Information processing system, information processing method, and computer readable medium

Abstract

Disclosed are an information processing system, an information processing method, and a computer readable medium, the information processing system having one or more processors that acquire behavior histories of a plurality of users in a website, accept settings of specific behaviors, i.e., transformations, performed by the users in the website, and determine whether the specific behaviors are transformations or not in the website. A second behavior is extracted from a first behavior, which is a behavior constituting the acquired behavior history, and information on whether or not the received conversion is included in the behavior history, and a relationship between a change in a numerical value relating to the extracted second behavior and a change in a numerical value relating to the conversion is simulated. And outputting the simulation result.

Description

Technical Field

[0001] The present invention relates to an information processing system, an information processing method and a computer readable medium. Background Art

[0002] Nowadays, companies use web pages to provide advertisements and other information or services. Therefore, each company analyzes historical data related to web page browsing by users and effectively uses the analysis results when updating the web page. (For example, Japanese Patent No. 6866069 can be cited.) Summary of the invention

[0003] An object of the present invention is to provide a system that enables a user to efficiently improve a website, compared to a system that outputs only user behavior estimated to contribute to conversion.

[0004] According to a first aspect of the present invention, there is provided an information processing system having one or more processors, wherein the one or more processors obtain behavior histories of multiple users in a website, accept settings of specific behaviors, i.e., conversions, performed by the users in the website, extract a second behavior from the first behavior based on information on behaviors constituting the obtained behavior history, i.e., a first behavior, and whether the accepted conversion is included in the behavior history, simulate the relationship between changes in a numerical value associated with the extracted second behavior and changes in a numerical value associated with the conversion, and output the simulation results.

[0005] According to a second aspect of the present invention, in the information processing system of the first aspect, when accepting the specification of the second behavior, the one or more processors output a result of the simulation regarding the accepted second behavior.

[0006] According to the third aspect of the present invention, in the information processing system of the first or second aspect, when the one or more processors accept the designation of the change in the numerical value associated with the second behavior, they output the simulation result regarding the change in the numerical value associated with the accepted second behavior.

[0007] According to the fourth aspect of the present invention, in the information processing system of any one of the first to third aspects, when the one or more processors accept the specification of the change in the numerical value associated with the conversion, they output the change in the numerical value associated with the second behavior to achieve the change as the result of the simulation.

[0008] According to the fifth aspect of the present invention, in the information processing system of any one of the first to fourth aspects, the one or more processors extract one or more features of the web pages contained in the website based on the results of the simulation, and output the relationship between the extracted features and the changes in numerical values ​​related to the conversion.

[0009] According to a sixth aspect of the present invention, in the information processing system of the fifth aspect, the one or more processors output the relationship between the changes in one or more features of the web page and the changes occurring in the conversion based on weighted operations corresponding to the features.

[0010] According to the seventh aspect of the present invention, in the information processing system of any one of the first to sixth aspects, the one or more processors extract the second behavior based on a learning model, and the learning model takes the behavior history as input and outputs the possibility of a hypothetical user performing a behavior set as a conversion based on the input behavior history.

[0011] According to the eighth aspect of the present invention, in the information processing system of the seventh aspect, the one or more processors update each weight set for the first behavior through learning during the calculation process performed by the learning model, and extract the second behavior based on the updated weight.

[0012] According to the ninth aspect of the present invention, in the information processing system of any one of the first to eighth aspects, the one or more processors assign a first behavior history as input to a learning model that takes the behavior history as input and outputs whether a hypothetical user has performed a behavior set as a conversion based on the input behavior history, and obtains a first result related to the conversion based on the assigned first behavior history, assigns a second behavior history generated based on the first behavior history as input to the learning model, and obtains a second result related to the conversion based on the assigned second behavior history, and simulates the relationship between the change in the numerical value related to the second behavior and the change in the numerical value related to the conversion based on the first result and the second result.

[0013] According to the tenth aspect of the present invention, there is provided a computer-readable medium storing a program causing a computer to execute processing, the processing comprising: a process of acquiring behavior histories of multiple users in a website; a process of accepting settings for specific behaviors, i.e., conversions, performed by the users in the website; a process of extracting a second behavior from the first behavior based on information on behaviors constituting the acquired behavior history, i.e., the first behavior, and whether the accepted conversion is included in the behavior history; a process of simulating the relationship between changes in a numerical value associated with the extracted second behavior and changes in a numerical value associated with the conversion; and a process of outputting the result of the simulation.

[0014] According to the eleventh aspect of the present invention, there is provided an information processing method, comprising: a process of acquiring behavior histories of a plurality of users in a website; a process of accepting settings of specific behaviors, i.e., conversions, performed by the users in the website; a process of extracting a second behavior from the first behavior based on information on behaviors constituting the acquired behavior history, i.e., the first behavior, and whether the accepted conversion is included in the behavior history; a process of simulating the relationship between changes in a numerical value associated with the extracted second behavior and changes in a numerical value associated with the conversion; and a process of outputting the result of the simulation.

[0015] (Effect)

[0016] According to the first aspect, it is possible to provide a system that allows a user to efficiently improve a website, compared with a system that outputs only the behavior of the user estimated to contribute to conversion.

[0017] According to the second aspect, the influence of accepting the designated second behavior on the conversion can be output as a numerical value.

[0018] According to the third aspect, the influence of accepting the designated second behavior on the conversion can be output as a numerical value.

[0019] According to the fourth aspect, a change in a numerical value associated with a second behavior can be output, and the second behavior implements a change in a numerical value designated in connection with a conversion acceptance.

[0020] According to the fifth aspect, it is possible to output a relationship between a change in a numerical value designated by a conversion acceptance and a change in a numerical value related to a feature of a web page.

[0021] According to the sixth aspect, the relationship between changes in the characteristics of the web page and the changes occurring during the conversion can be output.

[0022] According to the seventh aspect, even when the number of action histories to be processed is large, it is possible to extract the second action that affects the conversion.

[0023] According to the eighth aspect, even when the number of action histories to be processed is large, the accuracy of extraction of the second action that affects conversion can be improved.

[0024] According to the ninth aspect, the time required for outputting simulation results can be shortened.

[0025] According to the tenth aspect, it is possible to provide a system that allows a user to efficiently improve a website, compared to a system that outputs only the behavior of a user estimated to contribute to conversion.

