KR20220026356A - Method for providing session replay service using session information storage and rendering - Google Patents

Abstract

Provided is a method for providing a session replay service using session information storage and rendering, comprising: a step of detecting through a script when at least one preset event occurs in a user terminal accessing a webpage; a step of receiving and storing at least one event together with the document object model data of the webpage; and a step of reconstructing a screen of the user terminal into a video by performing time-series listing and restructuring rendering to at least one event on the structure information data. Therefore, the present invention is capable of providing in helping improve a user experience and improve the site.

Description

Method of providing session replay service using session information storage and rendering

The present invention relates to a method of providing a session replay service using session information storage and rendering, and provides a platform that can visualize time-series user events on structure information data of a web page and reproduce them like a video.

Mouse tracking analysis is one of the performance measurement methods, and it is a method of tracking and analyzing the movement of a mouse on a web browser along with eye tracking that tracks and analyzes the user's gaze. By examining the trajectory of the mouse used while the user accesses and navigates the web, the user's interests, movement route, and required time can be identified. After a goal is determined on the web interface, the web user defines a path to the location and tries to reach the target point by moving the cursor through device manipulation. In particular, it can be seen that the movement and manipulation of the cursor using the mouse occupies most of the practical interactions in the GUI-based web interface. As such, mouse tracking is useful for observing the movement of a web user that appears in the background of the monitor screen compared to video ethnography, which is suitable for observing user movement in the physical world through devices such as video and camcorders.

At this time, a method for analyzing user behavior by extracting session information has been researched and developed. In No. 0004207 (published on January 13, 2020), the web log is analyzed in real time to determine the web log for session processing, and the session based on the IP address and URL for the web log determined as the web log for session processing A configuration for determining the type and storing user session information according to the determined session type, and when the user terminal connects to the web server and forms a session, log data corresponding to the action performed by the user terminal within the session is recorded in real time It collects, detects log data corresponding to the trigger log from among the log data, extracts the accumulated log data from the start of the session to the detection time of the trigger log to generate the accumulated log data, and performs pattern analysis of the user terminal. Each of the analysis configurations for generating behavior information corresponding to the behavior is disclosed.

However, the above-described configuration is merely a history log that simply stores session information or collects logs up to the moment when a specific event occurs, and it is insufficient to intuitively understand exactly why users are flowing in, making purchases, or leaking using this. Do. In addition, using mouse tracking and video ethnography to record and reproduce the user's screen as it is is a violation of the Personal Information Protection Act and GDPR (General Data Protection Regulation) and is an illegal act. In addition, a number of tracking solutions are focused on tracking quantitative indicators on the web, but user experience is simply a quantitative indicator that corresponds to numbers. It's not enough, and really important information other than just numbers is not being recorded. Accordingly, research and development of a platform capable of analyzing user behavior data up to the level of qualitative data without specifying the user is required.

An embodiment of the present invention collects at least one type of event that occurs in a user's web browser using a script, collects event occurrence information as well as structure information (Document Object Model) of a web site together, All events generated by the user on the web page and the user's screen at that time are not read as it is, but the user's qualitative data is collected through reconfiguration, while at the same time complying with the Personal Information Protection Act and clustering (clustering) so that individuals are not identified or identified. ) by unsupervised learning, each user's event grouped into a cluster is statistically processed and de-identified, while providing a basis to help improve user experience and site improvement, and intuitively data on user behavior It is possible to provide a method of providing a session replay service using session information storage and rendering, which can perform rendering by listing events in a time series so as to be able to view them and reconstructing them in the form of a moving picture. However, the technical task to be achieved by the present embodiment is not limited to the technical task as described above, and other technical tasks may exist.

As a technical means for achieving the above-described technical problem, an embodiment of the present invention provides a step of detecting through a script when at least one preset event occurs in a user terminal accessing a web page, at least one event Reconstructing the screen of the user terminal into a moving image by performing the steps of receiving and storing together with the document object model data of the web page and rendering to list and restructure at least one event in time series on the structure information data includes steps.

According to any one of the above-described problem solving means of the present invention, at least one type of event occurring in a user's web browser is collected using a script, and not only event occurrence information but also structure information of a web site (Document Object Model) All events generated by each user on the web page and the user's qualitative data are collected through reconfiguration rather than reading the user's screen at that time, while at the same time complying with the Personal Information Protection Act, By performing clustering to prevent identification by unsupervised learning, the event of each user grouped into a cluster is statistically processed and de-identified while providing a basis for helping to improve user experience and improve the site, In order to intuitively view the user's behavior data, it is possible to perform rendering that lists events in a time series and reconfigures them in the form of a video, and ultimately uses the analysis of user events to create time-based web pages for UX designers. Analysis tools can be provided.

