SaaS software user stickiness analysis method and system based on multi-source data fusion

Abstract

The present application discloses a method and system for analyzing user stickiness of SaaS software based on multi-source data fusion. The server-side method includes: when the duration between the current moment and the last analysis end moment is equal to the preset polling duration, obtaining user behavior data, log information and business data collected on the platform where the SaaS software is deployed within a preset period, the user behavior data is automatically collected by the preset embedding point, and the preset embedding point is an event pre-set for the user interaction function in the SaaS software for automatically collecting user behavior data; pre-processing user behavior data, log information and business data to obtain data to be analyzed; generating parameter values ​​of each indicator in the pre-built multi-dimensional user stickiness tracking indicator according to the data to be analyzed; sending the parameter values ​​of each indicator to the client to be displayed through a preset dashboard. Therefore, the use of the embodiments of the present application can effectively integrate and analyze complex data sets from different channels, thereby improving the accuracy of user stickiness analysis.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method and system for analyzing user stickiness of SaaS software based on multi-source data fusion. Background Art

[0002] In the field of Software as a Service (SaaS), it is very important for providers of manufacturing SaaS software to analyze user stickiness. This analysis is crucial for optimizing user experience, improving user retention, and enhancing service personalization. Analysis of user stickiness includes but is not limited to analysis of user usage habits, identification of system function preferences, objective measurement of project implementation effects, and early warning of user churn.

[0003] Currently, the technical means used to analyze user stickiness include manual data collection, questionnaire-based user feedback, and simple data analysis tools. These technical means usually rely on users to actively provide feedback, lack automation and real-time performance, and cannot fully capture user usage data on the SaaS platform. In addition, existing data analysis methods are often limited to a single data source and cannot effectively integrate and analyze complex data sets from different channels, thereby reducing the accuracy of user stickiness analysis. Summary of the invention

[0004] The embodiments of the present application provide a method and system for analyzing SaaS software user stickiness based on multi-source data fusion. In order to have a basic understanding of some aspects of the disclosed embodiments, a simple summary is given below. This summary is not a general review, nor is it intended to identify key / important components or describe the scope of protection of these embodiments. Its only purpose is to present some concepts in a simple form as a preface to the detailed description that follows.

[0005] In a first aspect, an embodiment of the present application provides a SaaS software user stickiness analysis method based on multi-source data fusion, which is applied to a server, and the method includes:

[0006] When the time between the current moment and the end moment of the last SaaS software user stickiness analysis is equal to the preset polling time, obtain the user behavior data, log information and business data collected on the platform where the SaaS software is deployed within the preset period. The user behavior data is automatically collected by the preset tracking point. The preset tracking point is an event pre-set for the user interaction function in the SaaS software for automatically collecting user behavior data.

[0007] Preprocess user behavior data, log information, and business data to obtain data to be analyzed;

[0008] Generate parameter values ​​of each indicator in the pre-built multi-dimensional user stickiness tracking indicator based on the data to be analyzed; wherein the pre-built multi-dimensional user stickiness tracking indicator is constructed based on user behavior patterns and user preference trend information, which is obtained by analyzing sample data using a big data processing framework and a machine learning algorithm, and the sample data is obtained by obtaining and pre-processing historical user behavior data, historical log information, and historical business data collected on a platform where the SaaS software is deployed within a preset time period;

[0009] Send the parameter values ​​of each indicator to the client to be displayed through the preset dashboard.

[0010] Optionally, follow these steps to generate pre-built multi-dimensional user stickiness tracking metrics, including:

[0011] Obtain historical user behavior data, historical log information, and historical business data collected on the platform where the SaaS software is deployed within a preset time period;

[0012] Use the big data processing framework to process historical user behavior data, historical log information, and historical business data to obtain sample data;

[0013] Input the sample data into a pre-trained SaaS software user stickiness analysis model; wherein the pre-trained SaaS software user stickiness analysis model is trained based on a machine learning algorithm;

[0014] Output user behavior patterns and user preference trend information corresponding to sample data;

[0015] According to the user behavior pattern and the user preference trend information, a multi-dimensional user stickiness tracking indicator is constructed to obtain a pre-constructed multi-dimensional user stickiness tracking indicator.

[0016] Optional, big data processing frameworks include the MapReduce programming model and Spark’s distributed computing engine;

[0017] Using the big data processing framework, historical user behavior data, historical log information, and historical business data are processed to obtain sample data, including:

[0018] Perform data cleaning and data conversion on historical user behavior data, historical log information, and historical business data;

[0019] For data cleaning and data conversion of historical user behavior data, historical log information, and historical business data, use the MapReduce programming model to perform parallel data processing;

[0020] For historical user behavior data, historical log information, and historical business data that are processed in parallel, Spark's distributed computing engine is used to transform and aggregate the data to obtain multi-source fused data.

[0021] Feature data associated with user stickiness analysis is extracted from multi-source fusion data as sample data, and the feature data includes feature data for characterizing user activity, feature data for characterizing user satisfaction, and feature data for characterizing user loyalty.

[0022] Optionally, build multi-dimensional user stickiness tracking indicators based on user behavior patterns and user preference trend information, including:

[0023] Perform feature encoding on the user behavior information in the user behavior pattern to obtain user behavior features;

[0024] Perform feature encoding on user preference trend information to obtain user preference features;

[0025] Use user behavior characteristics as activity characteristics;

[0026] From the user preference features, the features related to user satisfaction and user loyalty are divided out respectively, and satisfaction features and loyalty features are obtained;

[0027] Obtaining a predefined initial activity indicator sequence, an initial satisfaction indicator sequence, and an initial loyalty indicator sequence;

[0028] Analyze the first correlation between the activity feature and each initial activity indicator in the initial activity indicator sequence, and use the initial activity indicator whose first correlation is greater than or equal to a preset correlation threshold as the target activity indicator;

[0029] Analyze the second correlation between the satisfaction feature and each initial satisfaction indicator in the initial satisfaction indicator sequence, and take the initial satisfaction indicator whose second correlation is greater than or equal to a preset correlation threshold as the target satisfaction indicator;

[0030] Analyze the third correlation between the loyalty feature and each initial loyalty indicator in the initial loyalty indicator sequence, and take the initial loyalty indicator whose third correlation is greater than or equal to a preset correlation threshold as the target loyalty indicator;

[0031] The target activity index, target satisfaction index and target loyalty index are used as multi-dimensional user stickiness tracking indicators; among them, correlation is used to characterize the relationship between different characteristics and known user stickiness.

[0032] Optionally, follow the steps below to generate a pre-trained SaaS software user stickiness analysis model, including:

[0033] Obtain sample user behavior data, sample log information, and sample business data collected on the platform where the SaaS software is deployed within a preset time period;

[0034] Use the big data processing framework to pre-process sample user behavior data, sample log information, and sample business data to obtain historical multi-source fusion data;

[0035] Extract historical feature data for representing user activity, historical feature data for representing user satisfaction, and historical feature data for representing user loyalty from historical multi-source fusion data;

[0036] Determine historical user behavior patterns and historical user preference trend information corresponding to the historical feature data;

[0037] Use historical user behavior patterns and historical user preference trend information as data labels for historical feature data to perform data annotation and obtain model training samples;

[0038] Based on the model training samples, a pre-trained SaaS software user stickiness analysis model is generated.