[0026] According to the eleventh aspect, it is possible to provide a system that allows a user to efficiently improve a website, compared to a system that outputs only the behavior of a user estimated to contribute to conversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A diagram for explaining a configuration example of an information processing system assumed in the first embodiment;

[0028] Figure 2 This is a diagram for explaining an example of the hardware configuration of a business support server;

[0029] Figure 3 This is a diagram for explaining an example of the hardware configuration of an information terminal;

[0030] Figure 4 A diagram that illustrates information related to conversion of a web page;

[0031] Figure 5 is a flowchart illustrating an example of a processing action performed by a business support server;

[0032] Figure 6 This is a diagram illustrating an example of a browsing user's behavior history;

[0033] Figure 7 This is a diagram illustrating an example of the behavior history after grouping;

[0034] Figure 8 This is a diagram for explaining an example of a screen displaying the top N behaviors that have the highest contribution to conversion;

[0035] Fig. 9 A diagram for explaining an example of a screen for inputting simulation conditions and displaying simulation results;

[0036] Fig.10 A diagram for explaining an example of a method for generating pseudo improvement data used in a simulation of the effect of a change in a behavior-related numerical value on a conversion-related numerical value;

[0037] Fig.11 A diagram for explaining another example of a method for generating pseudo improvement data used in simulating the effect of a change in a behavior-related numerical value on a conversion-related numerical value;

[0038] Fig.12 A diagram for explaining the continuation of another example of a method for generating pseudo improvement data used in simulating the influence of a change in a behavior-related numerical value on a conversion-related numerical value;

[0039] Fig.13A diagram for explaining processing actions related to prediction of improvement of conversion;

[0040] Fig.14 A diagram for explaining an example of a screen for inputting simulation conditions and displaying simulation results;

[0041] Fig.15 A diagram for explaining an example of a screen for inputting simulation conditions and displaying simulation results;

[0042] Fig.16 A diagram for explaining changes in screen display when modal information is added to simulation conditions;

[0043] Fig.17 It is a diagram that illustrates the process of machine learning for predictive models;

[0044] Fig.18 This is a diagram for explaining the extraction process of the top N behaviors;

[0045] Fig.19 A diagram illustrating an example of a simulation of a transformation using a prediction model. DETAILED DESCRIPTION

[0046] Embodiments of the present invention will be described below with reference to the accompanying drawings.

[0047] <Glossary>

[0048] In this specification, "conversion" is used to mean a specific action performed by a user on a website. Examples of specific actions include membership registration, e-magazine registration, product purchase, inquiry, download, application, reservation, and redirection to a specific page.

[0049] In the conversion, not only may a specific behavior be achieved without leaving the target site, but also a specific behavior may be achieved by temporarily leaving the target site and then returning to the target site.

[0050] In addition, conversion may also include behaviors before reaching the download, application, reservation, etc., such as behaviors that have not reached the download but have reached the download page, behaviors that have not reached the application or reservation but have reached the application page or reservation page, and behaviors that have reached the confirmation screen for the input content.

[0051] In the following embodiments, the specific behavior is described as a jump to a specific web page.

[0052] <Implementation method 1>

[0053] <System Structure>

[0054] Figure 1This is a diagram for explaining a configuration example of the information processing system 1 assumed in the first embodiment. Figure 1 The information processing system 1 shown is composed of a web server 10 , a server 20 that provides services for supporting the improvement of conversion to a web page (hereinafter referred to as a “business support server”), an information terminal 30 , and a network N.

[0055] The web server 10 is a computer that displays a web page that has been accessed on the web browser in response to a request from the web browser as a client. Figure 1 In the embodiment, the provided web pages are marked as WP1, WP2, ... In the present embodiment, a collection of a plurality of web pages is referred to as a website. The web page that serves as the entrance to the website is referred to as a homepage.

[0056] exist Figure 1 In the case of, the web browser is installed in at least one of the information terminals 30. The information terminal 30 in which the web browser is installed is referred to as a browsing terminal. In the present embodiment, a user who operates the browsing terminal is referred to as a browsing user.

[0057] The web server 10 records a history of web page browsing by a browsing user (hereinafter also referred to as “behavior history”). The web server 10 is an example of a terminal connected to the network N.

[0058] The business support server 20 is a computer that provides the user with the relationship between the change in numerical value of the transition to the web page specified by the engineer or the like and the change in numerical value related to the behavior of the browsing user.

[0059] In the present embodiment, the term "conversion improvement" is used to mean an increase in conversion.

[0060] exist Figure 1 In this case, the business support server 20 is operated by at least one of the information terminals 30.

[0061] In this embodiment, the information terminal 30 for accessing the business support server 20 is referred to as a business terminal. As a user operating the business terminal, an engineer who analyzes behavior that contributes to conversion improvement is assumed. In this embodiment, the engineer is referred to as a business user.

[0062] The business support server 20 is also an example of a terminal connected to the network N.

[0063] The information terminal 30 is, for example, a desktop computer, a notebook computer, a tablet computer, or a smartphone.

[0064] The information terminal 30 that operates as a client terminal operated by a browsing user and the information terminal 30 that operates as a user terminal operated by a business user may be one or more. The information terminal 30 is also an example of a terminal connected to the network N.

[0065] The network N is, for example, the Internet, wireless LAN (=Local area network), 4G, 5G, or other mobile communication systems.

[0066] exist Figure 1 In the example, the web server 10, the business support server 20, and the information terminal 30 are connected to one network N, but the network N for browsing the web page and the network N for accessing the business support server 20 may be different.

[0067] <Server Hardware Structure>

[0068] Figure 2 1 is a diagram for explaining an example of the hardware configuration of the business support server 20. The basic configuration of the hardware configuration of the network server 10 is also common to the business support server 20.

[0069] Figure 2 The business support server 20 shown has a processor 21, a ROM (=Read Only Memory) 22 storing BIOS (=Basic Input Output System) and the like, a RAM (=Random Access Memory) 23 used as a work area of ​​the processor 21, an auxiliary storage device 24, and a communication interface 25. Each device is connected via a bus or other signal lines 26.

[0070] The processor 21 is a device that realizes various functions by executing programs.

[0071] The processor 21 , the ROM 22 , and the RAM 23 function as a computer.

[0072] The auxiliary storage device 24 is constituted by, for example, a hard disk device or a semiconductor memory. Programs and various data are stored in the auxiliary storage device 24. The program here is used as a general term for OS (=Operating System) or application programs.

[0073] In the case of the business support server 20, there is a program for performing support as one of the application programs.

[0074] The communication interface 25 is an interface for communicating with the information terminal 30 that accesses information supporting the service through the network N. The communication interface 25 supports various communication standards. Examples of the communication standards include Ethernet (registered trademark), Wi-Fi (registered trademark), and mobile communication systems.

[0075] <Hardware structure of information terminal>

[0076] Figure 3 This is a diagram for explaining an example of the hardware configuration of the information terminal 30 .

[0077] Figure 3 The hardware structure of the information terminal 30 shown is similar to that of the business support server 20 (see Figure 2 ) are roughly the same.