1 is a diagram for explaining a system for providing a session replay service using session information storage and rendering according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating a session replay service providing server included in the system of FIG. 1 .
3 and 4 are diagrams for explaining an embodiment in which a session replay service using session information storage and rendering according to an embodiment of the present invention is implemented.
5 is an operation flowchart illustrating a method for providing a session replay service using session information storage and rendering according to an embodiment of the present invention.
6 is a diagram for explaining an embodiment of a screen on which a session replay service is implemented according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily implement them. However, the present invention may be embodied in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is "connected" with another part, this includes not only the case of being "directly connected" but also the case of being "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated, and one or more other features However, it is to be understood that the existence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded in advance.

The terms "about", "substantially", etc. to the extent used throughout the specification are used in or close to the numerical value when manufacturing and material tolerances inherent in the stated meaning are presented, and are intended to enhance the understanding of the present invention. To help, precise or absolute figures are used to prevent unfair use by unconscionable infringers of the stated disclosure. As used throughout the specification of the present invention, the term "step of (to)" or "step of" does not mean "step for".

In this specification, a "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be implemented using two or more hardware, and two or more units may be implemented by one hardware. Meanwhile, '~ unit' is not limited to software or hardware, and '~ unit' may be configured to be in an addressable storage medium or to reproduce one or more processors. Thus, as an example, '~' denotes components such as software components, object-oriented software components, class components, and task components, and processes, functions, properties, and procedures. , subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. The functions provided in the components and '~ units' may be combined into a smaller number of components and '~ units' or further separated into additional components and '~ units'. In addition, components and '~ units' may be implemented to play one or more CPUs in a device or secure multimedia card.

Some of the operations or functions described as being performed by the terminal, apparatus, or device in the present specification may be performed instead of by a server connected to the terminal, apparatus, or device. Similarly, some of the operations or functions described as being performed by the server may also be performed in a terminal, apparatus, or device connected to the server.

In this specification, some of the operations or functions described as mapping or matching with the terminal means mapping or matching the terminal's unique number or personal identification information, which is the identification data of the terminal. can be interpreted as

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram for explaining a system for providing a session replay service using session information storage and rendering according to an embodiment of the present invention. Referring to FIG. 1 , the session replay service providing system 1 using session information storage and rendering may include at least one user terminal 100 and a session replay service providing server 300 . However, since the session replay service providing system 1 using the session information storage and rendering of FIG. 1 is only an embodiment of the present invention, the present invention is not limitedly interpreted through FIG. 1 .

At this time, each component of FIG. 1 is generally connected through a network 200 . For example, as shown in FIG. 1 , at least one user terminal 100 may be connected to the session replay service providing server 300 through the network 200 . In addition, the session replay service providing server 300 may be connected to at least one user terminal 100 through the network 200 .

Here, the network refers to a connection structure in which information exchange is possible between each node, such as a plurality of terminals and servers, and an example of such a network includes a local area network (LAN), a wide area network (WAN: Wide Area Network), the Internet (WWW: World Wide Web), wired and wireless data communication networks, telephone networks, wired and wireless television networks, and the like. Examples of wireless data communication networks include 3G, 4G, 5G, 3rd Generation Partnership Project (3GPP), 5th Generation Partnership Project (5GPP), Long Term Evolution (LTE), World Interoperability for Microwave Access (WIMAX), Wi-Fi (Wi-Fi) , Internet, LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network), PAN (Personal Area Network), RF (Radio Frequency), Bluetooth (Bluetooth) network, NFC ( Near-Field Communication) networks, satellite broadcast networks, analog broadcast networks, Digital Multimedia Broadcasting (DMB) networks, and the like are included, but are not limited thereto.

In the following, the term at least one is defined as a term including the singular and the plural, and even if the at least one term does not exist, each element may exist in the singular or plural, and may mean the singular or plural. it will be self-evident In addition, that each component is provided in singular or plural may be changed according to embodiments.

The at least one user terminal 100 may be a terminal of a user using a session replay service related web page, an app page, a program, or an application using session information storage and rendering. In this case, the at least one user terminal 100 may be a terminal that generates at least one type of event and transmits the generated event to the session replay service providing server 300 together with the structure information data of the web page.

Here, the at least one user terminal 100 may be implemented as a computer that can access a remote server or terminal through a network. Here, the computer may include, for example, navigation, a laptop equipped with a web browser, a desktop, and a laptop. In this case, the at least one user terminal 100 may be implemented as a terminal capable of accessing a remote server or terminal through a network. At least one user terminal 100 is, for example, as a wireless communication device that guarantees portability and mobility, navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS(Personal Handyphone System), PDA(Personal Digital Assistant), IMT(International Mobile Telecommunication)-2000, CDMA(Code Division Multiple Access)-2000, W-CDMA(W-Code Division Multiple Access), Wibro(Wireless Broadband Internet) ) terminal, a smart phone, a smart pad, a tablet PC, etc. may include all kinds of handheld-based wireless communication devices.