[0039] Optionally, determining historical user behavior patterns and historical user preference trend information corresponding to the historical feature data includes:

[0040] For each historical feature data, receiving the first user behavior pattern and the first user preference trend information input by the client as the historical user behavior pattern and the historical user preference trend information corresponding to the historical feature data; wherein the first user behavior pattern and the first user preference trend information are obtained by analyzing each historical feature data by business experts and data analysts; or,

[0041] Obtain user behavior characteristics, user feedback characteristics or user comment characteristics from historical feature data; use a sequential pattern mining algorithm to mine frequent patterns of user behavior characteristics as historical user behavior patterns; generate historical user preference trend information based on user feedback characteristics or user comment characteristics.

[0042] Optionally, historical user preference trend information is generated based on user feedback features or user comment features, including:

[0043] Performing semantic analysis on the user feedback features or user comment features to extract a user preference keyword sequence corresponding to the user feedback features or user comment features;

[0044] The user's preferred keyword sequence is used as the demand text for guiding the preset large language model to obtain the preference prompt information;

[0045] The preference prompt information is input into a preset large language model, and historical user preference trend information is determined through the preset large language model.

[0046] Optionally, the model training samples include a mapping relationship between historical user behavior patterns and historical user preference trend information and historical feature data;

[0047] Based on the model training samples, a pre-trained SaaS software user stickiness analysis model is generated, including:

[0048] Use decision tree algorithm to create a SaaS software user stickiness analysis model;

[0049] Input the mapping relationship between historical user behavior patterns and historical user preference trend information and historical feature data into the SaaS software user stickiness analysis model, and output the model loss value;

[0050] When the model loss value reaches the minimum, a pre-trained SaaS software user stickiness analysis model is generated; or, when the model loss value does not reach the minimum, the model loss value is back-propagated to update the model parameters of the SaaS software user stickiness analysis model, and the step of inputting the mapping relationship into the SaaS software user stickiness analysis model is continued until the model loss value reaches the minimum; wherein, the loss function of the SaaS software user stickiness analysis model is:

[0051]

[0052] in, is the model loss value, is the label value of the historical user behavior pattern and historical user preference trend information in each mapping relationship, It is the probability that the historical feature data in each mapping relationship predicted by the model carries its own label.

[0053] Optionally, each of the pre-built multi-dimensional user stickiness tracking indicators is a target activity indicator, a target satisfaction indicator, and a target loyalty indicator;

[0054] Based on the data to be analyzed, the parameter values ​​of each indicator in the pre-built multi-dimensional user stickiness tracking indicator are generated, including:

[0055] Determine, from the data to be analyzed, the number of user logins within a preset period, the duration of each user session, and the total number of pages viewed by the user within the preset period;

[0056] The ratio of the number of user logins to the number of days in the preset cycle is taken as the average number of daily logins; the ratio of the duration of each user session to the total number of users is taken as the average session duration; the total number of pages browsed by the user is taken as the number of page views; the average number of daily logins, average session duration and page views are taken as the parameter values ​​of the target activity indicator;

[0057] From the data to be analyzed, determine the time from customer service request to service response, the proportion of customer issues resolved, and customer feedback rating data;

[0058] Calculate the average response time from the time it takes for a customer to request a service and get a service response as the customer service response time; calculate the average score based on customer feedback rating data as the customer feedback score; take the proportion of customer problems that have been solved as the problem resolution rate; take the customer service response time, customer feedback score, and problem resolution rate as parameter values ​​for the target satisfaction indicator;

[0059] From the data to be analyzed, count the proportion of users who make repeat purchases within a preset period, the proportion of users who recommend new users, and the proportion of users who continue to use the service within a preset period;

[0060] The repeat purchase rate is the proportion of users who make repeat purchases within a preset period, the recommendation rate is the proportion of users who recommend new users, and the long-term retention rate is the proportion of users who continue to use the service within a preset period.

[0061] Repeat purchase rate, recommendation rate and long-term retention rate are used as parameter values ​​of target loyalty indicators.

[0062] In a second aspect, the embodiment of the present application provides a SaaS software user stickiness analysis system based on multi-source data fusion, the system comprising:

[0063] The multi-source data acquisition module is used to acquire user behavior data, log information and business data collected on the platform where the SaaS software is deployed within a preset period when the time between the current moment and the end moment of the previous SaaS software user stickiness analysis is equal to the preset polling time. The user behavior data is automatically collected by the preset tracking point, which is an event pre-set for the user interaction function in the SaaS software for automatically collecting user behavior data;

[0064] Data preprocessing module, used to preprocess user behavior data, log information and business data to obtain data to be analyzed;

[0065] A parameter value generation module is used to generate parameter values ​​of each indicator in a pre-built multi-dimensional user stickiness tracking indicator according to the data to be analyzed; wherein the pre-built multi-dimensional user stickiness tracking indicator is constructed based on user behavior patterns and user preference trend information, and the user behavior patterns and user preference trend information are obtained by analyzing sample data using a big data processing framework and a machine learning algorithm, and the sample data is obtained by obtaining and pre-processing historical user behavior data, historical log information, and historical business data collected on a platform where the SaaS software is deployed within a preset time period;

[0066] The parameter value display module is used to send the parameter value of each indicator to the client for display through the preset dashboard.

[0067] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0068] In the embodiment of the present application, on the one hand, the server effectively integrates and analyzes complex data sets from different channels by using big data processing technology and machine learning algorithms, thereby deeply exploring user behavior patterns and preference trends, which are then used to construct accurate multi-dimensional user stickiness tracking indicators. The multi-dimensional user stickiness tracking indicators provide a clear direction for data quantification, significantly improving the accuracy of user stickiness analysis; on the other hand, the multi-dimensional user stickiness tracking indicators can quantify the performance of user behavior, logs and business data over a certain period of time, and intuitively display the quantification results in the form of parameter values ​​through a customized dashboard. Managers can quickly grasp key performance indicators (KPIs), intuitively understand the global view and dynamic changes of user stickiness, and then quickly formulate targeted user retention strategies and product optimization measures to improve user satisfaction and loyalty.

[0069] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0071] Figure 1 This is a method flow chart of a SaaS software user stickiness analysis method based on multi-source data fusion provided in an embodiment of the present application;

[0072] Figure 2 It is a data table about user behavior data, business data and log information provided by the embodiment of the present application;

[0073] Figure 3It is a data table of the indicator values ​​of each indicator in a multi-dimensional user stickiness tracking indicator provided in an embodiment of the present application;

[0074] Figure 4A This is a user interface diagram of a user stickiness analysis dashboard displayed by a client provided in an embodiment of the present application;

[0075] Figure 4B This is a user interface diagram of a user stickiness analysis dashboard displayed by a client provided in an embodiment of the present application;

[0076] Figure 5 This is a flowchart of a SaaS software user stickiness analysis model training method provided by this application;

[0077] Figure 6 This is a structural diagram of a SaaS software user stickiness analysis system based on multi-source data fusion provided by this application;

[0078] Figure 7 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0079] The following description and the accompanying drawings sufficiently illustrate specific embodiments of the present application to enable those skilled in the art to practice them.