[0078] Specifically, the information terminal 30 includes a processor 31, a ROM 32 storing BIOS, etc., a RAM 33 used as a work area of ​​the processor 31, an auxiliary storage device 34, a communication interface 35, a display 36, and an input receiving device 37. Each device is connected via a bus or other signal lines 38.

[0079] The display 36 is, for example, a liquid crystal display or an organic EL (=Electro Luminescence) display.

[0080] The input receiving device 37 is, for example, a mouse or a keyboard. Incidentally, when the information terminal 30 is a notebook computer, a tablet computer, or a smart phone, an electrostatic capacitive touch sensor having a transmittance that does not hinder the visibility of an image displayed on the display 36 is used as the input receiving device 37. A device combining such a touch sensor and a display is called a touch panel.

[0081] <Information related to web page conversion>

[0082] Figure 4 This diagram explains information related to web page conversion. Figure 4 , label the conversions "CV".

[0083] like Figure 4 As shown, transformation is related to the combination of "modality" and "behavior". Figure 4 In the example, the symbol X represents the combination of information.

[0084] "Modality" is information that gives characteristics to a web page, and is composed of a combination of page structure, content, and design, for example.

[0085] exist Figure 4In the example of page structure, "links" and "banners" are shown. Links are information that connects web pages and are also called "hyperlinks". Banners are information that introduces other websites or other web pages within the same website. Many banners are accompanied by links to the websites or web pages they introduce.

[0086] exist Figure 4 In the example, "document" and "image" are shown as contents. A document is composed of text and a style sheet written in HTML or the like. An image includes a still image and a moving image.

[0087] exist Figure 4 In FIG. 1 , “color” and “layout” are exemplified as designs.

[0088] "Behavior" is information related to the actions of users browsing the web pages. Figure 4 In the description, jump, stay time, reference source, time period, search engine, reference type, browser, device, and region are illustrated as examples of behaviors.

[0089] There are page jumps and page jumps in jumps. Page jumps are information indicating a jump to a specific web page. It does not necessarily need to be the web page that is the jump source. Page jumps are information indicating a jump from another web page to a specific web page. In addition to the case where the page that is the jump source and the web page that is the jump target are specified, page jumps can also be information indicating a jump between three or more web pages.

[0090] Dwell time is the time spent on a specific web page. In other words, it is the time a user spends browsing a specific web page.

[0091] The reference source is information indicating the web page of the link source when jumping to a specific web page by clicking a link.

[0092] The time period is information indicating a time period during which a specific web page is displayed.

[0093] The search engine is information indicating the name of a search engine used to refer to a specific web page.

[0094] The reference type is information indicating the type of method used to refer to a specific web page, such as a search engine, direct input of a URL, or clicking of a bookmark.

[0095] The browser is information indicating a web browser used to display a specific web page.

[0096] The device is information indicating the type of the information terminal 30 used by the user who refers to the web page.

[0097] The region is information indicating the location of the information terminal 30 where the web browser is running. The location here is not limited to the location located by the positioning system, but may be an estimated location. Furthermore, the location is not limited to the physical location, but may be a logical location on the network.

[0098] In addition, when a database that manages information about users who operate information terminals 30 for browsing web pages (hereinafter referred to as "user information") is accessible, user information such as the user's age, gender, and address may also be included as information related to conversion to a specific web page.

[0099] Even if it is only the combination of the exemplified information, the amount is enormous, and therefore it is not easy to estimate the behavior highly correlated with the conversion to a specific web page based on the experience and skills of engineers and the like.

[0100] <Processing actions of the business support server>

[0101] Figure 5 This is a flowchart for explaining an example of a processing action performed by the business support server 20. The symbol S shown in the figure means "step". Figure 5 The processing actions shown are performed by the processor 21 (refer to Figure 2 ) is implemented by executing the program.

[0102] In the case of this embodiment, Figure 5 The processing operation shown is started by an engineer or the like accessing the business support server 20 or the like.

[0103] <Step 1>

[0104] First, the processor 21 obtains the browsing user's behavior history from the network server 10 (step 1). The browsing user's behavior history is stored in the network server 10 as a so-called access log (see Figure 1 ), and also copied to the auxiliary storage device 24 of the business support server 20 (refer to Figure 2 )middle.

[0105] Figure 6 This is a diagram for explaining an example of the behavior history of a browsing user.

[0106] exist Figure 6 In the case of the above, the behaviors constituting the behavior history include, for example, browsing date and time 24B, entry page 24C, reference type 24D, search engine 24E, exit page 24F, previous page 24G, object page 24H, next page 24I, number of clicks 24J, browser 24K, and number of views 24L.

[0107] exist Figure 6In the case of , each row corresponds to the behavior history of each browsing performed by the browsing user 24A (for example, A, B, C, ...). Therefore, the behavior history based on the same browsing user appears in multiple rows. By the way, each column is the behavior that constitutes the behavior history. Figure 6 In order to illustrate the corresponding relationship, the browsing users corresponding to the behavior history are recorded. The same is true in other figures.

[0108] The browsing user 24A is information for identifying a user who browses a specific web page. For example, the browsing user 24A stores information on the information terminal 30 (see Figure 1 )’s IP address.

[0109] The browsing date and time 24B stores information on the date and time when a specific web page was browsed.

[0110] The portal page 24C stores information on the web page that is first accessed in a series of page jumps performed by the browsing user.

[0111] The exit page 24F stores information on the web page that was last visited in a series of page jumps performed by the browsing user.

[0112] The previous page 24G stores information indicating the web page browsed just before jumping to the specific web page.

[0113] The target page 24H stores information indicating a specific web page.

[0114] The next page 24I stores information indicating a web page that is a target for jumping from a specific web page.

[0115] The number of clicks 24J stores information indicating the number of clicks by the browsing user in a series of page jumps.

[0116] The browser 24K stores information indicating the name of the browser used by the browsing user for web browsing.

[0117] Information indicating the number of web page views performed by the browsing user is stored in the number of views 24L.

[0118] <Steps 2 and 3>

[0119] Next, the processor 21 receives the setting of the web page to be converted (step 2). Specifically, the processor 21 receives the designation of the web page to be converted from the information terminal 30 operated by an engineer or the like.

[0120] Next, the processor 21 performs grouping of behaviors related to the web page as the conversion target (step 3). In the grouping, the behaviors related to the web page as the conversion target are integrated for each browsing user.

[0121] Figure 7 This is a diagram illustrating an example of grouped behavior history. Figure 7 In, with Figure 6 The corresponding parts are marked with corresponding symbols.