The session replay service providing server 300 may be a server that provides a session replay service web page, an app page, a program, or an application using session information storage and rendering. And, the session replay service providing server 300, when at least one type of event occurs from the at least one user terminal 100, collects and stores the structure information data of the web page together with the lightweight structure information data and It may be a server that renders the player so that it can be played based on an event. In addition, the session replay service providing server 300 may be a server for clustering at least one user terminal 100 by unsupervised learning, or a server for performing session replay for each cluster. In addition, the session replay service providing server 300 may be a server that provides a result of analyzing a user's behavior to at least one client (not shown) requesting it.

Here, the session replay service providing server 300 may be implemented as a computer capable of accessing a remote server or terminal through a network. Here, the computer may include, for example, navigation, a laptop equipped with a web browser, a desktop, and a laptop.

2 is a block diagram for explaining a session replay service providing server included in the system of FIG. 1, and FIGS. 3 and 4 are session replay services using session information storage and rendering according to an embodiment of the present invention. It is a diagram for explaining an embodiment.

Referring to FIG. 2 , the session replay service providing server 300 includes a sensing unit 310 , a storage unit 320 , a reconstruction unit 330 , a mining unit 340 , and a funnel analysis unit 350 . may include

A session replay service application, program using the session replay service providing server 300 or another server (not shown) operating in conjunction with at least one user terminal 100 to store and render session information according to an embodiment of the present invention , when transmitting an app page, a web page, etc., at least one user terminal 100 may install or open a session replay service application, program, app page, web page, etc. using session information storage and rendering. In addition, a service program may be driven in at least one user terminal 100 using a script executed in a web browser. Here, the web browser is a program that enables the use of a web (WWW: World Wide Web) service, and refers to a program that receives and displays hypertext written in HTML (Hyper Text Mark-up Language), for example, Netscape. , Explorer, Chrome, and the like. In addition, the application means an application on the terminal, for example, includes an app (App) executed in a mobile terminal (smartphone).

Before describing Fig. 2, the basic concept of the present invention will be described.

A UX designer plays a role in delivering data and hints for web page operation to users and designing a design with high production cost efficiency. For effective design, UX designers are in charge of design research and design evaluation. Unlike market research, design research should consider the users who use the service as a target and present what makes people use it, and what message and style make people happy. And design evaluation should proceed in the direction of improving the design and demonstrating the effectiveness of the design. In this way, the UX designer who proceeds from concept proposal to design evaluation not only analyzes short-term quantitative data within the page, but also analyzes the user's fluid flow over time and the causal relationship of the data in the page to improve services and prove efficiency through change data Should be. However, diachronic comparative analysis between users and factors affected by time flow is difficult only with the analysis of the public viewpoint provided by the current visualization tool. From a diachronic point of view, it is possible to analyze the overall service flow over time, and it has the advantage of being able to improve UI elements with strong seasonality and manage user segments according to the time trend. Accordingly, in an embodiment of the present invention, an in-page analysis tool capable of efficient UX analysis focusing on trend data is provided.

Referring to FIG. 2 , when at least one preset event occurs in the user terminal 100 accessing the web page, the sensing unit 310 may detect it through a script. In this case, a session means a period during which a user interacts for a certain task or collaboration. For example, after accessing the A shopping mall page, an event of making a purchase on the A shopping mall page (purchase event) It may mean a period until it occurs or a period from when it flows into the A shopping mall page and then flows out. In this case, the session recording (Session Recording) refers to storing an event generated while participating in a session, and the Session Replay (Session Replay) refers to replaying the events stored as the session recording by sequentially re-executing the event. At this time, in order to perform session replay according to an embodiment of the present invention, the time the user stayed on the web page, which page he looked at, which one he selected (clicked or touched), which was viewed for a long time, and a search term You need to collect real-time feedback or behavioral data, such as what the outcome was. Here, the script may use JavaScript or Amazon's AWS session script, but is not limited thereto.

The storage unit 320 may receive and store at least one event together with document object model data of a web page. At this time, the at least one event may include page view, click, purchase conversion, and session conversion, and the structure information data may include HTML, CSS, and JS, but is not limited to the listed ones and excluded for reasons not listed. doesn't happen HTML or XML documents are human-readable by representing spreadsheets, address books, or other structured data as logical structures in the form of text. However, the computer cannot recognize the structural contents of the document only with the text information of the document. Accordingly, the World Wide Web Consortium (W3C) recommends a programming interface document called DOM (Document Object Model, Structural Information Data), an API that defines the structural object expression style and approach by classifying elements of HTML and XML documents into independent objects. provide specifications. This means that by using the DOM, all data in the document can be loaded into memory in the form of parent and child hierarchies of nodes.

