Server and computer program

Abstract

We provide technology that enables more efficient selection of influencers suited to advertised products and services. [Solution] A server that performs the following steps: acquire vectorized influencer information for each of a plurality of influencers; acquire product information for a product, which is a good or service being advertised; vectorize the product information; calculate a similarity between the vectorized influencer information and the vectorized product information; select a first influencer from the plurality of influencers that is highly relevant to the product based on the calculated similarity; and output information about the first influencer.

Description

[Technical field]

[0001] The present invention relates to a server and a computer program. [Background technology]

[0002] In recent years, advances in communication technology have led to the provision of social networking services (SNS) that allow a wide range of information to be shared among a large number of users. On SNS, any user can transmit a wide range of information, opinions, assertions, reviews, etc., and among these, users who transmit information with great influence are called "influencers."

[0003] An influencer is a user who has a high influence on a specific platform, and generally corresponds to a user with a large number of followers, subscribers, plays, views, etc. In recent years, "influencer marketing," which utilizes the influence of influencers to promote the target of promotion, has been attracting attention. Patent Document 1 discloses a technology that mediates between an advertiser and an influencer suitable for the product or service to be advertised. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2023-535974 A Summary of the Invention [Problem to be solved by the invention]

[0005] The present invention aims to provide a technology for more efficiently selecting influencers suitable for advertising. [Means for solving the problem]

[0006] The present invention, corresponding to an aspect for solving the above-mentioned problem, is a server that communicates with a client terminal, the server comprising one or more processors, a memory, and a program stored in the memory, the program, when executed by the one or more processors, causes the server to execute the following: acquire first user information related to a first post of a first influencer; acquire second user information related to a second post of an influencer other than the first influencer; calculate a similarity of followers of the other influencer to followers of the first influencer based on the first user information and the second user information; select a second influencer that is determined to be similar to the first influencer from among the other influencers in accordance with the similarity; and notify the client terminal of information of the second influencer; Calculating the similarity includes calculating a first ratio of the number of users common between the first user information and the second user information to the first number of users included in the first user information, and calculating the similarity by a confidence interval of a population ratio based on the first number of users and the first ratio, wherein the first number of users is a sample number n, the first ratio is a sample ratio p, a coefficient according to a reliability according to a normal distribution is k, and a confidence coefficient is a lower limit value and an upper limit value of the confidence interval of the population ratio, lower limit: pk×(p(1−p)) 1 / 2 / (n) 1 / 2 Upper limit: p+k×(p(1-p)) 1 / 2 / (n) 1 / 2 and selecting the second influencer includes selecting the second influencer by comparing at least one of the calculated upper limit value and the calculated lower limit value. server. Effect of the Invention

[0007] According to the present invention, it is possible to provide a technology for more efficiently selecting influencers suitable for advertising. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram showing an example of the configuration of a system 10 according to an embodiment. [Diagram 2] FIG. 2 is a diagram showing an example of a hardware configuration of an information processing apparatus according to an embodiment. [Diagram 3] 5 is a flowchart corresponding to an example of processing executed in the server 101 according to the embodiment. [Figure 4A] 4 is a flowchart corresponding to an example of processing executed between the server 101 and the client terminal 102 according to the embodiment. [Figure 4B]11 is a flowchart corresponding to an example of a search process executed in the server 101 according to the embodiment. [Figure 5A] FIG. 2 is a diagram showing an example of a display screen of a client terminal 102 according to the embodiment. [Figure 5B] FIG. 13 is a diagram showing another example of the display screen of the client terminal 102 according to the embodiment. [Figure 5C] FIG. 13 is a diagram showing yet another example of the display screen of the client terminal 102 according to the embodiment. [Figure 5D] FIG. 13 is a diagram showing yet another example of the display screen of the client terminal 102 according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the invention according to the claims, and not all combinations of features described in the embodiments are essential to the invention. Two or more features among the multiple features described in the embodiments may be arbitrarily combined. In addition, the same reference numbers are used for the same or similar configurations, and duplicated descriptions are omitted.

