; Collect access interval data for device switching ,The data is obtained by monitoring and recording the use of equipment; Calculate the mean switching time interval of all devices : ; Where N is the total number of device switching records; Calculate the standard deviation of a time interval : ; Calculate the average absolute value of the differences between adjacent time intervals: ; Substituting the above results into the coefficient of variation formula: ; Specific calculation example: Set the following equipment switching time interval data (unit: minutes) to be obtained by monitoring equipment usage: T=[12,15,14,10,13,16,11,14,15,13]; Calculate the mean : Calculate standard deviation :

[0027] Calculate the mean absolute value MAD of the difference between adjacent time intervals:

[0028] Calculate the coefficient of variation CV: ; The results show that the coefficient of variation of the equipment switching time interval is about 0.136, indicating that the dispersion of the time interval is relatively low.

[0029] The matching evaluation submodule uses the device switching interval parameter to calculate the matching degree of the device access mode, and combines the device interaction characteristics and time stability to obtain the identity matching confidence value; Calculate the degree of device access mode matching, traverse all device interaction records, extract the switching time interval of adjacent device access, and calculate the average switching time of different device pairs. At the same time, count the device access mode, including the proportion of single device access times, the number of times multiple devices are used alternately, the proportion of fast switching times, etc. Combined with the device interaction characteristics, analyze the stability of device usage, calculate the time stability, that is, the variance of the user's access time distribution in different time periods (such as morning, afternoon, and evening). If the variance is small (such as lower than the set threshold of 0.1), it is determined that the user's access time is relatively stable, calculate the identity matching confidence value, and use a weighted average method to comprehensively consider the device access mode matching, time stability and device switching characteristics, set the weights of various indicators (such as device access mode matching 0.4, time stability 0.3, device switching characteristics 0.3), and calculate the matching confidence value. If the confidence value is greater than the set threshold (such as 0.85), it is determined that the devices belong to the same user, and the identity matching confidence value is obtained.

[0030] Specifically, if Figure 2 , 5 As shown, the conversion intention assessment module includes: The user behavior trajectory extraction submodule uses the identity matching confidence value and SAAS user access data to extract the mouse hover time, click path and access degree, filter the interaction records, and obtain the user interaction trajectory data; Using SAAS user access data, the identity matching confidence value is calculated based on the user's unique identity (such as user ID, device ID, browser fingerprint, etc.). This value can be calculated based on similarity, such as Jaccard similarity, cosine similarity, etc., to measure the consistency of access behavior between different devices or accounts. When the confidence value exceeds the set threshold (such as 0.85), it is determined to be the same user, and the mouse hover time is extracted. The user access records are traversed, and the mouse hover time on each page element is calculated, that is, the timestamp of the mouse entering a certain element area minus the timestamp of leaving the area, and stored in an array to obtain the hover time distribution of all page elements. At the same time, the click path is extracted, and the user's click sequence from entering the website to leaving is recorded. It is stored in the form of page ID sequence, such as [homepage→product page→shopping cart→payment page]. The interaction records are filtered to eliminate low-interaction behaviors, such as records of only visiting a single page without click operations, and stored in the database, including identity matching information, mouse hover time, click path, access level, etc., to obtain user interaction trajectory data.

[0031] The interactive feature analysis submodule calculates the number and interval of multiple visits to the same product by users based on user interaction trajectory data, counts the proportion of actions that add products to the shopping cart but do not pay, analyzes the distribution of mouse hovering time, calculates the product page dwell rate and shopping path completeness, extracts core interactive features, and obtains key conversion behavior parameters; Calculate the number and interval of multiple visits to the same product by users, traverse user access records, count the number of visits with the same product ID, and calculate the time interval between visits, that is, the difference between the timestamps of two consecutive visits to the product, and store it in a hash table. At the same time, count the proportion of behaviors that are added to the shopping cart but not paid, filter out records that include the shopping cart page but do not enter the payment page in the access path, and calculate the proportion of such behaviors, that is, the number of unpaid shopping cart records divided by the total number of shopping cart visits. Analyze the distribution of mouse hover time, divide the mouse hover time of all page elements into intervals, such as [0-2s], [2-5s], and [more than 5s], count the number of visits and proportions in each interval, calculate the product page residence ratio, that is, the ratio of the residence time on a specific product page to the total visit time, calculate the shopping path completeness, that is, the proportion of the number of occurrences of the complete shopping path (such as [product page→shopping cart→payment page]) to the total number of shopping paths, extract core interaction features, including visit frequency, hover time distribution, non-payment rate, residence ratio, and path completeness, and obtain key conversion behavior parameters.