For example, when the XML document is parsed, a DOM tree structure is formed in the memory of the user terminal 100 . Objects constituting this tree structure are called nodes, and each node has a unique ID (ID) so that it can be easily accessed and edited in memory. Nodes using the DOM tree can be classified into element nodes, attribute nodes, and text nodes. For example, an element node represents an information container. This is a basic node that makes XML, and it means a node that includes other element nodes or text nodes and can have attributes uniquely. Also, the attribute node contains information about the element node. By including a number of properties, the property node accesses the properties of the created object to control each element. Also, the text node contains meaningful text information or is blank. In addition, these DOM objects and nodes and their properties can be accessed and manipulated according to the W3C DOM standard, so node creation and access such as createElement(), create TextNode(), identification of nodes such as nodeName, nodeType, tagName, childNode(), Methods such as understanding tree structure such as firstChild() are used. Table 1 below summarizes descriptions of representative DOM methods provided by the W3C DOM standard.

createElement create element node createTextNode create text node appendchild add to element node insertBefore Insert new node into old node removeChilde remove target node setAttribute change attribute value getAttribute return property value getElementById Search for element corresponding to ID getElementsByTagName Find the element corresponding to the tag and return it as an array

At this time, the information constituting the web page is defined based on XML, so a parser for XML text processing is required to reconstruct the web page. However, the XML data analyzed through the XML parser creates a DOM tree in the memory of the user terminal 100 to form a logical structure, and this information outputs a scene configured to the user through rendering. There is a disadvantage in that only one page can be expressed, and new web page configuration information must be generated every time a scene change is made by an event designated by a service provider or a user's request. For example, the service provider (web page provider) transmits initial web page configuration information composed of a web page image object having a content ID of V1 and a text object having a content ID of A1 as data in XML format, and the user terminal 100 ) analyzes the received configuration information, creates it in the form of a DOM tree in memory, and consumes the contents while structuring and decoding each element. However, if the consumer selects an icon or menu to move, switch, or change the web page, for example, when an icon image object needs to be inserted because an icon should be displayed on the web page, not only additional objects but also existing It is necessary to create a duplicate XML document up to the web page configuration information of

Accordingly, the storage unit 320 decides to use the new web page configuration information data by using the DOM Method, which is a web page dynamic change technology. First, initial web page configuration information (DOM) data is transmitted to the user terminal 100 in the form of XML, and the user terminal 100 receives the configuration information data and analyzes it through an XML file parser on the memory. Creates and consumes the screen of the initial web page by creating a DOM tree in At this time, when an event occurs in the user terminal 100 and a web page is converted, changed, and moved, it enables more efficient configuration information data change using DOM Update, and the storage unit 320 uses this By collecting data, duplicate data is not received. At this time, the service producer (web page provider) generates and transmits a DOM Update Message to the user terminal 100, and the user terminal 100 transmits the DOM update message while preserving the DOM tree in the existing memory. The DOM tree is changed by performing a partial update on a specific element by analyzing it with an XML file parser, and the storage unit 320 also partially updates the already collected configuration information data to prevent duplicate collection of structure information data. there is.

For example, when the web page the user sees is a shopping mall page, a video of a model wearing clothes is played, and when the video is clicked, the user's event (click) is extracted from the web page where detailed information about clothes is overlaid with text. Suppose we want to store configuration information data. At this time, the video and audio data decoded by the service provider are input to the scene generator, and the scene generator provides information about when and how each media will be consumed based on the input information to the scene composition information as XML documents or binary created in the form of information. Scene configuration information including information on synchronization of each video and audio is input to the packetizer and packetized, and again multiplexed with media information through a multiplexer and delivered to the user terminal 100 .

The user terminal 100 receives media information in the form of a packet from a service provider (web page provider) using a transmission network and inputs the media information to the demultiplexer, and the demultiplexer analyzes the packet, and the packet with video and audio information is Media Decoding is accomplished by sending it to the decoder, and the scene configuration information is transmitted to the scene parser. The scene parser creates a DOM tree in the memory of the user terminal 100 and delivers it to the renderer, and media consumption is made according to the playback position and timing of each video and audio according to the DOM tree. In this way, when an event of clicking a video in the user terminal 100 occurs while using the media service, that is, when a text element needs to be added to the screen, DOM update information for the text is collected from the storage unit 320 and the video + By collecting text, the structure information data of the video is not collected again, but only text is collected and added to the DOM tree and rendered.

In this way, when DOM update is used, it is possible to efficiently process data because it is possible to collect only the elements to be changed while maintaining the independence of the existing elements. In addition, since updates can be performed only with small update data, less server resources, that is, networking and computing resources, can be used, and even in complex scenes in a streaming environment where text-based scene composition information must be converted to binary and delivered, it is also useful for resource utilization. The time consumed can be greatly reduced.

The reconfiguration unit 330 may reconstruct the screen of the user terminal 100 into a moving image by performing rendering of time-series listing and restructuring of at least one event on the structure information data. In this case, the user terminal 100 is a plurality of user terminals 100 belonging to different subjects, and the video may be reconstructed for each clustered group. At this time, the average value for each cluster and the value of one user may be calculated based on the metrics for each cluster, and at least one type of unique identification information of the user terminal 100 may be de-identified. . In addition, as described above, since each node and element are defined from the DOM data, reconstructing them only needs to go through rendering, which is a visualization process. At this time, it should be assumed that at least one element constituting the structure information data has a visualization image, a frame, or a layout stored in advance, respectively. For example, if element A is extracted and listed, but there is no image of how it will be expressed, it cannot be expressed. Therefore, each element must be stored and stored in a database mapped with each image, frame or layout. [Z web page - A element - B image - C coordinate (or C frame or C layout)] is saved, and when element A is extracted, it is placed on the reconstruction screen to compose the image B, and the user By arranging events on top of it in time series, it performs rendering that "reconstructs" the user's screen rather than recording it as it is. This reconstruction process is a pre-processing and refining process for subsequent data mining, but it can also serve as a de-identification process to comply with the Personal Information Protection Act.