[0010] <System configuration> 1 shows a general configuration of a system 10 corresponding to the embodiment. Here, a server 101 which is an information processing device implementing a service corresponding to the embodiment, a client terminal 102 which is an information processing device used by a client receiving information provided from the server 101, and an SNS server 103 which provides information on influencers to the server 101 are connected via a network 104 such as the Internet. An SNS user information database 105 is connected to the server 101, and an SNS information database 106 is connected to the SNS server 103.

[0011] The server 101 is an information processing device that mainly executes the process corresponding to this embodiment. The server 101 collects and acquires information of registered users who are posting information on the SNS server 103 using the API provided by the SNS server 103 via the network 104, and stores the information in the SNS user information database 105. Regarding the information of SNS users to be registered in the SNS user information database 105, only information of SNS users who satisfy a predetermined condition may be registered. The predetermined condition includes, for example, the number of posts in the channel or account of the SNS user is a predetermined number of posts or more, the number of follower registrations (also simply called the number of followers) of the channel, etc. is a predetermined number of followers or more, the total number of views of the content of the channel, etc., or the maximum number of views of a single channel, etc. is a predetermined number of views or more, and the number of high ratings (so-called "likes") for the content of the channel, etc. is a predetermined number of ratings or more. The acquired SNS user information may include information of the followers of the SNS user. The information of the followers includes, for example, the number of followers, follower ID, etc. In this embodiment, the SNS users registered in the SNS user information database 105 are also called "influencers". Moreover, information about influencers registered in the SNS user information database 105 is collectively referred to as "influencer information."

[0012] The server 101 can also update the information of the influencers stored in the SNS user information database 105, for example, at regular intervals.

[0013] The server 101 can also accept a designation of an influencer from the client terminal 102, and based on the influencer information stored in the SNS user information database 105, select (or choose, the same applies below) other influencers who are highly similar to the influencer designated by the client, and present a list of influencers to the client terminal 102.

[0014] The client terminal 102 can be configured as any information processing device such as a general-purpose personal computer or a smartphone. The client terminal 102 transmits information on an influencer designated by the client to the server 101, and receives a list of other influencers with high similarity based on the designated influencer from the server 101. Although only one client terminal 102 is illustrated in FIG. 1, this is merely for illustrative purposes, and it is possible for multiple client terminals 102 to be connected to the network 104, and multiple different clients to receive information provided from the server 101 via their respective client terminals 102.

[0015] The SNS server 103 is a server that provides a social networking service (SNS) on the network 104. The SNS information database 106 registers information for the SNS server 103 to provide the SNS service, and includes information about users of the SNS (SNS users) (including information about the followers of the SNS user (number of followers and follower ID)). The SNS server 103 also manages, as information about SNS users, information about videos, comments, photos, etc. posted by the SNS users on the SNS service, as well as comments posted by other SNS users in response to the posts and information about users who posted comments (hereinafter referred to as "comment users"). The SNS server 103 provides the server 101 with information about SNS users contained in the SNS information database 106 via an API.

[0016] In this embodiment, a client using the client terminal 102 sells products and services in Japan, for example, and places advertisements on SNS and other media to promote the sales of the products and services and to raise awareness of them. In addition, the aim is to have influencers feature the products and services on their own SNS accounts and channels, thereby further increasing awareness. Here, in this embodiment, the types of products that are the subject of advertising are not limited. Any type of product or service is eligible.

[0017] In addition, other purposes for selecting influencers are also envisioned, such as the following: - To select the destination for placement-type TrueView ads on video streaming sites, which specify the channel to which the ad is delivered in programmatic advertising. Here, TrueView ads refer to ads that are displayed in an interruptive manner when a user is trying to watch a video. -To find influencers who are a good fit for the target product as interview subjects for a new product monitoring survey. - Enter the PR composition plan (video flow) you are considering and search for influencers that match that composition plan

[0018] When requesting advertising from influencers, the effectiveness of the advertising varies depending on who is requested, so the selection of influencers is important. Although it is possible to request only influencers known to the client, there are limitations, and it is not easy to select influencers suitable for advertising from among them.