[0080] It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0081] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Instead, they are only examples of systems and methods consistent with some aspects of the present application as detailed in the attached claims.

[0082] In the description of the present application, it should be understood that the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances. In addition, in the description of the present application, unless otherwise specified, "multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects associated before and after are in an "or" relationship.

[0083] Currently, the technical means used to analyze user stickiness include manual data collection, questionnaire-based user feedback, and simple data analysis tools.

[0084] The inventors realize that these technical means usually rely on users to actively provide feedback, lack automation and real-time performance, and cannot fully capture users' usage data on the SaaS platform. In addition, existing data analysis methods are often limited to a single data source and cannot effectively integrate and analyze complex data sets from different channels, thereby reducing the accuracy of user stickiness analysis.

[0085] In order to solve the above problems, the present application provides a SaaS software user stickiness analysis method and system based on multi-source data fusion to solve the problems existing in the above-mentioned related technical problems. In the embodiment of the present application, on the one hand, the server effectively integrates and analyzes complex data sets from different channels by using big data processing technology and machine learning algorithms, so as to deeply explore user behavior patterns and preference trends. User behavior patterns and preference trends are then used to construct accurate multi-dimensional user stickiness tracking indicators. The multi-dimensional user stickiness tracking indicators provide a clear direction for data quantification and significantly improve the accuracy of user stickiness analysis; on the other hand, the multi-dimensional user stickiness tracking indicators can quantify the performance of user behavior, logs and business data in a certain period of time, and intuitively display the quantification results in the form of parameter values ​​through a customized dashboard. Managers can quickly grasp key performance indicators (KPIs), intuitively understand the global view and dynamic changes of user stickiness, and then quickly formulate targeted user retention strategies and product optimization measures to improve user satisfaction and loyalty. The following uses an exemplary embodiment for detailed description.

[0086] The following will be combined with the attached Figure 1 -Attached Figure 5 , a SaaS software user stickiness analysis method based on multi-source data fusion provided in an embodiment of the present application is introduced in detail. The method can be implemented by a computer program and can be run on a SaaS software user stickiness analysis system based on multi-source data fusion based on the von Neumann system. The computer program can be integrated into an application or run as an independent tool application.

[0087] See also Figure 1 , a flowchart of a SaaS software user stickiness analysis method based on multi-source data fusion is provided for an embodiment of the present application, which is applied to the server. Figure 1 As shown, the method of the embodiment of the present application may include the following steps:

[0088] S101, when the time between the current moment and the end moment of the last SaaS software user stickiness analysis is equal to the preset polling time, obtain user behavior data, log information, and business data collected on the platform where the SaaS software is deployed within a preset period, wherein the user behavior data is automatically collected by a preset tracking point, and the preset tracking point is an event pre-set for a user interaction function in the SaaS software for automatically collecting user behavior data;

[0089] Among them, SaaS software (Software as a Service) allows users to access applications hosted by providers through the Internet. The preset polling duration is the time interval at which the system regularly performs an operation, such as automatically performing data collection every hour or every day. The preset cycle is a specified time period for collecting and analyzing data, such as the past week, month, or quarter. User behavior data is data generated when users interact with SaaS software, such as clicks, browsing, transactions, and other behaviors. Log information is a record generated by system, application, or user activity, containing information such as timestamps and event types. Business data is data directly related to the business processes of SaaS software, such as order information, customer feedback, financial data, etc. Preset tracking points are monitoring points pre-set in the code to capture and record user interaction behaviors on specific functions.

[0090] In some embodiments of the present application, for example, there is a SaaS software that provides online collaborative document editing services, and tracking points are deployed at key interaction points of the document editing function, such as saving, sharing, commenting, etc. Whenever a user performs the above interaction, the preset tracking point is automatically triggered, and the event is recorded and stored in the database. The system automatically records log information such as user login, document access, system errors, etc., and collects business data related to user document editing, such as the number of documents created, the number of edits, etc. All collected data is stored in the database. Assuming that it is decided to poll once every 24 hours, and the preset period is the past 7 days, when the time between the current moment and the end moment of the previous SaaS software user stickiness analysis is equal to 24 hours, obtain the user behavior data, log information, and business data collected on the platform where the SaaS software is deployed in the past 7 days, for example Figure 2 shown.

[0091] S102, pre-processing user behavior data, log information and business data to obtain data to be analyzed;

[0092] In some embodiments of the present application, duplicate records in user behavior data, log information, and business data are deleted through SQL queries or programming scripts. For missing user behavior data, log information, or business data, these records are selected to be filled (such as using the average, median, or mode) or deleted, and all timestamps are converted to a unified format, such as the ISO8601 standard (YYYY-MM-DDTHH:MM:SS). After the processing is completed, the data to be analyzed is obtained.

[0093] S103, generating parameter values ​​of each indicator in a pre-constructed multi-dimensional user stickiness tracking indicator according to the data to be analyzed; wherein the pre-constructed multi-dimensional user stickiness tracking indicator is constructed based on user behavior patterns and user preference trend information, and the user behavior patterns and user preference trend information are obtained by analyzing sample data using a big data processing framework and a machine learning algorithm, and the sample data is obtained by obtaining and pre-processing historical user behavior data, historical log information, and historical business data collected on a platform where the SaaS software is deployed within a preset time period;

[0094] The pre-built multi-dimensional user stickiness tracking indicators include target activity indicator, target satisfaction indicator and target loyalty indicator. The parameter value of each indicator refers to the specific value of each user stickiness tracking indicator, such as the average number of daily logins, average session duration, page views, etc.

[0095] In some embodiments of the present application, the specific process of generating the parameter values ​​of each indicator in the pre-built multi-dimensional user stickiness tracking indicator according to the data to be analyzed includes: determining the number of user logins within a preset period, the duration of each user session, and the total number of pages browsed by the user within the preset period from the data to be analyzed; taking the ratio of the number of user logins to the number of days in the preset period as the average daily number of logins; taking the ratio of the duration of each user session to the total number of users as the average session duration; taking the total number of pages browsed by the user as the number of page views; taking the average daily number of logins, the average session duration, and the number of page views as the parameter values ​​of the target activity indicator; determining the time from customer request for service to service response, the proportion of customer problems solved, and customer feedback rating data from the data to be analyzed; and calculating the target activity indicator according to the number of days from the customer request for service. The average response time is calculated from the time from service to service response, which is used as the customer service response time; the average score is calculated based on the customer feedback score data as the customer feedback score; the proportion of customer problems that have been solved is used as the problem resolution rate; the customer service response time, customer feedback score and problem resolution rate are used as parameter values ​​of the target satisfaction indicator; from the data to be analyzed, the proportion of users who repeat purchases within a preset period, the proportion of users who recommend new users and the proportion of users who continue to use the service within a preset period are counted; the proportion of users who repeat purchases within a preset period is used as the repeat purchase rate, the proportion of users who recommend new users is used as the recommendation rate, and the proportion of users who continue to use the service within a preset period is used as the long-term retention rate; the repeat purchase rate, recommendation rate and long-term retention rate are used as parameter values ​​of the target loyalty indicator. Among them, the parameter values ​​of each indicator in the pre-constructed multi-dimensional user stickiness tracking indicator are as follows: Figure 3 shown.