The National Science and Technology Knowledge Information Service (NTIS) of the Korea Institute of Science and Technology Information conducts an impact assessment on personal information protection, and assigns grades according to the degree of impact on personal information and manages them. Hereinafter, Table 2 shows the personal information impact ratings of NTIS. Level 1 is a very important personal information item that can be individually identified by itself or can be individually identified when combined.

rank classification Personal information items explanation Grade 1 unique identification information Resident registration number, passport number, driver's license number, alien registration number *Personal Information Protection Act Article 24 and Enforcement Decree Article 19 Personal information that can be identified by itself or is very sensitive, or personal information whose processing is strictly restricted in accordance with relevant laws and regulations sensitive information Information that may significantly infringe privacy, such as ideology and belief, union or party membership or withdrawal, political opinion, medical history, physical or mental disability, sexual orientation, genetic test information, criminal history information, etc.
*Personal Information Protection Act Article 23 and Enforcement Decree Article 18 authentication information Password, bio information (fingerprint, iris, vein, etc.)
* Article 7 of the Standard Notification of Measures to Secure the Stability of Personal Information credit information
financial information Credit information, credit card number, account number, etc.
* Articles 2 and 19 of the Credit Information Use and Protection Act and Articles 2, 16, and 21 of the Enforcement Decree of the same Act location information *Act on the protection and movement of location information, Article 2, Article 16, etc. 2nd grade personally identifiable information Name, address, phone number, mobile phone number, email, date of birth, gender, etc. Personal information that can be clearly identified when combined Personal information Education, occupation, height, weight, marital status, family situation, hobbies, etc. 3rd grade Auto-generated information IP information, MAC information, site visit history, cookies, etc. Indirect personal information that provides additional information when combined with personally identifiable information processing information Statistical information, subscriber propensity, etc. Limited personally identifiable information Member information, company number, personally identifiable information for internal use, etc.

At this time, it can be seen that most of the unique identification information is defined as belonging to the first grade, and other personal information is also subject to security in the second and third grades. Here, since site visit records and cookie or session information are also subject to security, reconfiguration for session replay according to an embodiment of the present invention that de-identifies them is a very important process. Of course, through this reconstruction, as described above, not raw data, but pre-processed data, so there is no need to add or tag metadata for identification based on this, and data mining, that is, No preprocessing is required in the process of clustering each user.

The mining unit 340 reconstructs the screen of the user terminal 100 into a moving image by performing rendering to time-series listing and restructuring at least one event on the structure information data in the reconstruction unit 330, then at least A video of a plurality of user terminals 100 may be data-mined using a single artificial intelligence algorithm. In this case, the user terminal 100 is a plurality of user terminals 100 belonging to different subjects. That is, instead of outputting a restructured video (including gif) by specifying one person, statistical processing is performed for a plurality of users. For example, if there are more than 100 people who have pressed [Menu A - Menu B - Buy button], group these users into one group (cluster) "A" and statisticalize the group "A" data can be output. Also, if [Menu C - Menu D - Exit] has 200 people, for example, you can group these users into one "me" group and use it to find out why customers are leaving the website. Also, if there are 150 users [F menu - G button - screen freeze], you can group this group into a "many" group and analyze whether there are errors or errors in the F menu or G button, and this You can instruct the developer to make an update that fixes it.

At this time, data mining is a classification that predicts a new data class by learning a training data set with a known class by exploring the inherent relationship between preprocessed data, or grouping data based on similarity without class information. clustering) can be performed. In addition, the at least one artificial intelligence algorithm may be an algorithm based on unsupervised learning. Clustering is a representative unsupervised learning method widely applied in research in text mining and IR theory. is divided into

At this time, two approaches can be used to perform data mining without exposing data to the outside. Since the second approach can expose the user's personal information data, only the first method will be described. The first method is a method in which a distributed database obtains a clustering result from an aggregated database by secret sharing without disclosing data collected to each server to each other, and may be as shown in FIG. 4A . Each data is provided from a horizontally distributed or vertically partitioned database composed of two or more servers, and each data is computed cooperatively to obtain only clustered results without exposing each data to the other party. In this case, the K-means algorithm is performed on the premise that K center points in the vector space can be expressed as K clusters, and in general, clusters are formed by calculating the cosine similarity of each vector. The K-means algorithm can further use various extension algorithms such as Bisecting, Antipole, and Multipole depending on the cluster generation method.

The above-described process will be described in detail as follows.