[0019] In this embodiment, the server 101 provides information on other influencers that are highly similar to an influencer specified by the client of the client terminal 102 in response to a request from the client terminal 102. In this embodiment, when selecting an influencer, attention is paid to the similarity of the influencer's followers and the degree of overlap of users who commented on the influencer's most recent post. An influencer that has followers similar to the followers of the client-specified influencer can be considered to have a high similarity to the client-specified influencer.

[0020] For example, if the majority of followers of influencer A, who is well-known in the field of cosmetics, follow influencer B in addition to influencer A, it can be estimated that influencer B is also well-known in the field of cosmetics. Therefore, influencer B can be recognized as an influencer with a high degree of similarity to influencer A. In addition, the degree of overlap of comment users commenting on the influencer's most recent post proves that they are interested in common users, so users with a high degree of overlap of comment users can be considered to have a high degree of similarity. In this way, other influencers with a high degree of similarity to the influencer specified by the client can be selected based on information about followers.

[0021] In this embodiment, the server 101 selects other influencers (second influencers) that are highly similar to the influencer (first influencer) specified by the client based on the SNS user information managed in the SNS user information database 105, and provides the information to the client terminal 102 as an influencer list.

[0022] 1 shows the SNS server 103 as a single server, the embodiment is not limited to this form, and multiple different SNS servers 103 may be provided depending on the form and platform of the SNS service. Platforms include a video distribution platform, a still image and video distribution platform, a text message distribution platform, a multimedia distribution platform including still images, videos, and text, etc. One platform may include multiple servers and databases.

[0023] <Hardware configuration> Next, a hardware configuration of the server 101 will be described with reference to Fig. 2. The server 101 can be configured as an information processing device, for example, from one or more personal computers. In Fig. 2, a CPU / GPU 201 as one or more processors controls the server 101 using programs and data stored in a RAM (random access memory) 202 and a ROM (read only memory) 205 as memories, an internal storage device 207, etc., and executes processing corresponding to the embodiments described below. The RAM 202 has an area for reading processing programs stored in the internal storage device 207 and information stored in an external storage device 208, and also has a work area used by the CPU / GPU 201 when executing various processes.

[0024] The input unit 203 is an input means for receiving input from an administrator of the server 101, and is composed of a keyboard, a mouse, etc. The communication I / F (interface) 204 functions as an I / F for connecting to the network 104, etc. The ROM 205 stores programs (e.g., a boot program, etc.) that control the entire server 101, etc. The display unit 206 is a display unit serving as a display screen, and is composed of a liquid crystal display device, etc.

[0025] The internal memory device 207 is mainly composed of a hard disk and stores programs and various application data for the server 101 to execute processes. The data stored here is read into the RAM 202 as needed. The external memory device 208 is a database that corresponds to the SNS user information database 105 and stores various information etc. at any time. The bus 209 provides interconnection of the above-mentioned respective blocks.

[0026] Although FIG. 2 has been described as the hardware configuration of the server 101, the basic hardware configuration of the client terminal 102 and the SNS server 103 can also be made the same as that shown in FIG. 2.

[0027] <Collection of SNS User Information> The processes in the server 101 will be described. FIG. 3 is a flowchart showing an example of the processes in the server 101. The processes corresponding to the flowchart can be realized, for example, by the CPU / GPU 201 of the server 101 executing a corresponding program (stored in the ROM 205, the internal memory device 207, etc.).

[0028] In S301, the server 101 uses the API of the SNS server 103 to collect and acquire information of a predetermined registered user in the service provided by the SNS server 103. The services include, for example, YouTube (registered trademark), Instagram (registered trademark), TikTok (registered trademark), Facebook (registered trademark), X (formerly Twitter (registered trademark)), etc., but are not limited thereto. The information of the registered user includes the profile information of the registrant and information of other users following the registrant (including the number of followers and the follower user IDs). The information of the registered user further includes information regarding the posts of the registered user, specifically, the posting date and time, information regarding the posting content (post title, summary information, hashtags), and information regarding comments posted by other SNS users on the post (comment user name, ID, comment posting date and time), etc.

[0029] The method of acquiring information on SNS users is not limited to the above-mentioned method, and may be performed by sending / providing a predetermined questionnaire to SNS users and receiving the responses from the SNS users by the server 101. Alternatively, the SNS users may be interviewed and the responses obtained therefrom may be acquired as the SNS user information.