[0096] In some embodiments of the present application, the specific process of generating a pre-built multi-dimensional user stickiness tracking indicator includes: obtaining historical user behavior data, historical log information, and historical business data collected on a platform where SaaS software is deployed within a preset time period; using a big data processing framework to process the historical user behavior data, historical log information, and historical business data to obtain sample data; inputting the sample data into a pre-trained SaaS software user stickiness analysis model; wherein the pre-trained SaaS software user stickiness analysis model is trained based on a machine learning algorithm; outputting user behavior patterns and user preference trend information corresponding to the sample data; based on the user behavior patterns and user preference trend information, constructing a multi-dimensional user stickiness tracking indicator to obtain a pre-built multi-dimensional user stickiness tracking indicator.

[0097] Among them, the pre-trained SaaS software user stickiness analysis model refers to a model trained using historical data to analyze user stickiness to SaaS software. User behavior patterns are the behavioral habits and rules of users on the SaaS platform. User preference trend information is the user's preference and trend for SaaS software functions or services. Multi-dimensional user stickiness tracking indicators are a set of indicators that measure user stickiness from multiple perspectives, such as activity, satisfaction, and loyalty.

[0098] In the embodiment of the present application, by collecting historical data on the SaaS platform within a preset time period and using a big data processing framework for cleaning and processing, high-quality sample data can be obtained. By inputting this data into a pre-trained model trained based on a machine learning algorithm, user behavior patterns and preference trend information can be accurately output. Furthermore, the multi-dimensional user stickiness tracking indicators constructed based on this information can help companies fully understand and improve user stickiness to SaaS software, optimize user experience, and develop more effective user retention and growth strategies.

[0099] Among them, the big data processing framework includes the MapReduce programming model and Spark's distributed computing engine.

[0100] In some embodiments of the present application, a big data processing framework is used to process historical user behavior data, historical log information, and historical business data, and the specific process of obtaining sample data includes: data cleaning and data conversion of historical user behavior data, historical log information, and historical business data; using the MapReduce programming model to perform parallel processing of the data cleaned and converted historical user behavior data, historical log information, and historical business data; using Spark's distributed computing engine to convert and aggregate the parallel processed historical user behavior data, historical log information, and historical business data to obtain multi-source fused data; extracting feature data associated with user stickiness analysis from the multi-source fused data as sample data, the feature data including feature data for characterizing user activity, feature data for characterizing user satisfaction, and feature data for characterizing user loyalty.

[0101] In some embodiments of the present application, the specific process of constructing a multi-dimensional user stickiness tracking indicator based on user behavior patterns and user preference trend information includes: feature encoding the user behavior information in the user behavior pattern to obtain user behavior features; feature encoding the user preference trend information to obtain user preference features; using the user behavior features as activity features; from the user preference features, respectively dividing the features related to user satisfaction and user loyalty to obtain satisfaction features and loyalty features; obtaining a pre-defined initial activity indicator sequence, an initial satisfaction indicator sequence, and an initial loyalty indicator sequence; analyzing the first correlation between the activity features and each initial activity indicator in the initial activity indicator sequence, and The initial activity index whose first correlation is greater than or equal to a preset correlation threshold is used as the target activity index; the second correlation between the satisfaction feature and each initial satisfaction index in the initial satisfaction index sequence is analyzed, and the initial satisfaction index whose second correlation is greater than or equal to a preset correlation threshold is used as the target satisfaction index; the third correlation between the loyalty feature and each initial loyalty index in the initial loyalty index sequence is analyzed, and the initial loyalty index whose third correlation is greater than or equal to a preset correlation threshold is used as the target loyalty index; the target activity index, the target satisfaction index and the target loyalty index are used as multi-dimensional user stickiness tracking indicators; wherein the correlation is used to characterize the relationship between different features and known user stickiness.

[0102] The correlation can be measured using the Pearson correlation coefficient.

[0103] S104, sending the parameter value of each indicator to the client to be displayed through a preset dashboard.

[0104] The client here refers to the user interface of the SaaS software, or the interface used within the enterprise to view data analysis results. The preset dashboard refers to a pre-designed data display interface used to intuitively display and monitor key performance indicators (KPIs) and other important data.

[0105] In some embodiments of the present application, for example, it is calculated that the average number of daily logins is 3 times, the average session duration is 30 minutes, and the number of page views is 15 pages. These parameter values ​​are transmitted over the network and sent to the client. On the client, there is a preset dashboard, which includes multiple data views, such as bar charts, line charts, etc. The dashboard is updated in real time to display indicators such as the average number of daily logins, average session duration, and page views. Managers analyze user behavior and preferences based on the data on the dashboard and develop strategies to improve product features and enhance user satisfaction. For example, if it is found that the average session duration is far below the industry standard, you may consider optimizing the user experience or adding new features to increase user stickiness. Among them, the information displayed by the preset dashboard is, for example, Figure 4A and Figure 4B shown.

[0106] In the embodiment of the present application, on the one hand, the server effectively integrates and analyzes complex data sets from different channels by using big data processing technology and machine learning algorithms, thereby deeply mining user behavior patterns and preference trends, which are then used to construct accurate multi-dimensional user stickiness tracking indicators. The multi-dimensional user stickiness tracking indicators provide a clear direction for data quantification, significantly improving the accuracy of user stickiness analysis; on the other hand, the multi-dimensional user stickiness tracking indicators can quantify the performance of user behavior, logs and business data over a certain period of time, and intuitively display the quantification results in the form of parameter values ​​through a customized dashboard. Managers can quickly grasp key performance indicators (KPIs), intuitively understand the global view and dynamic changes of user stickiness, and then quickly formulate targeted user retention strategies and product optimization measures to improve user satisfaction and loyalty.

[0107] See also Figure 5 , which is a flowchart of a method for training a SaaS software user stickiness analysis model according to an embodiment of the present application. Figure 5 As shown, the method of the embodiment of the present application may include the following steps:

[0108] S201, obtaining sample user behavior data, sample log information, and sample business data collected on a platform where the SaaS software is deployed within a preset time period;

[0109] S202, using a big data processing framework to pre-process sample user behavior data, sample log information, and sample business data to obtain historical multi-source fusion data;

[0110] The preprocessing process is the same as the process of using a big data processing framework to process historical user behavior data, historical log information, and historical business data to obtain sample data. For details, please refer to the explanation of step S103, which will not be repeated here.

[0111] S203, extracting historical feature data for representing user activity, historical feature data for representing user satisfaction, and historical feature data for representing user loyalty from the historical multi-source fusion data;

[0112] Among them, historical multi-source fusion data refers to historical data collected and integrated from different sources, which may include user behavior data, log information, business data, etc. The historical feature data of user activity is used to describe the historical data of the frequency and intensity of user activities on the platform, such as the number of logins, session duration, page views, etc. The historical feature data of user satisfaction is used to describe the historical data of user satisfaction with products or services, such as customer service response time, customer feedback score, problem resolution rate, etc. The historical feature data of user loyalty is used to describe the historical data of user loyalty to the brand, such as repeat purchase rate, recommendation rate, long-term retention rate, etc.