[0032] The willingness classification submodule calls the key behavior parameters of conversion, matches the preset classification criteria according to the access frequency, interaction degree and payment behavior pattern, and generates the conversion willingness classification results; According to the visit frequency, interaction level and payment behavior pattern, the preset classification standards are matched and the visit frequency threshold is set. For example, the average daily visit times of high-willing users are ≥3 times, 1-3 times for medium-willing users, and <1 time for low-willing users. The interaction level is calculated, that is, the weighted average of the mouse hovering time, the page dwelling ratio and the shopping path completeness. The weight can be determined according to the user conversion rate, such as the hovering time weight 0.3, the page dwelling ratio weight 0.4, and the path completeness weight 0.3. The payment behavior pattern is analyzed and the payment conversion rate is calculated, that is, the number of payments / total number of visits. If the conversion rate is ≥50%, it is judged as high willingness, 20%-50% is medium willingness, and <20% is low willingness. The user interaction parameters are substituted into the classification rules, matched with the corresponding conversion intention categories, and stored in the user behavior data to generate the conversion intention classification results.

[0033] Specifically, if Figure 2 , 6 As shown, the marketing link optimization module includes: The marketing matching submodule selects user groups with differentiated intentions based on the conversion intention classification results, calculates the interaction difference values ​​of the groups for differentiated marketing methods, selects the contact methods with interaction deviations, matches the optimal marketing contact methods for the groups, and obtains the user marketing contact method matching results; Divide the conversion intention of user groups into intervals. According to the distribution of conversion intention, set three intervals of high, medium and low intention. The definition of high-intention users can be calculated by conversion behavior data, such as users who browsed similar products more than 10 times and added to shopping carts more than 3 times in the past 30 days, and the conversion rate is higher than 60%. Medium-intention users are users with a conversion rate between 30%-60% and 5-10 browsing times. Low-intention users are users with less than 5 browsing times or a conversion rate lower than 30%. For the selected differentiated intention user groups, the click-through rate, conversion rate and stay time under different marketing methods are counted through the user's past interaction data, and the interaction difference values ​​of each marketing method for different user groups are calculated; The interaction difference value E can be calculated by the formula: ; in, Indicates the conversion rate of high willingness groups, Indicates the conversion rate of the low willingness group; When E>1.5, it means that the marketing method has a greater impact on the high-willingness group, and the reach method for the low-willingness group needs to be optimized. When E<0.5, it means that the marketing method is more attractive to the low-willingness group, and the interaction methods of high-willing users need to be increased. Then, the reach methods with large interaction deviations are screened, and the reach effects of different marketing methods are compared, including SMS, email, social media push, in-site messages, etc. The click-through rate and conversion rate of users in each method are counted, and the data is combined to analyze whether there is a significant interaction deviation in a certain method. If the click-through rate and conversion rate of a certain method are 30% lower than the average of all methods, it is considered that the method is not suitable for the current user group and needs to be adjusted to match the optimal marketing reach method for the group. The method with the smallest interaction deviation and click-through rate and conversion rate above the average is selected to obtain the user marketing reach method matching result.