[0122] Figure 7 Each behavior in the behavior history shown is a summary of behaviors related to the web page that is the conversion target. Figure 7 In the case where the behaviors constituting the behavior history are shown, for example, browsing users 24A, number of page jumps (or rate) 24M, number of page inflows (or rate) 24N, number of page jumps (or rate) 24O, number of reference sources (or rate) 24P, and number of session time periods (or rate) 24Q.

[0123] The number (or rate) of page jumps 24M is information on the number of jumps from web page A to web page B per browsing user or the jump rate obtained by dividing the number of jumps by the number of sessions.

[0124] The page inflow number (or rate) 24N is information on the inflow number to the web page A per browsing user or the bounce rate obtained by dividing the inflow number by the number of sessions.

[0125] The page jump number (or rate) 240 is information on the number of jumps to web page A by each browsing user or the jump rate obtained by dividing the number of jumps by the number of sessions.

[0126] The number (or rate) of reference sources 24P is information on the number of inflows from the web page A that is the reference source for each browsing user, or the inflow rate obtained by dividing the number of inflows by the number of sessions.

[0127] The session time slot number (or rate) 24Q is information on the time slot number of sessions of each browsing user or a time slot rate obtained by dividing the time slot number by the number of sessions.

[0128] <Step 4>

[0129] Next, the processor 21 extracts the top N behaviors with the highest contribution to conversion (step 4). The contribution is a numerical value indicating the possibility that a specific behavior contributes to conversion.

[0130] In the present embodiment, the processors 21 arrange the processors in the order of contribution. Figure 7 For each behavior (ie, each column) shown, N behaviors are extracted from the uppermost position.

[0131] The number of upper N here may be a fixed value or may be specified by an engineer or the like. When it can be specified by an engineer or the like, the display 36 (see Figure 3 ) displays a field for specifying the quantity to be extracted.

[0132] The top N behaviors are examples of second behaviors extracted from the first behaviors.

[0133] In the case of this embodiment, each time a new behavior history is obtained, the contribution value is updated according to the following rules. When a specific web page has been reached after one or more screen jumps, the contribution value of the related first behavior is increased by, for example, "1", and when the specific web page has not been reached, the contribution value of the related first behavior is reduced by, for example, "1".

[0134] As a result, the contribution of a specific behavior that is likely to contribute to conversion increases, while the contribution of a behavior that is less likely to contribute to conversion decreases. For example, the contribution of behavior Y that appears ten times increases compared to behavior X that appears only once in the history of jumps to web page A.

[0135] In this way, the top N behaviors are extracted based on the behavior (first behavior) constituting the behavior history and the information on whether the conversion is included in the behavior history.

[0136] <Steps 5, 6, 7>

[0137] Next, the processor 21 displays the extracted top N behaviors (step 5). Specifically, the top N behaviors are displayed on the display 36 of the information terminal 30 operated by the engineer or the like. Through this display, the engineer or the like can confirm the behaviors that contribute most to the browsing of the web page that is the conversion target. In the case of this embodiment, the contribution of each behavior is also displayed.

[0138] Next, the processor 21 simulates the relationship between the change of the value associated with the specific behavior and the change of the value associated with the conversion (step 6).

[0139] In the present embodiment, the change in the converted numerical value when the numerical value related to the specific behavior changes is simulated. For example, the change amount of the converted numerical value expected when the numerical value related to the specific behavior increases by 10% is calculated.

[0140] When the simulation is completed, the processor 21 displays the simulation result (step 7). Specifically, the amount of change in the numerical value associated with the specific behavior and the amount of change in the numerical value occurring in the transition are displayed on the display 36 of the information terminal 30 operated by the engineer or the like.

[0141] <Specific example of processing action>

[0142] Hereinafter, a specific example of the processing operation executed by the business support server 20 will be described.

[0143] <Display screen of top N behaviors>

[0144] Figure 8This is a diagram for explaining an example of a screen 300 that displays the top N behaviors that have the highest contribution to conversion. Figure 8 The screen 300 shown is displayed on the information terminal 30 (see Figure 1 ) of the display 36 (refer to Figure 3 )superior.

[0145] Figure 8 The screen 300 shown is composed of a conversion input field 301, a display number input field 302, a behavior designation field 303 to be used as a simulation object, an improvement value input field 304 required for the designated behavior, a display field 305 for the top N behaviors, and a total number of browsing users field 306. Of course, the contents and numbers of items displayed are only examples.

[0146] exist Figure 8 In the case of , "visit page XX" is displayed in the conversion input field 301. Page XX refers to web page XX. For example, it is "product list".

[0147] Incidentally, a list of one or more web pages constituting a website is prepared in advance in the input field 301. When the pull-down menu is opened, the list of one or more web pages is displayed.

[0148] The input field 302 for the number of displays is used to specify the number of behaviors to be displayed as behaviors with high contribution to conversion. The value here is the same as that in step 4 (see Figure 5 ) or step 5 (refer to Figure 5 ) in N corresponding. Figure 8 In this case, enter "12". Figure 8 The "12" in is the initial value. Therefore, engineers can change the displayed number.

[0149] exist Figure 8 In this case, the value of the input field 302 can be selected by opening a pull-down menu.

[0150] In the behavior specification column 303, one or more behaviors used in the simulation are specified. Figure 8 In the case of , a total of seven check boxes are displayed, corresponding to "All", "Jump", "Reference Source", "Inflow", "Time Period", "Jump Between Pages", and "Detach". Because they are check boxes, you can not only select one, but also select multiple. In addition, when "All" is selected, the other six actions become selected. Of course, because Figure 8 The time point is just after the top N behaviors are displayed, so any check box is empty.

[0151] The improvement value input field 304 is used to input the improvement degree of the numerical requirement related to the selected behavior. Figure 8The time point is just after the top N behaviors are displayed, so the numerical value indicating the improvement degree is not displayed.

[0152] In the display column 305 of the top N behaviors, the Figure 5 ) extracted from the top N (in Figure 8 In addition, in the right column of the top N behaviors, the contribution is shown by the length of the bar graph.

[0153] exist Figure 8 In this case, it can be seen that the contribution of "Number of Jumps_Page A" is much greater than the other eleven.

[0154] The total number of browsing users column 306 indicates that a total of 200,000 browsing users were used in the analysis, and the target period was five months from “20220801” to “20221231”.

[0155] <Simulation result display screen>

[0156] Fig. 9 3 is a diagram for explaining an example of a screen 310 for inputting simulation conditions and displaying simulation results. Fig. 9 In, with Figure 8 The corresponding parts are marked with corresponding symbols.

[0157] Fig. 9 The screen 310 shown is also displayed on the information terminal 30 (see Figure 1 ) of the display 36 (refer to Figure 3 )superior.