[0113] In some embodiments of the present application, the historical feature data extraction of user activity includes the number of logins, session duration, and page views. Number of logins: Count the number of logins of each user within a certain period (such as the past 30 days). Session duration: Calculate the average session duration of each user for each login. Page views: Record the total number of pages browsed by each user on the platform. The historical feature data of user satisfaction include customer service response time, customer feedback score, and problem resolution rate. Customer service response time: Record the average time from when a user submits a service request to when the service team responds. Customer feedback score: Collect user evaluation scores for purchased products or services. Problem resolution rate: Count the proportion of problems raised by users that have been resolved. The historical feature data of user loyalty include repeat purchase rate, recommendation rate, and long-term retention rate. Repeat purchase rate: Calculate the proportion of users who repeat purchase within a certain period. Recommendation rate: Determine the proportion of users who recommend new users by investigating or analyzing user behavior data. Long-term retention rate: Count the proportion of users who have been continuously active on the platform for more than a certain period of time (such as one year).

[0114] S204, determining historical user behavior patterns and historical user preference trend information corresponding to the historical feature data;

[0115] In some embodiments of the present application, the specific process of determining the historical user behavior patterns and historical user preference trend information corresponding to the historical feature data includes: performing semantic analysis on the user feedback features or the user comment features to extract the user preference keyword sequence corresponding to the user feedback features or the user comment features; using the user preference keyword sequence as the demand text for guiding the preset large language model to obtain preference prompt information; inputting the preference prompt information into the preset large language model, and determining the historical user preference trend information through the preset large language model; wherein the preset large language model is ChatGPT.

[0116] S205, using historical user behavior patterns and historical user preference trend information as data labels for historical feature data to perform data labeling to obtain model training samples;

[0117] S206: Generate a pre-trained SaaS software user stickiness analysis model based on the model training samples.

[0118] Among them, the model training samples include the mapping relationship between historical user behavior patterns and historical user preference trend information and historical feature data.

[0119] In some embodiments of the present application, the specific process of generating a pre-trained SaaS software user stickiness analysis model based on the model training sample includes: using a decision tree algorithm to create a SaaS software user stickiness analysis model; inputting the mapping relationship between historical user behavior patterns and historical user preference trend information and historical feature data into the SaaS software user stickiness analysis model, and outputting a model loss value; when the model loss value reaches a minimum, generating a pre-trained SaaS software user stickiness analysis model; or, when the model loss value does not reach a minimum, back-propagating the model loss value to update the model parameters of the SaaS software user stickiness analysis model, and continuing to execute the step of inputting the mapping relationship into the SaaS software user stickiness analysis model until the model loss value reaches a minimum; wherein the loss function of the SaaS software user stickiness analysis model is:

[0120]

[0121] in, is the model loss value, is the label value of the historical user behavior pattern and historical user preference trend information in each mapping relationship, It is the probability that the historical feature data in each mapping relationship predicted by the model carries its own label.

[0122] In the embodiment of the present application, on the one hand, the server effectively integrates and analyzes complex data sets from different channels by using big data processing technology and machine learning algorithms, thereby deeply exploring user behavior patterns and preference trends, which are then used to construct accurate multi-dimensional user stickiness tracking indicators. The multi-dimensional user stickiness tracking indicators provide a clear direction for data quantification, significantly improving the accuracy of user stickiness analysis; on the other hand, the multi-dimensional user stickiness tracking indicators can quantify the performance of user behavior, logs and business data over a certain period of time, and intuitively display the quantification results in the form of parameter values ​​through a customized dashboard. Managers can quickly grasp key performance indicators (KPIs), intuitively understand the global view and dynamic changes of user stickiness, and then quickly formulate targeted user retention strategies and product optimization measures to improve user satisfaction and loyalty.

[0123] The following is a system embodiment of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the system embodiment of the present application, please refer to the method embodiment of the present application.

[0124] See also Figure 6 , which shows a schematic diagram of the structure of a SaaS software user stickiness analysis system based on multi-source data fusion provided by an exemplary embodiment of the present application. The SaaS software user stickiness analysis system based on multi-source data fusion can be implemented as all or part of an electronic device through software, hardware, or a combination of both. The system 1 includes a multi-source data acquisition module 10, a data preprocessing module 20, a parameter value generation module 30, and a parameter value display module 40.

[0125] The multi-source data acquisition module 10 is used to acquire user behavior data, log information and business data collected on the platform where the SaaS software is deployed within a preset period when the time between the current moment and the end moment of the previous SaaS software user stickiness analysis is equal to the preset polling time. The user behavior data is automatically collected by the preset embedding point, and the preset embedding point is an event pre-set for the user interaction function in the SaaS software for automatically collecting user behavior data;

[0126] The data preprocessing module 20 is used to preprocess user behavior data, log information and business data to obtain data to be analyzed;

[0127] The parameter value generating module 30 is used to generate the parameter value of each indicator in the pre-built multi-dimensional user stickiness tracking indicator according to the data to be analyzed; wherein the pre-built multi-dimensional user stickiness tracking indicator is constructed based on the user behavior pattern and the user preference trend information, and the user behavior pattern and the user preference trend information are obtained by analyzing the sample data using the big data processing framework and the machine learning algorithm, and the sample data is obtained by obtaining and pre-processing the historical user behavior data, historical log information and historical business data collected on the platform where the SaaS software is deployed within a preset time period;

[0128] The parameter value display module 40 is used to send the parameter value of each indicator to the client for display through a preset dashboard.

[0129] It should be noted that the SaaS software user stickiness analysis system based on multi-source data fusion provided in the above embodiment only uses the division of the above functional modules as an example when executing the SaaS software user stickiness analysis method based on multi-source data fusion. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the SaaS software user stickiness analysis system based on multi-source data fusion provided in the above embodiment and the SaaS software user stickiness analysis method based on multi-source data fusion embodiment belong to the same concept. The implementation process is detailed in the method embodiment, which will not be repeated here.

[0130] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0131] In the embodiment of the present application, on the one hand, the server effectively integrates and analyzes complex data sets from different channels by using big data processing technology and machine learning algorithms, thereby deeply mining user behavior patterns and preference trends, which are then used to construct accurate multi-dimensional user stickiness tracking indicators. The multi-dimensional user stickiness tracking indicators provide a clear direction for data quantification, significantly improving the accuracy of user stickiness analysis; on the other hand, the multi-dimensional user stickiness tracking indicators can quantify the performance of user behavior, logs and business data over a certain period of time, and intuitively display the quantification results in the form of parameter values ​​through a customized dashboard. Managers can quickly grasp key performance indicators (KPIs), intuitively understand the global view and dynamic changes of user stickiness, and then quickly formulate targeted user retention strategies and product optimization measures to improve user satisfaction and loyalty.

[0132] The present application also provides a computer-readable medium having program instructions stored thereon, which, when executed by a processor, implement the SaaS software user stickiness analysis method based on multi-source data fusion provided by the above-mentioned various method embodiments.

[0133] The present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the SaaS software user stickiness analysis method based on multi-source data fusion of the above-mentioned various method embodiments.