[0034] The promotion push optimization submodule matches the results through user marketing contact methods, screens users with high conversion intention, calculates the limited-time discount strength and inventory reminder interval, calculates the recommended content weight and promotion push period for interested users, calculates the optimal push method for evaluation, review and case of low-interest users, and obtains user promotion push records; Screen users with high conversion intention, that is, calculate the limited-time discount strength based on their behaviors, such as users who browse more than 15 times, add to shopping cart more than 5 times, and have uncompleted order records in the past week. The discount strength D can be calculated by the formula: ; in, represents the standard profit margin, Indicates the target profit margin, which is set according to the inventory turnover rate. For example, if a product currently has 200 pieces in stock and the normal sales cycle is expected to be 30 days, the target profit margin is set to 10%, and the standard profit margin is 25%, then the discount strength is calculated as follows: ; At the same time, the inventory reminder interval is calculated, that is, the user's sensitivity to inventory changes. The reminder interval is established based on the user's browsing frequency and the dynamic relationship between inventory. If a user browses more than 5 times within 24 hours before and after the inventory change, the inventory reminder interval is set to 6 hours. If the number of views is between 2-5 times, it is set to 12 hours. If the number of views is less than 2 times, no inventory reminder is pushed. The weight of the recommended content matched by the interested user is calculated, and it is layered according to the click content type, such as product details page, user evaluation, discount information, brand story, etc. The weight W is calculated according to the click rate of each content type. The weight calculation formula is: ; in, Indicates the number of clicks of a certain type. Indicates the total number of clicks; Select the content with the highest weight for push, and calculate the promotion push time period at the same time, and determine the optimal push time based on the user's active time period. For example, if the user's click-through rate from 20:00 to 22:00 is 20% higher than the daily average, then the push time period is set during this time period, and calculate the optimal push method for evaluations, reviews, and cases for low-interest users. Through data analysis, users' click preferences for evaluations, reviews, and cases, for example, if the evaluation content click-through rate is 40%, the evaluation content click-through rate is 30%, and the case click-through rate is 30%, then the evaluation content will be pushed first, and the click-through rate changes will be tested through SMS, email, etc. Select the method that increases the click-through rate by more than 5% for push, and obtain the user's promotion push record.

[0035] The marketing content response monitoring submodule uses user promotion push records to obtain user interaction data after marketing reach, calculates the marketing response deviation value, screens user groups with reach effect deviation, adjusts the push method, and obtains the optimized marketing reach path; Statistics on users’ interactive behaviors after receiving push notifications, including clicks, purchases, favorites, and bounces, etc., calculate the marketing response deviation value R, the formula is: ; in, Indicates the conversion rate after push. Indicates the conversion rate before push; If R<0.5, it means that the push method has a low impact on users, and the push strategy needs to be optimized to screen user groups with deviations in reach effects, that is, to compare the changes in click-through rate before and after the push. If a user's click-through rate before the push is 5%, and the click-through rate after the push is still lower than 7.5%, it is considered that the user has no significant interest in the current push method, and the push method needs to be adjusted. The adjustment methods include optimizing the push content, modifying the push time, changing the push channel, etc. Push again after optimization, and analyze the interaction data of the optimized users. If the click-through rate of the optimized push increases by more than 10% compared with that before optimization, the optimization is judged to be effective. Otherwise, continue to adjust to obtain the optimized marketing reach path.

[0036] Specifically, if Figure 2 , 7 As shown, the reach strategy adjustment module includes: The user response monitoring submodule detects the user response in SAAS push data through the optimized marketing reach path, extracts push click records, stay duration and interaction times, and obtains user push response results; Detect user response in SAAS push data, traverse SAAS push data, extract push reception records of all users, including push timestamp, push content type (such as product recommendation, inventory reminder, discount information), push channel (such as APP notification, SMS, email), extract push click records, that is, determine the user's click after receiving the push, calculate the push click rate (number of clicks / number of pushes), and store it in a hash table. The key is the user ID and the value is the number of push clicks of the user. Calculate the stay time. For the page entered by clicking the push, record the entry timestamp and exit timestamp, calculate the stay time and store it in the database. At the same time, extract the number of interactions, count the user's operations on the push page, such as button clicks, sliding browsing, page switching, etc., and calculate the interaction depth, that is, the number of interactions / page stay time, including push click rate, stay time, number of interactions and interaction depth, etc., for push strategy optimization analysis to obtain user push response results.