[0158] exist Fig. 9 In the case of , "10" is input in the input field 302 of the display number. Therefore, ten behaviors reflecting the simulation results are displayed in the display field 305 of the top N behaviors.

[0159] In addition, Fig. 9 In the case of , in the behavior designation column 303, "jump", "inflow" and "jump between pages" are selected, and in the improvement value input column 304, "10%" is input as the numerical value to be given to the improvement degree.

[0160] As a result, in the display column 305 of the top N behaviors, the predicted value of the conversion improvement number predicted for each behavior is displayed in a bar graph under the title "Prediction of the improvement number of CV when the value of each behavior is improved by 10%".

[0161] Depend on Fig. 9 It can be seen from the display that when the "number of page jumps" from web page C to web page A increases by 10%, the conversion to web page XX increases by more than 35 points.

[0162] By displaying the display field 305 of the screen 310 , engineers and the like can easily confirm not only the behaviors of browsing users estimated to contribute to conversion, but also the differences in the degree to which each behavior contributes to the improvement of conversion.

[0163] As a result, compared with a system that outputs only user behaviors estimated to contribute to conversion, engineers and the like can efficiently advance work related to behaviors that contribute to conversion improvement or modifications to web page XX.

[0164] <Generation of suspected improvement data required for simulation>

[0165] Fig.10 This is a diagram for explaining an example of a method for generating pseudo improvement data used in simulating the influence of a change in a behavior-related numerical value on a conversion-related numerical value.

[0166] exist Fig.10 The processing stage ST1 and the processing stage ST2 are shown in FIG.

[0167] The processing stage ST1 shows the relationship between the number of jumps of each web page and the browsing users before the simulation is performed.

[0168] In addition, "Page A Jump Count" represents the statistical value of the number of jumps to web page A, and other columns "Page B Jump Count", "Page C Jump Count", "Page D Jump Count", and "Page XX Jump Count" also represent the statistical values ​​of the number of jumps for each web page.

[0169] Here, the “page A jump count” is an example of the second behavior extracted from the first behavior.

[0170] exist Fig.10 In the case of , the number of jumps to web page A made by browsing user A is one, the number of jumps to web page A made by browsing user B is two, the number of jumps to web page A made by browsing user C is two, the number of jumps to web page A made by browsing user D is one, and the number of jumps to web page A made by browsing user E is four. In this case, the average value of the number of jumps to web page A is calculated as "2".

[0171] The processing stage ST2 is a diagram for explaining an example of generating pseudo-improvement data when the “number of jumps” is required to increase by 20%.

[0172] The value obtained by increasing the average value "2" by 20% is "2.4". In order to increase the average value, the data of browsing users whose jump count is less than the average value can be deleted, and the data of browsing users whose jump count is greater than the average value can be added.

[0173] In the case of processing stage ST2, the data of browsing user C is deleted and replaced with the data of browsing user E.

[0174] As a result, the average value of the number of jumps to the web page A calculated using the pseudo improvement data is calculated to be "2.4." That is, pseudo improvement data satisfying a 20% increase in the number of jumps can be created by replacing the browsing user data once.

[0175] However, in the processing stage ST1, even if the data of browsing users with a jump number smaller than the average value is replaced with the data of browsing users with a jump number larger than the average value, the average value of the suspected improvement data after replacement may not necessarily satisfy the specified increase of 20%.

[0176] Fig.11 This is a diagram for explaining another example of a method for generating pseudo improvement data used in simulating the effect of a change in a behavior-related numerical value on a conversion-related numerical value. Fig.11 In, with Fig.10 The corresponding parts are marked with corresponding symbols.

[0177] exist Fig.11 The processing stage ST1 and the processing stage ST2 are also shown.

[0178] The difference is that in the processing stage ST2, the data of browsing user C is replaced by the data of browsing user D.

[0179] exist Fig.11 In order to facilitate the explanation, the number of jumps from browsing user D to web page A is changed to "3". Fig.10 same.

[0180] exist Fig.11 In the case of , the data of browsing user D whose jump number is "3" is used as the browsing user whose jump number is greater than "2.4".

[0181] Therefore, the average value of the number of jumps to the web page A in the generated suspected improvement data is calculated to be "2.6".

[0182] However, the average value of the generated suspected improvement data does not satisfy the specified "jump count" increase of 20%. Therefore, the deletion of the browsing user's data and the copying of the data of other browsing users are repeated until the jump count increase of 20% is satisfied.

[0183] Fig.12 This is a diagram for explaining the continuation of another example of a method for generating pseudo improvement data used in simulating the effect of a change in a behavior-related numerical value on a conversion-related numerical value. Fig.12 In, with Fig.11 The corresponding parts are marked with corresponding symbols.

[0184] exist Fig.12The processing stage ST2 and the processing stage ST3 are shown in FIG.

[0185] exist Fig.12 In the processing stage ST3, the data of browsing user A is replaced with the data of browsing user D.

[0186] Therefore, the average value of the number of jumps to the web page A in the generated suspected improvement data is calculated to be "3". This value corresponds to a 50% increase in the "number of jumps", which satisfies a 20% increase.

[0187] As described above, the creation of pseudo improvement data necessary for calculating the amount of change in the numerical value related to the conversion is completed.

[0188] <Forecast of conversion improvement>

[0189] Fig.13 This is a diagram for explaining processing operations related to prediction of conversion improvement.

[0190] As described above, "Page A Jump Number" is an action that contributes greatly to the conversion of web page XX. However, it is always one of the actions that contribute greatly, and the number of conversions to web page XX is also related to actions other than "Page A Jump Number".

[0191] exist Fig.13 In the example, the numerical value of the second behavior (ie, "page A jump number") used for generating the suspected improvement data is marked as the source data.

[0192] exist Fig.13 In the case of , the browsing users whose behavior history includes conversion to web page XX are browsing users D and E. Therefore, the actual value of the number of conversions in the source data is "2", and the actual value of the number of conversions in the suspected improvement data is "3".

[0193] Processor 21 (see Figure 2 ) By comparing the required conversion numbers, the conversion improvement rate and the conversion improvement number are calculated and displayed on the display 36 of the information terminal 30 operated by engineers, etc.

[0194] exist Fig.13 In the case of , the improvement number of conversion is "1". In addition, the improvement rate of conversion is calculated as 50% by dividing the improvement number "1" by the conversion number "2" of the source data.

[0195] <Implementation method 2>

[0196] This embodiment is also based on the information processing system 1 described in the first embodiment.

[0197] In the above-mentioned first embodiment, the case where the amount of change in the numerical value related to the displayed upper N behaviors is specified and the influence of the change in behavior on the numerical value related to conversion is simulated is described, but in this embodiment, the relationship between the premise and the result is changed. That is, the case where the amount of change in the numerical value related to conversion is specified and the amount of change in the upper N behaviors that can realize the change is simulated is described.