First, the process of listing each user's events time-series is carried out and converted into a form suitable for clustering in order to perform K-means. To this end, we perform a digitization work that converts each event into a vector. For quantification, morphemes are analyzed for each event, and as a result, each morpheme forms the dimension of a vector. Each vector element has a value of 1 or 0 depending on whether a corresponding morpheme is included, and a morpheme analyzer can be used in this process.

The second is to create a tag pool. A measure of distance and a comparable criterion between the time-series enumeration of each event is achieved through vector multiplication. For this, it is necessary to define a standard vector containing information on all morphemes, and with this standard vector, the tag pool extracted from the time-series listing of all events can be used. The tag pool refers to a pool that contains all tags extracted from all events without duplication.

The third is the process of creating a tag hierarchy. The tag hierarchy is created based on the previously created tag pool. A vector for generating a tag hierarchy can be assigned a vector value depending on how many times arbitrary tags are included in the same tag pool while comparing elements of the tag pool. For example, if a plurality of users generate an event through the process of A-B-C-D, a vector value is assigned according to the degree of simultaneous use of A-B-C-D. Each tag vector calculated according to this method is converted into a unit vector and then clustered through K-means, and vectors with high similarity are grouped through cluster analysis to be classified into several tag groups. And again, by repeatedly applying the above clustering operation to each group, it is possible to create a tag hierarchy for the entire tag set.

Finally, the tag vector needs to be calculated. Finally, through the process of creating the tag vector, all preparations for performing the clustering algorithm are completed. Through this process, a standard for inferring how similar the time-series arrangement of each event has to each other is prepared. In addition, a value calculated based on the hierarchical structure may be used to reflect the tag hierarchy to the tag vector. In this case, the assigned vector value may use, for example, Leacock-Chodorow Measure, but is not limited thereto. . Since Leacock-Chodorow Measure expresses lexical similarity with values ranging from 0 to infinity, it can be applied to update the weight of the tag vector.

When all preparations for clustering are completed in this way, you can start tying users through each time series listing. In this case, when K-means clustering is applied to determine the first clustering K value, the number of clustering result groups (K) has an important influence on the clustering performance. In most cases, K is set as a static variable to derive clustering results according to the number, but it is very difficult to determine a constant K value considering the direction and specificity of different individuals.

Accordingly, in the method according to an embodiment of the present invention, it is also possible to set different K values according to different tags according to events listed in time series. K-means regenerates clusters iteratively until an optimal cluster is generated. That is, once a cluster is created by K-means, the center value is recalculated for all clustered groups, and vector multiplication is sequentially performed on K clusters for all tag pools. After that, the tag pool is again assigned to the group with the largest vector product as a result of the execution, and the vector product and assignment process is performed until the last tag pool. As a result, when a new clustering result is calculated, the process of obtaining the center value is repeated to achieve an optimal clustering state. In this case, the end point is when the newly formed results do not change when compared with the previous clustering results.

The funnel analysis unit 350, after data mining the video of the plurality of user terminals 100 using at least one artificial intelligence algorithm in the mining unit 340, the identification code of the web page from which the user leaves for each group , when a user leaves for each cluster, a funnel analysis is performed to analyze the identification code of the incoming web page and the departing web page, and the rate at which the user leaves for each cluster, and the analyzed data is visualized and output and can be as shown in FIG. 4C. At this time, the word funnel means a funnel, but the meaning is different in marketing. Funnel analysis is a marketing strategy for customer analysis. In other words, it is an analytical method that numerically confirms the key steps from the influx of customers to conversion. The goal is to improve the cause of churn by dividing the customer experience process leading to conversion into stages by identifying when, where, and how they churn, and based on this, increase the conversion rate. In this process, it is also useful to improve customer experience because the stage of service vulnerability can be identified.

For example, suppose your conversion rate has decreased by 10%. First, you will find the cause of the drop. You may suspect that there are measurement errors, server failures, changes in your company's and competitors' activities, and if your conversion rate is decreasing I can inquire with each department about any changes or review negative issues both inside and outside the company. . For this reason, companies are providing services that can find the cause of customer churn. For example, by providing a visualization of the churn rate at each stage, it is possible to check the customer churn rate at each stage of the funnel at a glance. For example, if customer churn continues to occur on a specific mobile page, it is confirmed that mobile optimization has not been done, and this can reduce customer inconvenience and provide a better customer experience.

In addition, an embodiment of the present invention may use an optimization model for designing a GUI design customized to a user based on user usage data to provide a customized user experience (UX). First, since the events occurring in the user terminal 100 have already been collected, it can be used as usage data, and separately, the number of times each object (GUI) is used, clickstream data, time taken to select an object, eye tracking data, etc. can be collected, or by identifying the range and form of data that can be collected, all data that fits the format can be collected.