[0037] The reach effect analysis submodule calculates the click-through rate of differentiated push content based on the user push response results, counts the average length of time users stay on the push page, analyzes the impact of push frequency on user response, calculates the push effect difference value, and generates marketing reach effect parameters; Obtain the user's push response data, including the number of clicks after push, dwell time, conversion behavior, etc., and calculate the click-through rate and click-through rate of differentiated push content The calculation formula is ; in, Indicates the number of clicks on the pushed content. Indicates the total number of pushes; If the click-through rate is lower than the global average of 30% of the pushed content, the push content is judged to be of poor effect. The average length of time users stay on the push page is counted. The calculation method is: ; in, Indicates the single page dwelling time of each user. Indicates the total number of users who clicked on the pushed content. If the average dwell time of a pushed content is more than 20% lower than the average of similar pushed content, it is considered that the pushed content is less attractive to users. Analyze the impact of push frequency on user response and count the changes in click-through rate under different push frequencies. For example, record the click-through rate when the content is pushed once, twice, and three times on the same day, and calculate the marginal click-through rate of change of push frequency. , the calculation formula is: ; in, Indicates the click rate after increasing the push frequency. Indicates the click rate of the original push frequency; like If it is lower than 5%, it means that increasing the push frequency has no obvious effect on improving user response. It is necessary to reduce the push frequency and calculate the push effect difference value. If the push effect difference value is lower than -0.2, it means that the current push effect is significantly lower than the average level. It is necessary to optimize the push strategy and generate marketing reach effect parameters.

[0038] The push strategy optimization submodule uses marketing reach effect parameters to adjust push frequency and content, optimize push time windows and information presentation methods, match user interaction characteristics, and generate marketing reach adjustment results; Adjust push frequency and content, using the formula: ; in, Represents push optimization parameters, Represents the content weight of the oth push, represents the average weight of the pushed content, n represents the total number of pushed content, represents the jth user interaction intensity, represents the time interval between the jth user interactions, and m represents the total number of user interactions; Push optimization parameters Aims to measure the matching degree between push content and user interaction in order to optimize push strategy; Push content weight Reflects the importance or relevance of each push content, which can be quantified through indicators such as the click rate and reading time of the pushed content by users. The click rate can be expressed as the ratio of the number of clicks to the number of pushes, and the reading time can be obtained by monitoring the time users stay on the pushed content page; The average weight of all pushed content It is the average weight of all pushed content, which is calculated by dividing the sum of all pushed content weights by the total number of pushed content times n; User interaction intensity Indicates the degree of user response to the pushed content, which can be quantified through the user's likes, comments, shares and other behaviors. The number of likes, comments and shares can be recorded separately, and different weights can be assigned according to the degree of influence on the interaction intensity to calculate the comprehensive interaction intensity value; User interaction interval It refers to the time it takes for a user to interact with a push notification, which can be obtained by recording the difference between the push notification sending time and the user interaction time. The total number of user interactions m refers to the total number of times a user interacts with the pushed content within a specific time period; For calculation , we need to calculate the mean absolute deviation of the pushed content weight: Calculate the average weight of all pushed content : It is set that 5 push notifications are sent within a week, and the push content weights are 0.8, 0.6, 0.9, 0.7 and 0.85 respectively. ; Calculate the absolute value of the difference between the weight of each pushed content and the average weight, and sum them up:

[0039] Calculate the mean absolute deviation: ; Calculate the weighted average of user interaction intensity and time interval: In the above push notifications, the user's interaction intensity (such as the number of likes) and interaction interval (the time from push notification to interaction, measured in hours) are as follows: First push: =50 times, =2 hours; Second push: =30 times, = 3 hours; The third push: = 60 times, = 1.5 hours; 4th push: =40 times, = 2.5 hours; The 5th push: =55 times, =2 hours; Calculate the product of the interaction intensity of each push and the time interval, and sum them up:

[0040] Calculate the weighted average: ; Calculate the square root: ; Calculate push optimization parameters: ; The results show that the comprehensive performance of the current push strategy in terms of content matching and user interaction is If the value is low, it means that the push content is poorly matched or the user's enthusiasm for interaction is low. It is necessary to adjust the quality of the pushed content, optimize the push time window, or adjust the push frequency. A higher value indicates that the current push content has a better match, the user's interaction intensity is higher, and the push strategy is relatively better. It serves as the basis for dynamically adjusting push strategies, so as to optimize push content and timing according to actual conditions and improve the overall effect of marketing reach.