[0198] Fig.14 3 is a diagram for explaining an example of a screen 320 for inputting simulation conditions and displaying simulation results. Fig.14 In, with Fig. 9 The corresponding parts are marked with corresponding symbols.

[0199] Fig.14 The screen 320 shown is also displayed on the information terminal 30 (see Figure 1 ) of the display 36 (refer to Figure 3 )superior.

[0200] and Fig. 9 The screen 310 shown is different in that the title of the improvement value input field 304 is changed to "CV improvement rate" and the behaviors displayed in the display field 305 are changed to "the top N behaviors with the improvement rate that achieves conversion".

[0201] In addition, Fig.14 In the case of , “Visited page XX” is displayed in the conversion input field 301 , “5” is displayed in the display number input field 302 , and “10%” is input in the improvement value input field 304 .

[0202] Therefore, a title “Improvement rate of each behavior required to improve conversion by 10%” is given in display column 305 , and five behaviors are displayed.

[0203] In the display field 305, the behavior with a smaller change in the numerical value is displayed at the top. Fig.14 For example, "Number of page jumps_ " is displayed at the top. In addition, "Number of page jumps_ The length of the bar graph for " is 20%.

[0204] This means that if you increase the number of page jumps from page C to page A by 20%, you can improve conversions by 10%.

[0205] In addition, in order to achieve the same improvement, the number of jumps from web page D to web page A needs to increase by 50%, the number of jumps from web page E to web page A needs to increase by 60%, the number of inflows to web page A needs to increase by 70%, and the number of jumps to web page A needs to increase by 120%.

[0206] Incidentally, the improvement rates required for the top N behaviors (ie, the second behavior) in order to satisfy the designated improvement rate of the conversion are extracted by reverse lookup from the relationship calculated by the method described in Embodiment 1. This method is described below.

[0207] First, various improvement rates are assigned to the top N behaviors, and the improvement rate of conversion and the like are calculated.

[0208] Next, samples satisfying the value in input field 304 are extracted from the calculated improvement rates of conversion, and the top N behaviors used in the calculation of the extracted samples and the corresponding improvement rates are extracted.

[0209] Through the above processing, we can achieve Fig.14 The display of screen 320 is shown.

[0210] <Implementation method 3>

[0211] This embodiment is also based on the information processing system 1 described in the first embodiment.

[0212] In the present embodiment, a case will be described in which a weighting based on a modality indicating a feature of a web page can be selected in addition to the improvement rates of the top N behaviors.

[0213] Fig.15 3 is a diagram for explaining an example of a screen 330 for inputting simulation conditions and displaying simulation results. Fig.15 In, with Fig. 9 The corresponding parts are marked with corresponding symbols.

[0214] exist Fig.15 The screen 330 shown in the figure has a selection bar 331 added thereto for specifying whether or not to perform weighting based on the modality. Fig. 9 The other screen structures of screen 330 are different from those of screen 310. Fig. 9 common.

[0215] exist Fig.15 In the case of , the question "Weighting of modalities" is marked in the selection column 331. Fig.15 In the case of , no check mark is marked in the selection column 331. That is, the modality-based weighting is not selected.

[0216] therefore, Fig.15 The display of the display column 305 in Fig. 9 The display of the display column 305 in is the same.

[0217] Fig.16This is a diagram for explaining the change in the display of the screen 330 when the modal information is added to the simulation conditions. Fig.16 In, with Fig.15 The corresponding parts are marked with corresponding symbols.

[0218] As mentioned above, in Fig.16 In the case of , a check mark is marked in the selection column 331. That is, the modality-based weighting is selected.

[0219] If used Figure 4 As explained above, modality is information about the page structure, content, and design of a web page, for example. Therefore, when modality-based weighting is enabled, it means evaluating the impact of the page structure, content, design, and browsing user behavior of a web page on conversion.

[0220] For example, the impact of the modality on the conversion is evaluated as follows: As a prerequisite for the evaluation, for example, the user pre-sets a weight coefficient for each modality for each behavior.

[0221] For example, for “jump between pages”, the weight coefficient of “page structure” is set to 1.0, the weight coefficient of “content” is set to 0.7, and the weight coefficient of “design” is set to 0.3.

[0222] In addition, for “inflow”, the weight coefficient of “page structure” is set to 0.8, the weight coefficient of “content” is set to 0.8, and the weight coefficient of “design” is set to 0.3.

[0223] In this setting, 1.0 is used as the weight coefficient representing the influence of "page structure" on "inter-page jump (for example, from page A to page B)", and 0.8 is used as the weight coefficient representing the influence of "content" on "inflow page".

[0224] These weight coefficients may be stored in the system as a matrix and may be changed by the user.

[0225] exist Fig.16 In the case of the display field 305 shown, the top order of behaviors having a high contribution to the improvement in conversion does not change, but the length of the bar graph indicating the degree of improvement in conversion changes.

[0226] For example, when "Structure" is included as one of the modes, the conversion improvement rate is set to the second "Number of Page Jumps_ The improvement in conversions for a 10% improvement dropped from over 35 points to under 35 points.

[0227] In addition, the improvement rate of conversion is below the fourth place, and the improvement numbers are all lower than before the modality-based weighting.

[0228] Engineers and the like can easily understand the relationship between behaviors and modes that affect conversion improvement by observing changes in the length of the bar graph or changes in the arrangement of behaviors.

[0229] <Implementation method 4>

[0230] This embodiment is also based on the information processing system 1 described in the first embodiment.

[0231] In this embodiment, step 4 (refer to Figure 5 ) is described below with one specific example of the processing action in FIG.

[0232] In this embodiment, a case is described where the top N behaviors (second behaviors) are extracted using a prediction model that performs machine learning on the relationship between behaviors that constitute the behavior history (first behaviors) and behaviors with a high possibility of achieving conversion (i.e., contribution).

[0233] Fig.17 It is a diagram for explaining the process of performing machine learning on the prediction model 200.

[0234] In the present embodiment, a learning model based on, for example, a gradient boosting decision tree (GBDT) or other decision trees is used as the prediction model 200 .

[0235] exist Fig.17 In the machine learning of the prediction model 200 shown, for example, in step 1 (refer to Figure 5 ) as training data.

[0236] Fig.17 The prediction model 200 shown is a model that learns the following relationship: taking any behavior history (i.e., a set of behaviors (first behaviors)) as input, and outputting the possibility of achieving conversion in association with the behavior (first behavior). In addition, the possibility here is equivalent to the above-mentioned contribution. Therefore, the higher the possibility, the higher the probability of conversion to the specified web page.