[0134] See also Figure 7 , is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 7 As shown, the electronic device 1000 may include: at least one processor 1001 , at least one network interface 1004 , a user interface 1003 , a memory 1005 , and at least one communication bus 1002 .

[0135] The communication bus 1002 is used to realize the connection and communication between these components.

[0136] The user interface 1003 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 1003 may also include a standard wired interface and a wireless interface.

[0137] The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0138] Among them, the processor 1001 may include one or more processing cores. The processor 1001 uses various interfaces and lines to connect various parts of the entire electronic device 1000, and executes various functions and processes data of the electronic device 1000 by running or executing instructions, programs, code sets or instruction sets stored in the memory 1005, and calling data stored in the memory 1005. Optionally, the processor 1001 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 1001 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 1001, and it can be implemented separately through a chip.

[0139] Among them, the memory 1005 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 1005 includes a non-transitory computer-readable storage medium. The memory 1005 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 1005 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 1005 may also be optionally at least one storage system located away from the aforementioned processor 1001. As Figure 7 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a SaaS software user stickiness analysis application based on multi-source data fusion.

[0140] exist Figure 7 In the electronic device 1000 shown, the user interface 1003 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 1001 can be used to call the SaaS software user stickiness analysis application based on multi-source data fusion stored in the memory 1005, and specifically perform the following operations:

[0141] When the time between the current moment and the end moment of the last SaaS software user stickiness analysis is equal to the preset polling time, obtain the user behavior data, log information and business data collected on the platform where the SaaS software is deployed within the preset period. The user behavior data is automatically collected by the preset tracking point. The preset tracking point is an event pre-set for the user interaction function in the SaaS software for automatically collecting user behavior data.

[0142] Preprocess user behavior data, log information, and business data to obtain data to be analyzed;

[0143] Generate parameter values ​​of each indicator in the pre-built multi-dimensional user stickiness tracking indicator based on the data to be analyzed; wherein the pre-built multi-dimensional user stickiness tracking indicator is constructed based on user behavior patterns and user preference trend information, which is obtained by analyzing sample data using a big data processing framework and a machine learning algorithm, and the sample data is obtained by obtaining and pre-processing historical user behavior data, historical log information, and historical business data collected on a platform where the SaaS software is deployed within a preset time period;

[0144] Send the parameter values ​​of each indicator to the client to be displayed through the preset dashboard.

[0145] In one embodiment, when the processor 1001 generates a pre-built multi-dimensional user stickiness tracking indicator, the processor 1001 specifically performs the following operations:

[0146] Obtain historical user behavior data, historical log information, and historical business data collected on the platform where the SaaS software is deployed within a preset time period;

[0147] Use the big data processing framework to process historical user behavior data, historical log information, and historical business data to obtain sample data;

[0148] Input the sample data into a pre-trained SaaS software user stickiness analysis model; wherein the pre-trained SaaS software user stickiness analysis model is trained based on a machine learning algorithm;

[0149] Output user behavior patterns and user preference trend information corresponding to sample data;

[0150] According to the user behavior pattern and the user preference trend information, a multi-dimensional user stickiness tracking indicator is constructed to obtain a pre-constructed multi-dimensional user stickiness tracking indicator.

[0151] In one embodiment, when the processor 1001 processes historical user behavior data, historical log information, and historical business data using a big data processing framework to obtain sample data, the processor 1001 specifically performs the following operations:

[0152] Perform data cleaning and data conversion on historical user behavior data, historical log information, and historical business data;

[0153] For data cleaning and data conversion of historical user behavior data, historical log information, and historical business data, use the MapReduce programming model to perform parallel data processing;

[0154] For historical user behavior data, historical log information, and historical business data that are processed in parallel, Spark's distributed computing engine is used to transform and aggregate the data to obtain multi-source fused data.

[0155] Feature data associated with user stickiness analysis is extracted from multi-source fusion data as sample data, and the feature data includes feature data for characterizing user activity, feature data for characterizing user satisfaction, and feature data for characterizing user loyalty.

[0156] In one embodiment, when the processor 1001 constructs a multi-dimensional user stickiness tracking indicator according to the user behavior pattern and the user preference trend information, the processor 1001 specifically performs the following operations:

[0157] Perform feature encoding on the user behavior information in the user behavior pattern to obtain user behavior features;

[0158] Perform feature encoding on user preference trend information to obtain user preference features;

[0159] Use user behavior characteristics as activity characteristics;

[0160] From the user preference features, the features related to user satisfaction and user loyalty are divided out respectively, and satisfaction features and loyalty features are obtained;

[0161] Obtaining a predefined initial activity indicator sequence, an initial satisfaction indicator sequence, and an initial loyalty indicator sequence;

[0162] Analyze the first correlation between the activity feature and each initial activity indicator in the initial activity indicator sequence, and use the initial activity indicator whose first correlation is greater than or equal to a preset correlation threshold as the target activity indicator;

[0163] Analyze the second correlation between the satisfaction feature and each initial satisfaction indicator in the initial satisfaction indicator sequence, and take the initial satisfaction indicator whose second correlation is greater than or equal to a preset correlation threshold as the target satisfaction indicator;

[0164] Analyze the third correlation between the loyalty feature and each initial loyalty indicator in the initial loyalty indicator sequence, and take the initial loyalty indicator whose third correlation is greater than or equal to a preset correlation threshold as the target loyalty indicator;

[0165] The target activity index, target satisfaction index and target loyalty index are used as multi-dimensional user stickiness tracking indicators; among them, correlation is used to characterize the relationship between different characteristics and known user stickiness.

[0166] In one embodiment, when the processor 1001 generates a pre-trained SaaS software user stickiness analysis model, the processor 1001 specifically performs the following operations:

[0167] Obtain sample user behavior data, sample log information, and sample business data collected on the platform where the SaaS software is deployed within a preset time period;

[0168] Use the big data processing framework to pre-process sample user behavior data, sample log information, and sample business data to obtain historical multi-source fusion data;

[0169] Extract historical feature data for representing user activity, historical feature data for representing user satisfaction, and historical feature data for representing user loyalty from historical multi-source fusion data;

[0170] Determine historical user behavior patterns and historical user preference trend information corresponding to the historical feature data;

[0171] Use historical user behavior patterns and historical user preference trend information as data labels for historical feature data to perform data annotation and obtain model training samples;

[0172] Based on the model training samples, a pre-trained SaaS software user stickiness analysis model is generated.

[0173] In one embodiment, the processor 1001 specifically performs the following operations when determining the historical user behavior pattern and the historical user preference trend information corresponding to the historical feature data:

[0174] For each historical feature data, receiving the first user behavior pattern and the first user preference trend information input by the client as the historical user behavior pattern and the historical user preference trend information corresponding to the historical feature data; wherein the first user behavior pattern and the first user preference trend information are obtained by analyzing each historical feature data by business experts and data analysts; or,

[0175] Obtain user behavior characteristics, user feedback characteristics or user comment characteristics from historical feature data; use a sequential pattern mining algorithm to mine frequent patterns of user behavior characteristics as historical user behavior patterns; generate historical user preference trend information based on user feedback characteristics or user comment characteristics.