[0041] See also Figure 8 The SAAS method based on mobile digital marketing is executed based on the above SAAS platform based on mobile digital marketing, and includes the following steps: S1: Obtain SAAS user access data, collect page access sequence, click frequency and stay time, calculate device switching interval, analyze access patterns, and obtain stable results of device behavior sequence; S2: Based on the stable results of device behavior sequence, extract device interaction features, calculate device switching interval, count the access frequency of differentiated devices, normalize the interval time, and obtain the identity matching confidence value; S3: According to the identity matching confidence value, the mouse hover time and click path are obtained, the users who repeatedly visit the product are screened, the visit repetition rate is calculated, and the conversion intention classification result is obtained; S4: Based on the conversion intention classification results, filter users with differentiated interest levels, match digital marketing reach methods, set inventory reminders and product recommendations for high-interest users, and obtain optimized marketing reach paths; S5: Based on the optimized marketing reach path, detect user responses in the push data, calculate click-through rate and dwell time, adjust push frequency and content, and obtain marketing reach adjustment results.

[0042] 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. SAAS platform based on mobile digital marketing, characterized by: The platform includes: The behavior sequence module obtains SAAS user access data, sorts the page access order, click frequency and stay time by timestamp, determines whether the device switching time interval meets the continuous access characteristics, analyzes the device access mode and access frequency, and obtains the stability result of the device behavior sequence; The identity matching module extracts the device interaction features of the SAAS user access data according to the device behavior sequence stabilization result, calculates the device switching interval time, and obtains the identity matching confidence value; The conversion intention evaluation module obtains the mouse hover time and click path in the SAAS user access data based on the identity matching confidence value, determines whether the same product is visited multiple times and added to the shopping cart without payment, and obtains the conversion intention classification result; The marketing link optimization module uses the conversion intention classification results to trigger limited-time discounts and inventory reminders for users with high conversion intentions, and push product reviews and user comments to users with low interest, thereby obtaining an optimized marketing reach path; The reach strategy adjustment module detects the user response in the SAAS push data based on the optimized marketing reach path, calculates the click rate and dwell time, adjusts the push frequency and content, and obtains the marketing reach adjustment result.

2. The SAAS platform based on mobile digital marketing according to claim 1, characterized in that: The device behavior sequence stability results include page access order, page access frequency, and multi-device access coordination; the identity matching confidence value includes device recognition degree, user interaction consistency, and identity authentication strength; the conversion intention classification results include product attention, shopping cart behavior analysis, and purchase potential rating; the optimized marketing reach path includes limited-time discount notifications, inventory shortage reminders, product evaluation displays, and user evaluation push; the marketing reach adjustment results include advertising delivery efficiency, content matching degree, and response time optimization results.

3. The SAAS platform based on mobile digital marketing according to claim 1, characterized in that: The behavior sequence module includes: The access record screening submodule obtains SAAS user access data, sorts the page access sequence by timestamp, calculates the page click frequency and stay time, screens the continuous access records of the same device, and obtains the page access behavior sequence; The device switching identification submodule calculates the device identification changes of adjacent access records based on the page access behavior sequence, extracts the device switching time interval, determines whether it is lower than the device continuous access threshold, screens the device switching records that meet the continuous access characteristics, calculates the continuous device switching ratio, analyzes the access interval distribution of differentiated device switching, and establishes a device continuous access identification; The access pattern analysis submodule calculates the device access frequency through the device continuous access identification, counts the access time distribution of differentiated devices, screens the device behavior matching the stable characteristics, and generates a stable result of the device behavior sequence.

4. The SAAS platform based on mobile digital marketing according to claim 1, characterized in that: The identity matching module comprises: The interaction frequency extraction submodule utilizes the device behavior sequence stabilization result and extracts the device interaction frequency, access sequence and switching features between devices according to the SAAS user access data to obtain the device interaction feature data; The device switching time calculation submodule calculates the adjacent access time intervals in the device switching records based on the device interaction feature data, counts the time interval distribution of device switching, calculates the concentration of access time between devices, calculates the mean and coefficient of variation of device switching, and generates device switching interval parameters; The matching degree evaluation submodule uses the device switching interval parameter to calculate the matching degree of the device access mode, and obtains the identity matching confidence value by combining the device interaction characteristics and time stability.