In addition, based on the collected usage data, users are classified into N clusters through the above-mentioned clustering, that is, clustering, and the types of collected data, the purpose to be achieved by the service provider and designer, and constraints to be considered, etc. are reflected. An optimization model can be established, and an optimal GUI design for N user clusters is derived using the optimization model. The next step is GUI reconfiguration. Based on the N optimal GUI designs derived through GUI optimization, final GUIs can be determined and reflected in the system through discussion with administrators, designers, and developers.

In addition, in an embodiment of the present invention, when a preset keyword is input by voice, events before and after the voice is input are listed in time series to report errors, errors, or user mistakes, and a method for transmitting to an administrator or designer more may be used. For example, if users do not know well due to inexperienced use, or if an error occurs or an error occurs because the UI is too complexly arranged, users sigh in frustration. to be able to be modified. If it does not operate properly or the user does not recognize it well, the user thinks that the site does not run well and leaks rapidly. You can report to a designer or admin, and cluster them to change the UI that is most problematic. Alternatively, when such a sigh or profanity is recognized, it is possible to know how to use it by outputting an animation of which menu to use and how to use it as an icon to show a tutorial.

Hereinafter, an operation process according to the configuration of the session replay service providing server of FIG. 2 will be described in detail with reference to FIGS. 3 and 4 as an example. However, it will be apparent that the embodiment is only one of various embodiments of the present invention and is not limited thereto.

Referring to FIG. 3A , the session replay service providing server 300 receives the event and structure information data received from the user terminal 100 together, and places the event on the lightweight structure information data as shown in FIG. 3B. Render it to be playable in . Since it is impossible beyond inefficiency to store a video that shows the user's screen as it is for each user or group, it is impossible to design a data structure to store the data, and to play it like a video based on the data. At this time, the important point is not to record or play a video, but to store the DOM (HTML, CSS, JS, structured website data) and make it look like a website again. By installing (<script>...<script>), we use a script installed in the browser that delivers all the events that have occurred, initializes the DOM, and delivers changes. In addition, the player can output a session replay with a technology that shows it like a video based on data stored in the snapshot server.

Also, in the clustering (segmentation) of FIG. 4A, the group is automatically grouped based on the user's behavioral characteristics, rather than by an administrator or operator dividing the group in advance. In addition to K-means, there are DBSCAN, EM (Expectation-Maxximization), etc., but it is not limited to algorithms and can be clustered by applying various techniques. The average value for each group, segment or cluster can be calculated, and since metrics for each cluster are stored, the value of one user can be judged based on the median or average value of each group. For example, “Group #1” has the highest value per user because the number of pageviews per session (pageview/session) is high and the unit price is 45,000 won (or not necessarily purchase, but the number of applications and reservations). may be high). It is possible to calculate and provide the behavior F funnel data for each group, as shown in FIG. 4C . For each cluster, it provides a visual representation of the funnel, i.e., which page they leave from, where they come from and where they leave, and what the percentage is.

In addition to the visualization of FIGS. 4B and 4C , a user experience heat map that visualizes data based on the analysis page screen and Reporting Heatmaps that summarize heat map results and issues can be used. In addition, user analysis (User Analytics), a user index chart board for each page, funnel comparison (Comparing Referrers) that can compare the user experience status by source, and CTA classification (Segmenting CTA) that can group user information by major elements , Activity Stream that simulates the overall user browsing status, A/B test that can measure UX improvement performance, Funnels that identify the exit status by stage, It may further include a user session report (User Session Report) and an attention graph (Attention Graph) capable of grasping the degree and distribution of user interest.

The details that have not been described with respect to the method of providing the session replay service using the session information storage and rendering of FIGS. 2 to 4 are described above with respect to the session replay service providing method using the session information storage and rendering through FIG. 1 , and Since it can be easily inferred from the same or described content, the following description will be omitted.

5 is a diagram illustrating a process in which data is transmitted/received between components included in the system for providing a session replay service using session information storage and rendering of FIG. 1 according to an embodiment of the present invention. Hereinafter, an example of a process in which data is transmitted and received between each component will be described with reference to FIG. 5, but the present application is not limited to such an embodiment, and the example shown in FIG. 5 according to various embodiments described above will be described. It is apparent to those skilled in the art that the data transmission/reception process may be changed.

Referring to FIG. 5 , the session replay service providing server detects through a script when at least one preset event occurs in the user terminal accessing the web page (S5100).

In addition, the session replay service providing server receives and stores at least one event together with the document object model data of the web page (S5200), and lists at least one event on the structure information data in time series and By performing restructuring rendering, the screen of the user terminal is reconstructed into a moving picture (S5300).

The order between the above-described steps ( S5100 to S5300 ) is merely an example and is not limited thereto. That is, the order between the above-described steps ( S5100 to S5300 ) may be mutually changed, and some of these steps may be simultaneously executed or deleted.

As described above, the method for providing the session replay service using the session information storage and rendering of FIG. 5 is described above with reference to FIGS. 1 to 4 and the session replay service providing method using the session information storage and rendering. Since it can be easily inferred from the same or described content, the following description will be omitted.