[0176] In one embodiment, when the processor 1001 generates historical user preference trend information according to user feedback features or user comment features, the processor 1001 specifically performs the following operations:

[0177] Performing semantic analysis on the user feedback features or user comment features to extract a user preference keyword sequence corresponding to the user feedback features or user comment features;

[0178] The user's preferred keyword sequence is used as the demand text for guiding the preset large language model to obtain the preference prompt information;

[0179] The preference prompt information is input into a preset large language model, and the historical user preference trend information is determined through the preset large language model; wherein the preset large language model is ChatGPT.

[0180] In one embodiment, when the processor 1001 generates a pre-trained SaaS software user stickiness analysis model according to the model training sample, the processor 1001 specifically performs the following operations:

[0181] Use decision tree algorithm to create a SaaS software user stickiness analysis model;

[0182] Input the mapping relationship between historical user behavior patterns and historical user preference trend information and historical feature data into the SaaS software user stickiness analysis model, and output the model loss value;

[0183] When the model loss value reaches the minimum, a pre-trained SaaS software user stickiness analysis model is generated; or, when the model loss value does not reach the minimum, the model loss value is back-propagated to update the model parameters of the SaaS software user stickiness analysis model, and the step of inputting the mapping relationship into the SaaS software user stickiness analysis model is continued until the model loss value reaches the minimum; wherein, the loss function of the SaaS software user stickiness analysis model is:

[0184]

[0185] in, is the model loss value, is the label value of the historical user behavior pattern and historical user preference trend information in each mapping relationship, It is the probability that the historical feature data in each mapping relationship predicted by the model carries its own label.

[0186] In one embodiment, when the processor 1001 generates the parameter value of each indicator in the pre-built multi-dimensional user stickiness tracking indicator according to the data to be analyzed, the processor 1001 specifically performs the following operations:

[0187] Determine, from the data to be analyzed, the number of user logins within a preset period, the duration of each user session, and the total number of pages viewed by the user within the preset period;

[0188] The ratio of the number of user logins to the number of days in the preset cycle is taken as the average number of daily logins; the ratio of the duration of each user session to the total number of users is taken as the average session duration; the total number of pages browsed by the user is taken as the number of page views; the average number of daily logins, average session duration and page views are taken as the parameter values ​​of the target activity indicator;

[0189] From the data to be analyzed, determine the time from customer service request to service response, the proportion of customer issues resolved, and customer feedback rating data;

[0190] Calculate the average response time from the time it takes for a customer to request a service and get a service response as the customer service response time; calculate the average score based on customer feedback rating data as the customer feedback score; take the proportion of customer problems that have been solved as the problem resolution rate; take the customer service response time, customer feedback score, and problem resolution rate as parameter values ​​for the target satisfaction indicator;

[0191] From the data to be analyzed, count the proportion of users who make repeat purchases within a preset period, the proportion of users who recommend new users, and the proportion of users who continue to use the service within a preset period;

[0192] The repeat purchase rate is the proportion of users who make repeat purchases within a preset period, the recommendation rate is the proportion of users who recommend new users, and the long-term retention rate is the proportion of users who continue to use the service within a preset period.

[0193] Repeat purchase rate, recommendation rate and long-term retention rate are used as parameter values ​​of target loyalty indicators.

[0194] In the embodiment of the present application, on the one hand, the server effectively integrates and analyzes complex data sets from different channels by using big data processing technology and machine learning algorithms, thereby deeply mining user behavior patterns and preference trends, which are then used to construct accurate multi-dimensional user stickiness tracking indicators. The multi-dimensional user stickiness tracking indicators provide a clear direction for data quantification, significantly improving the accuracy of user stickiness analysis; on the other hand, the multi-dimensional user stickiness tracking indicators can quantify the performance of user behavior, logs and business data over a certain period of time, and intuitively display the quantification results in the form of parameter values ​​through a customized dashboard. Managers can quickly grasp key performance indicators (KPIs), intuitively understand the global view and dynamic changes of user stickiness, and then quickly formulate targeted user retention strategies and product optimization measures to improve user satisfaction and loyalty.

[0195] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through computer programs. The program for analyzing user stickiness of SaaS software based on multi-source data fusion can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium of the program for analyzing user stickiness of SaaS software based on multi-source data fusion can be a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0196] The above disclosure is only the preferred embodiment of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A SaaS software user stickiness analysis method based on multi-source data fusion, characterized in that: Applied to the server, the method includes: When the time between the current moment and the end moment of the last SaaS software user stickiness analysis is equal to the preset polling time, obtain user behavior data, log information and business data collected on the platform where the SaaS software is deployed within the preset period, wherein the user behavior data is automatically collected by the preset embedding point, and the preset embedding point is an event pre-set for the user interaction function in the SaaS software for automatically collecting user behavior data; Preprocessing the user behavior data, log information, and business data to obtain data to be analyzed; According to the data to be analyzed, the parameter value of each indicator in the pre-constructed multi-dimensional user stickiness tracking indicator is generated; wherein the pre-constructed multi-dimensional user stickiness tracking indicator is constructed based on user behavior patterns and user preference trend information, and the user behavior patterns and user preference trend information are obtained by analyzing sample data using a big data processing framework and a machine learning algorithm, and the sample data is obtained by obtaining and pre-processing historical user behavior data, historical log information, and historical business data collected on a platform where SaaS software is deployed within a preset time period; wherein, The pre-built multi-dimensional user stickiness tracking indicators include target activity indicator, target satisfaction indicator and target loyalty indicator. The parameter value of each indicator is the specific value of each user stickiness tracking indicator. The parameter values ​​of the indicators are sent to the client to be displayed through a preset dashboard.

2. The method according to claim 1, characterized in that Follow these steps to generate pre-built multi-dimensional user stickiness tracking metrics, including: Obtain historical user behavior data, historical log information, and historical business data collected on the platform where the SaaS software is deployed within a preset time period; Using a big data processing framework, the historical user behavior data, historical log information, and historical business data are processed to obtain sample data; Inputting the sample data into a pre-trained SaaS software user stickiness analysis model; wherein the pre-trained SaaS software user stickiness analysis model is trained based on a machine learning algorithm; Outputting user behavior patterns and user preference trend information corresponding to the sample data; A multi-dimensional user stickiness tracking indicator is constructed according to the user behavior pattern and the user preference trend information to obtain a pre-constructed multi-dimensional user stickiness tracking indicator.

3. The method according to claim 2, characterized in that The big data processing framework includes the MapReduce programming model and Spark's distributed computing engine; The big data processing framework is used to process the historical user behavior data, historical log information, and historical business data to obtain sample data, including: Performing data cleaning and data conversion on the historical user behavior data, historical log information, and historical business data; For data cleaning and data conversion of historical user behavior data, historical log information, and historical business data, use the MapReduce programming model to perform parallel data processing; For historical user behavior data, historical log information, and historical business data that are processed in parallel, Spark's distributed computing engine is used to transform and aggregate the data to obtain multi-source fused data. Feature data associated with user stickiness analysis is extracted from the multi-source fusion data as sample data, wherein the feature data includes feature data for characterizing user activity, feature data for characterizing user satisfaction, and feature data for characterizing user loyalty.