5. The SAAS platform based on mobile digital marketing according to claim 4 is characterized in that: The formula for calculating the coefficient of variation of the device switching is: ； Where CV represents the coefficient of variation of device switching, represents the access time interval of the ith device switching, represents the access time interval of the i-1th device switching, represents the mean of the access time interval of device switching, N represents the total number of device switching, Represents the absolute difference between two adjacent device switching time intervals.

6. The SAAS platform based on mobile digital marketing according to claim 1, characterized in that: The conversion willingness evaluation module includes: The user behavior trajectory extraction submodule uses the identity matching confidence value and SAAS user access data to extract the mouse hover time, click path and access degree, screen the interaction records, and obtain user interaction trajectory data; The interactive feature analysis submodule calculates the number and interval of multiple visits of users to the same product based on the user interaction trajectory data, counts the proportion of actions of adding to the shopping cart but not paying, analyzes the distribution of mouse hovering time, calculates the product page stay rate and shopping path completeness, extracts core interactive features, and obtains key conversion behavior parameters; The willingness classification submodule calls the key conversion behavior parameters, matches the preset classification standards according to the access frequency, interaction degree and payment behavior pattern, and generates a conversion willingness classification result.

7. The SAAS platform based on mobile digital marketing according to claim 1, characterized in that: The marketing link optimization module includes: The marketing matching submodule screens user groups with differentiated intentions based on the conversion intention classification results, calculates the interaction difference values ​​of the groups to differentiated marketing methods, screens the contact methods with interaction deviations, matches the optimal marketing contact methods of the groups, and obtains the user marketing contact method matching results; The promotion push optimization submodule matches the results of the user marketing contact method, screens users with high conversion intention, calculates the limited-time discount strength and inventory reminder interval, calculates the recommended content weight and promotion push period matched by interested users, calculates the optimal push method for evaluation, review and case of low-interest users, and obtains user promotion push records; The marketing content response monitoring submodule utilizes the user promotion push records to obtain user interaction data after marketing reach, calculates the marketing response deviation value, screens user groups with reach effect deviation, adjusts the push method, and obtains an optimized marketing reach path.

8. The SAAS platform based on mobile digital marketing according to claim 1, characterized in that: The reach strategy adjustment module includes: The user response monitoring submodule detects the user response in the SAAS push data through the optimized marketing reach path, extracts push click records, stay duration and interaction times, and obtains the user push response results; The reach effect analysis submodule calculates the click-through rate of differentiated push content based on the user push response results, counts the average stay time of users on the push page, analyzes the impact of push frequency on user response, calculates the push effect difference value, and generates marketing reach effect parameters; The push strategy optimization submodule uses the marketing reach effect parameters to adjust the push frequency and content, optimize the push time window and information presentation method, match user interaction characteristics, and generate marketing reach adjustment results.

9. The SAAS platform based on mobile digital marketing according to claim 8, characterized in that: The adjustment of push frequency and content adopts the formula: ； in, Represents push optimization parameters, Represents the weight of the content pushed for the oth time, represents the average weight of the pushed content, n represents the total number of pushed content, represents the jth user interaction intensity, represents the j-th user interaction time interval, and m represents the total number of user interactions.

10. SAAS method based on mobile digital marketing, characterized by: The SAAS platform based on mobile digital marketing according to any one of claims 1 to 9 comprises the following steps: S1: Obtain SAAS user access data, collect page access sequence, click frequency and stay time, calculate device switching interval, analyze access patterns, and obtain stable results of device behavior sequence; S2: Based on the stable result of the device behavior sequence, extract the device interaction features, calculate the device switching interval, count the access frequency of differentiated devices, normalize the interval time, and obtain the identity matching confidence value; S3: According to the identity matching confidence value, the mouse hover time and click path are obtained, users who repeatedly visit the product are screened, the visit repetition rate is calculated, and the conversion intention classification result is obtained; S4: Based on the conversion intention classification results, screen users with differentiated interest levels, match digital marketing reach methods, set inventory reminders and product recommendations for high-interest users, and obtain an optimized marketing reach path; S5: According to the optimized marketing reach path, detect the user response in the push data, calculate the click rate and dwell time, adjust the push frequency and content, and obtain the marketing reach adjustment result.

Images