6 is a diagram for explaining an embodiment of a screen on which a session replay service is implemented according to an embodiment of the present invention. Referring to FIG. 6A , a layout of a screen displaying information about automatically classified clusters, the number of index sessions/expected revenue for each cluster, and the ratio of each cluster session to the total sessions may be displayed at corresponding positions, respectively. Referring to FIG. 6B , customers may be classified into loyal customers, customers to be most focused on, and the like, and the number of sessions, average session time, expected revenue, and session ratio for each customer may be output as numbers and visualization graphs.

Referring to FIG. 6C , the number of incoming sessions, the funnel for each page, detailed information for each page, etc. can be visualized and output, and a cluster summary and a list of each user can be displayed. Referring to FIG. 6D , the ratio of moving from each page to the shopping cart and the ratio of customers who even placed an order may be displayed using numbers and figures. Referring to FIG. 6E , user information and user behavior data listed in chronological order by page for each session may be output, and when the replay button is pressed, the user's behavior may be output as described above. As shown in FIG. 6F , the number of sessions, the session time, and a replay button are provided, respectively, so that data analyzed for each user can be output. 6G shows that after the replay button is selected, the description, title, and URL of the corresponding page, as well as the user's behavior, can be guessed and provided as an image. Behavior data for each page may be described in sentences, or cluster information or user information may be output on the right side.

As shown in FIG. 6H , when the user's session is replayed, the user's actions on the screen can be viewed as the movement flow of the mouse cursor. In this case, the user can see that after viewing the shopping cart page, the user moves the mouse to the cancel button as shown in FIG. 6i, and then moves back to the payment button as shown in FIG. 6j to make the payment.

The method for providing a session replay service using session information storage and rendering according to an embodiment described with reference to FIG. 5 is in the form of a recording medium including instructions executable by a computer, such as an application or program module executed by a computer. can also be implemented. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer-readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The method of providing a session replay service using session information storage and rendering according to an embodiment of the present invention described above includes an application basically installed in a terminal (which may include a program included in a platform or operating system basically installed in the terminal) ), and may be executed by an application (ie, a program) installed directly on the master terminal by a user through an application providing server such as an application store server, an application, or a web server related to the corresponding service. In this sense, the method for providing a session replay service using session information storage and rendering according to an embodiment of the present invention described above is implemented as an application (ie, program) installed by default in a terminal or directly installed by a user, and installed in a terminal, etc. It can be recorded on a computer-readable recording medium of

The above description of the present invention is for illustration, and those of ordinary skill in the art to which the present invention pertains can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a dispersed form, and likewise components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention. do.

Claims (10)

A method for providing a session replay service executed in a session replay service providing server, the method comprising:
Detecting through a script when at least one preset event occurs in the user terminal accessing the web page;
receiving and storing the at least one event together with document object model data of the web page; and
reconstructing the screen of the user terminal into a moving image by performing rendering of time-series listing and restructuring of the at least one event on the structure information data;
A method of providing a session replay service using session information storage and rendering comprising a.
The method of claim 1,
The at least one event includes a page view, a click, a purchase conversion, and a session conversion,
The method of providing a session replay service using the session information storage and rendering, wherein the structure information data includes HTML, CSS and JS.
The method of claim 1,
The user terminal is a plurality of user terminals belonging to different subjects,
After the step of reconstructing the screen of the user terminal into a moving image by performing rendering of time-series listing and restructuring of the at least one event on the structure information data,
data mining the videos of the plurality of user terminals using at least one artificial intelligence algorithm;
The method of providing a session replay service using session information storage and rendering further comprising a.
4. The method of claim 3,
The data mining is a classification that predicts a class of new data by learning a training data set with a known class by exploring the inherent relationship between the preprocessed data, or grouping data based on similarity without class information (Classification) Clustering), a method of providing a session replay service using session information storage and rendering, comprising performing.
4. The method of claim 3,
The at least one artificial intelligence algorithm is an algorithm based on unsupervised learning, a session replay service providing method using session information storage and rendering.
5. The method of claim 4,
After data mining the video of the plurality of user terminals using the at least one artificial intelligence algorithm,
Funnel analysis that analyzes the identification code of the web page from which the user departs for each cluster, the identification code of the web page that is introduced and the web page that leaves when the user leaves for each cluster, and the rate at which the user leaves for each cluster ( performing funnel analysis), and visualizing and outputting the analyzed data;
The method of providing a session replay service using session information storage and rendering further comprising a.
The method of claim 1,
The user terminal is a plurality of user terminals belonging to different subjects,
The method of providing a session replay service using session information storage and rendering, wherein the video is reconstructed for each clustered group.
8. The method of claim 7,
A method for providing a session replay service using session information storage and rendering, wherein the average value for each cluster and the value of one user are calculated based on the metrics for each cluster.
The method of claim 1,
At least one type of unique identification information of the user terminal is to be de-identified, a session replay service providing method using session information storage and rendering.
A computer-readable recording medium in which a program for executing the method of any one of claims 1 to 9 is recorded.

Images