4. The method according to claim 2, characterized in that: The multi-dimensional user stickiness tracking index is constructed according to the user behavior pattern and user preference trend information, including: Performing feature encoding on the user behavior information in the user behavior pattern to obtain user behavior features; Performing feature encoding on the user preference trend information to obtain user preference features; Using the user behavior characteristics as activity characteristics; From the user preference features, features related to user satisfaction and user loyalty are respectively divided to obtain satisfaction features and loyalty features; Obtaining a predefined initial activity indicator sequence, an initial satisfaction indicator sequence, and an initial loyalty indicator sequence; Analyzing a first correlation between the activity feature and each initial activity indicator in the initial activity indicator sequence, and taking an initial activity indicator whose first correlation is greater than or equal to a preset correlation threshold as a target activity indicator; Analyzing a second correlation between the satisfaction feature and each initial satisfaction indicator in the initial satisfaction indicator sequence, and taking an initial satisfaction indicator whose second correlation is greater than or equal to a preset correlation threshold as a target satisfaction indicator; Analyzing a third correlation between the loyalty feature and each initial loyalty indicator in the initial loyalty indicator sequence, and taking the initial loyalty indicator whose third correlation is greater than or equal to a preset correlation threshold as a target loyalty indicator; The target activity index, the target satisfaction index and the target loyalty index are used as multi-dimensional user stickiness tracking indicators; wherein the correlation is used to characterize the relationship between different features and known user stickiness.

5. The method according to claim 2, characterized in that: Follow these steps to generate a pre-trained SaaS software user stickiness analysis model, including: Obtain sample user behavior data, sample log information, and sample business data collected on the platform where the SaaS software is deployed within a preset time period; Preprocessing the sample user behavior data, sample log information, and sample business data using a big data processing framework to obtain historical multi-source fusion data; Extracting historical feature data for characterizing user activity, historical feature data for characterizing user satisfaction, and historical feature data for characterizing user loyalty from the historical multi-source fusion data; Determine historical user behavior patterns and historical user preference trend information corresponding to the historical feature data; The historical user behavior pattern and the historical user preference trend information are used as data labels of the historical feature data to perform data annotation to obtain a model training sample; A pre-trained SaaS software user stickiness analysis model is generated based on the model training samples.

6. The method according to claim 5, characterized in that The determining of the historical user behavior pattern and historical user preference trend information corresponding to the historical feature data includes: For each historical feature data, receiving a first user behavior pattern and a first user preference trend information inputted by a client as the historical user behavior pattern and the historical user preference trend information corresponding to the historical feature data; wherein the first user behavior pattern and the first user preference trend information are obtained by analyzing each historical feature data by a business expert and a data analyst; or, From the historical feature data, user behavior features, user feedback features or user comment features are obtained; a sequence pattern mining algorithm is used to mine frequent patterns of the user behavior features as historical user behavior patterns; and historical user preference trend information is generated based on the user feedback features or user comment features.

7. The method according to claim 6, characterized in that The generating historical user preference trend information according to the user feedback feature or the user comment feature includes: Performing semantic analysis on the user feedback feature or the user comment feature to extract a user preference keyword sequence corresponding to the user feedback feature or the user comment feature; Using the user's preferred keyword sequence as a demand text for guiding a preset large language model to obtain preference prompt information; The preference prompt information is input into a preset large language model, and historical user preference trend information is determined through the preset large language model.

8. The method according to claim 5, characterized in that The model training samples include the mapping relationship between the historical user behavior pattern and the historical user preference trend information and the historical feature data; The generating a pre-trained SaaS software user stickiness analysis model according to the model training sample comprises: Use decision tree algorithm to create a SaaS software user stickiness analysis model; Inputting the mapping relationship between the historical user behavior pattern and the historical user preference trend information and the historical feature data into the SaaS software user stickiness analysis model, and outputting the model loss value; When the model loss value reaches the minimum, a pre-trained SaaS software user stickiness analysis model is generated; or, when the model loss value does not reach the minimum, the model loss value is back-propagated to update the model parameters of the SaaS software user stickiness analysis model, and the step of inputting the mapping relationship into the SaaS software user stickiness analysis model is continued until the model loss value reaches the minimum; wherein the loss function of the SaaS software user stickiness analysis model is: in, is the model loss value, is the label value of the historical user behavior pattern and historical user preference trend information in each mapping relationship, It is the probability that the historical feature data in each mapping relationship predicted by the model carries its own label.

9. The method according to claim 1, characterized in that: The pre-built multi-dimensional user stickiness tracking indicators include a target activity indicator, a target satisfaction indicator, and a target loyalty indicator. The generating, based on the data to be analyzed, parameter values ​​of each indicator in the pre-built multi-dimensional user stickiness tracking indicator includes: Determine, from the data to be analyzed, the number of user logins within a preset period, the duration of each user session, and the total number of pages browsed by the user within the preset period; The ratio of the number of user logins to the number of days in the preset period is taken as the average number of daily logins; the ratio of the duration of each user session to the total number of users is taken as the average session duration; the total number of pages browsed by the user is taken as the number of page views; the average number of daily logins, the average session duration and the number of page views are taken as parameter values ​​of the target activity indicator; Determine the time from customer service request to service response, the proportion of customer problems solved, and customer feedback rating data from the data to be analyzed; Calculate the average response time based on the time from the customer requesting service to the service response as the customer service response time; calculate the average score based on the customer feedback score data as the customer feedback score; take the proportion of customer problems solved as the problem solving rate; take the customer service response time, the customer feedback score and the problem solving rate as the parameter values ​​of the target satisfaction index; From the data to be analyzed, count the proportion of users who make repeated purchases within a preset period, the proportion of users who recommend new users, and the proportion of users who continuously use the service within a preset period; The repeat purchase rate is the proportion of users who make repeated purchases within a preset period, the recommendation rate is the proportion of new users recommended by the users, and the long-term retention rate is the proportion of users who continue to use the service within a preset period. The repeat purchase rate, recommendation rate and long-term retention rate are used as parameter values ​​of the target loyalty index.

10. A SaaS software user stickiness analysis system based on multi-source data fusion implemented using the method described in any one of claims 1 to 9, characterized in that: The system comprises: A multi-source data acquisition module is used to acquire user behavior data, log information and business data collected on a platform where the SaaS software is deployed within a preset period when the time between the current moment and the end moment of the previous SaaS software user stickiness analysis is equal to a preset polling time, wherein the user behavior data is automatically collected by a preset embedding point, and the preset embedding point is an event pre-set for a user interaction function in the SaaS software for automatically collecting user behavior data; A data preprocessing module, used to preprocess the user behavior data, log information and business data to obtain data to be analyzed; A parameter value generation module, used to generate parameter values ​​of each indicator in a pre-built multi-dimensional user stickiness tracking indicator according to the data to be analyzed; wherein the pre-built multi-dimensional user stickiness tracking indicator is constructed based on user behavior patterns and user preference trend information, and the user behavior patterns and user preference trend information are obtained by analyzing sample data using a big data processing framework and a machine learning algorithm, and the sample data is obtained by acquiring and pre-processing historical user behavior data, historical log information, and historical business data collected on a platform where SaaS software is deployed within a preset time period; The parameter value display module is used to send the parameter values ​​of the indicators to the client for display through a preset dashboard.

Images