[0237] The output (prediction data) of the prediction model 200 is given to the evaluation unit 210 and compared with the correct answer data. In this embodiment, the comparison here is called evaluation.

[0238] When the evaluation unit 210 obtains the difference between the two data, it updates the weight set for the first behavior in the calculation process of the prediction model 200 so that the difference becomes smaller.

[0239] In addition, in the correct answer data, for example, in step 1 (refer to Figure 5 ) is part of the behavioral history obtained in .

[0240] By repeating the prediction by the prediction model 200 and the evaluation by the evaluation unit 210 , a prediction model 200 with high prediction accuracy is generated.

[0241] Fig.18 This is a diagram for explaining the extraction process of the top N behaviors (second behaviors). Fig.18 In, with Fig.17 The corresponding parts are marked with corresponding symbols.

[0242] The prediction model 200 here is obtained by Fig.17 The learning model generated by machine learning shown.

[0243] When a behavior history is input to the prediction model 200, a behavior associated with the likelihood of reaching a conversion is output.

[0244] The extraction unit 220 sorts the input behaviors in the order of numerical values ​​indicating the likelihood of achieving conversion, and outputs the top N behaviors.

[0245] By generating the prediction model 200 using machine learning, even when the number of behavior histories to be processed or behaviors constituting the behavior histories is large, it is possible to accurately predict a behavior (second behavior) that has a high possibility of achieving conversion.

[0246] <Implementation method 5>

[0247] This embodiment is also based on the information processing system 1 described in the first embodiment.

[0248] In this embodiment, step 6 (refer to Figure 5 ) is described below with one specific example of the processing action in FIG.

[0249] Fig.19 FIG. 2 is a diagram for explaining an example of a simulation of a transformation using the prediction model 230. Fig.19 In, with Fig.13 The corresponding parts are marked with corresponding symbols.

[0250] exist Fig.19 The prediction model 230 shown also uses, for example, a learning model based on a gradient boosting decision tree (GBDT: GradientBoosting Decision Tree) or other decision trees.

[0251] The prediction model 230 here is to convert the behavior history into the specified column 303 (refer to Figure 8 ) is a learning model that takes the behavior history specified by the user as input and the predicted value of the number of conversions achieved as output. The "predicted value of the number of conversions achieved" here is an example of a result related to conversion.

[0252] The machine learning of the prediction model 230 is achieved by comparing it with the prediction model 200 (refer to Fig.17 )The same method is used.

[0253] In this embodiment, the "source data" is an example of the first behavior history, and the "suspected improvement data" is an example of the second behavior history. Therefore, the result output by the prediction model 230 for the "source data" is called the first result, and the result output by the prediction model 230 for the "suspected improvement data" is called the second result.

[0254] exist Fig.19 In the case of , the number of conversions in the source data is also predicted to be "2", and the number of conversions in the suspected improvement data is also predicted to be "3". Fig.19 , the outputs of the prediction model 230 are labeled as “predicted CV number 2” and “predicted CV number 3”.

[0255] When the predicted CV number is obtained, the processor 21 (refer to Figure 2 ) calculates the conversion improvement rate and the conversion improvement number, and displays them on the display 36 of the information terminal 30 operated by engineers, etc.

[0256] exist Fig.19 In the case of , the improvement number of conversion is also "1". In addition, the improvement rate of conversion is calculated as 50% by dividing the improvement number "1" by the conversion number "2" of the source data.

[0257] In the case of this embodiment, the time required for calculating the conversion number is shortened compared with Embodiment 1. As a result, the time required for outputting the simulation result can be shortened.

[0258] <Other embodiments>

[0259] (1) The embodiments of the present invention are described above, but the technical scope of the present invention is not limited to the scope described in the embodiments. As can be seen from the claims, embodiments in which various changes or improvements are added to the embodiments described above are also included in the technical scope of the present invention.

[0260] (2) In the screen 330 (see Fig.15 ) describes the possibility of choosing whether or not to use modal-based weighting when calculating the impact on conversion improvement, but it may also be possible to set weights for each item that constitutes the modality, such as for page structure, content, and design, or to set weights for lower-level structures, such as using links, banners, documents, images, colors, and layouts as units.

[0261] In addition, even when weighting is selected, the size of the weight may be settable.

[0262] (3) In the above-mentioned embodiments 4 and 5, the case where a gradient boosting decision tree is used in the prediction model is described, but a neural network composed of an input layer, an intermediate layer, and an output layer may also be used.

[0263] (4) The processor in the above-mentioned embodiments refers to a processor in a broad sense, which includes, in addition to general-purpose processors (such as CPU (=Central Processing Unit)), special-purpose processors (such as GPU (=Graphical Processing Unit), ASIC (=Application Specific Integrated Circuit), FPGA (=Field Programmable Gate Array), programmable logic devices, etc.).

[0264] In addition, the actions of the processors of the above-mentioned embodiments may be performed by a single processor alone, or may be performed collaboratively by multiple processors located in physically separated locations. In addition, the order in which the actions of the processors are performed is not limited to the order described in the above-mentioned embodiments, and may also be changed individually.

[0265] (5) In the above-mentioned embodiment, in the business support server 20 (refer to Figure 1 ) performs all the processing of providing services to support the business of improving the conversion to the web page, but it can also be performed by the information terminal 30 (refer to Figure 1 ) executes all the corresponding processing. In addition, the same processing can also be executed through the cooperation of the business support server 20 and the information terminal 30. That is, the processing action can also be implemented through the cooperation of multiple processors.

[0266] <Note> (((1)))

[0268] An information processing system has one or more processors, which obtain the behavior history of multiple users in a website, accept the setting of specific behavior, i.e., conversion, performed by the users in the website, extract a second behavior from the first behavior based on the behavior constituting the obtained behavior history, i.e., the first behavior, and information on whether the accepted conversion is included in the behavior history, simulate the relationship between the change of a numerical value related to the extracted second behavior and the change of a numerical value related to the conversion, and output the result of the simulation. (((2)))

[0270] The information processing system according to (((1))), wherein, when the one or more processors accept the specification of the second behavior, they output the result of the simulation regarding the accepted second behavior. (((3)))

[0272] An information processing system according to (((1))) or (((2))), wherein, when the one or more processors accept the specification of the change in the numerical value associated with the second behavior, they output the result of the simulation regarding the change in the numerical value associated with the accepted second behavior. (((4)))

[0274] An information processing system according to any one of (((1))) to (((3))), wherein, when the one or more processors accept the specification of a change in a numerical value associated with the conversion, output the change in a numerical value associated with the second behavior that achieves the change as a result of the simulation. (((5)))

[0276] An information processing system according to any one of (((1))) to (((4))), wherein the one or more processors extract one or more features of the web pages contained in the website based on the results of the simulation, and output the relationship between the extracted features and the change in numerical values ​​related to the conversion. (((6)))

[0278] An information processing system according to (((5))), wherein the one or more processors output a relationship between changes in one or more features of the web page and changes occurring in the conversion based on weighted operations corresponding to the features. (((7)))

[0280] An information processing system according to any one of (((1))) to (((6))), wherein the one or more processors extract the second behavior based on a learning model, wherein the learning model takes a behavior history as input and outputs a likelihood of a hypothetical user performing a behavior set as a conversion based on the input behavior history. (((8)))

[0282] An information processing system according to (((7))), wherein the one or more processors update each weight set for the first behavior through learning during the calculation process performed by the learning model, and extract the second behavior based on the updated weight. (((9)))

[0284] An information processing system according to any one of (((1))) to (((8))), wherein the one or more processors assign a first behavior history as input to a learning model that takes a behavior history as input and outputs whether a hypothetical user performs a behavior set as a conversion based on the input behavior history, and obtains a first result related to the conversion based on the assigned first behavior history, assigns a second behavior history generated based on the first behavior history as input to the learning model, and obtains a second result related to the conversion based on the assigned second behavior history, and simulates the relationship between the change in a numerical value related to the second behavior and the change in the numerical value related to the conversion based on the first result and the second result. (((10)))

[0286] A program for causing a computer to implement the following functions: a function of acquiring the behavior histories of multiple users in a website; a function of accepting settings for specific behaviors, i.e., conversions, performed by the users in the website; a function of extracting a second behavior from the first behavior based on information about the behaviors constituting the acquired behavior history and whether the accepted conversion is included in the behavior history; a function of simulating the relationship between changes in a numerical value associated with the extracted second behavior and changes in a numerical value associated with the conversion; and a function of outputting the results of the simulation.

[0287] According to the information processing system of (((1))), compared with a system that outputs only the behavior of the user estimated to contribute to conversion, the user can efficiently improve the website.

[0288] According to the information processing system of (((2))), the influence of the designated second behavior on the conversion can be output as a numerical value.

[0289] According to the information processing system of (((3))), the influence of the designated second behavior on the conversion can be output as a numerical value.

[0290] According to the information processing system of (((4))), it is possible to output a change in a numerical value associated with the second behavior, the second behavior realizing a change in a numerical value designated with respect to a conversion acceptance.

[0291] According to the information processing system of (((5))), it is possible to output the relationship between the change in the numerical value designated by the conversion acceptance and the change in the numerical value related to the feature of the web page.

[0292] According to the information processing system of (((6))), it is possible to output the relationship between the change in the feature of the web page and the change occurring in the conversion.

[0293] According to the information processing system of (((7))), even when the number of action histories to be processed is large, it is possible to extract the second action that affects the conversion.

[0294] According to the information processing system of (((8))), even when the number of action histories to be processed is large, it is possible to improve the accuracy of extraction of the second action that affects conversion.

[0295] According to the information processing system of (((9))), the time required for outputting simulation results can be shortened.

[0296] According to the procedure of (((10))), compared with a system that outputs only the behavior of the user estimated to contribute to conversion, the user can efficiently improve the website.

Claims

1. An information processing system, characterized in that: having one or more processors, The one or more processors Get the behavior history of multiple users on the website, accepting the setting of a specific action, i.e., conversion, performed by the user on the website, extracting a second behavior from the first behavior based on the behavior constituting the acquired behavior history, namely the first behavior and information on whether the accepted conversion is included in the behavior history, Simulating the relationship between the change of the value associated with the extracted second behavior and the change of the value associated with the conversion, And output the simulation results.

2. The information processing system according to claim 1, wherein: The one or more processors When the designation of the second behavior is accepted, a result of the simulation regarding the accepted second behavior is output.

3. The information processing system according to claim 1 or 2, wherein: The one or more processors When a designation of a change amount of a numerical value associated with the second behavior is accepted, a result of the simulation regarding the change amount of the numerical value associated with the accepted second behavior is output.

4. The information processing system according to any one of claims 1 to 3, wherein: The one or more processors When the specification of the amount of change of the numerical value associated with the conversion is accepted, the amount of change of the numerical value associated with the second behavior to achieve the amount of change is output as the result of the simulation.

5. The information processing system according to any one of claims 1 to 4, wherein: The one or more processors Based on the result of the simulation, one or more features of the web pages contained in the website are extracted, The relationship between the extracted features and the change in the numerical value associated with the conversion is output.

6. The information processing system according to claim 5, wherein: The one or more processors Based on the weighted operation corresponding to the features, the relationship between the change of one or more features of the web page and the change occurring in the conversion is output.

7. The information processing system according to any one of claims 1 to 6, wherein: The one or more processors The second behavior is extracted based on a learning model that takes a behavior history as an input and outputs a probability that a hypothetical user will perform a behavior set as a conversion based on the input behavior history.

8. The information processing system according to claim 7, wherein: The one or more processors During the calculation process performed by the learning model, each weight set for the first behavior is updated through learning, And extracting the second behavior based on the updated weight.

9. The information processing system according to any one of claims 1 to 8, wherein: the one or more processors, Assigning a first behavior history as input to a learning model that takes a behavior history as input and outputs whether a hypothetical user has performed a behavior set as a conversion based on the input behavior history, and obtaining a first result related to the conversion based on the assigned first behavior history, Assigning a second behavior history generated based on the first behavior history as input to the learning model, and obtaining a second result related to the conversion based on the assigned second behavior history, Based on the first result and the second result, the relationship between the change of the numerical value associated with the second behavior and the change of the numerical value associated with the conversion is simulated.

10. A computer-readable medium storing a program for causing a computer to execute a process, characterized in that: The process has: The process of obtaining the behavior history of multiple users on the website; A step of receiving a setting of a specific action, i.e., a conversion, performed by the user on the website; A step of extracting a second behavior from the first behavior based on the behavior constituting the acquired behavior history, that is, the first behavior and information on whether the accepted conversion is included in the behavior history; A step of simulating a relationship between a change in a numerical value associated with the extracted second behavior and a change in a numerical value associated with the conversion; as well as A step of outputting the simulation result.

11. An information processing method, characterized in that: have: The process of obtaining the behavior history of multiple users on the website; A step of receiving a setting of a specific action, i.e., a conversion, performed by the user on the website; A step of extracting a second behavior from the first behavior based on the behavior constituting the acquired behavior history, that is, the first behavior and information on whether the accepted conversion is included in the behavior history; A step of simulating a relationship between a change in a numerical value associated with the extracted second behavior and a change in a numerical value associated with the conversion; as well as A step of outputting the simulation result.

Images