[0030] The SNS users whose information is collected and acquired in S301 may be SNS users who satisfy the following conditions: the number of posts made to a specific SNS is equal to or greater than a specific number of posts, the number of followers is equal to or greater than a specific number of followers, the total number of views of posts is equal to or greater than a specific number of views, the number of likes is equal to or greater than a specific number of ratings, etc. In this embodiment, SNS users who satisfy these conditions and are registered in the SNS user information database 105 are called influencers.

[0031] In the next S302, information related to the content posted on the SNS by the influencer whose information was acquired in S301 (e.g., post title, summary information, hashtag, comment, etc.), information on the comment user, etc. are acquired for each SNS service. The information on the comment user includes ID information for identifying the comment user and a user name.

[0032] In the next S303, the various information acquired in S301 and S302 is stored in the SNS user information database 105. The information stored here is collectively called influencer information. At that time, the accounts of the same influencer are associated with each other even in different SNS services based on the mutual link information of the accounts between the SNS services. For example, in recent years, it is common for an influencer to have accounts on each of YouTube (registered trademark), Instagram (registered trademark), TikTok (registered trademark), Facebook (registered trademark), and X (formerly Twitter (registered trademark)), so that each account is associated with the same influencer and managed. Therefore, the influencer information of each influencer includes information related to accounts of different SNS services.

[0033] In S304, the information of the SNS user stored in S303 is processed, such as by adding an index for search. This is to enable the influencer to be searched for by any keyword in the process of FIG. 4A described later. The above process can be performed periodically for each piece of influencer information.

[0034] <Influencer Selection> Next, a method for selecting an influencer will be described. In this embodiment, the server 101 selects a second influencer that is considered to have a high similarity to a first influencer designated by a client. The first influencer may be a single person or multiple people, and it is preferable that the second influencer be multiple people.

[0035] 4A is a timing chart showing an outline of the flow of processing executed in the system 10 according to this embodiment. This processing can be realized, for example, by the CPU / GPU 201 of each of the server 101 and the client terminal 102 executing a corresponding program (stored in the ROM 205, the internal storage device 207, etc.).

[0036] In S401, the client terminal 102 displays a search screen on the display unit 206. The search screen is provided in advance by the server 101 and displayed. An example of the search screen displayed here is as shown in Fig. 5A. In the following S402, the client terminal 102 accepts an operation input from the client via the input unit 203. The operation input is an input to the screen 500 shown in Fig. 5A.

[0037] 5A shows an example of a search screen for specifying a first influencer and searching for similar influencers. FIG. 5A shows a state where no information has been entered. In FIG. 5A, a screen 500 includes a keyword input area 501, a narrowing down condition setting button 502, a search button 503, a display area 504 of designated candidates, a display area 505 of designated influencers, and a search button 506 for similar influencers.

[0038] Any keyword for identifying a client that the client wishes to specify can be input in the keyword input area 501. For example, in the case of beauty, a keyword for identifying a category such as "beauty influencer," "beauty YouTuber," "beauty related," or "beauty specialist" may be input, or the name or ID of a specific influencer, or any other information for identifying an influencer may be input.

[0039] When the filter condition button 502 is selected, the screen shown in FIG. 5B is displayed, and filter conditions related to the influencer to be searched can be set in addition to the keywords. Screen 510 shown in FIG. 5B is an example of a filter condition input screen. Various filter conditions can be set in filter condition 511. Once the filter condition 511 has been set, the user can return to the original screen 500 by selecting OK button 513 or cancel button 512.

[0040] On screen 510, first, the type of SNS of the influencer can be specified. For example, any one of YouTube (registered trademark), Instagram (registered trademark), TikTok (registered trademark), Facebook (registered trademark), and X (formerly Twitter (registered trademark)), or any combination thereof, can be specified.

[0041] Next, the gender and age group can be set as attributes. The number of followers can also be specified. The field of expertise, number of most recent posts, and number of comments can also be specified. These can be in the form of selecting one of the options prepared in advance, or any item can be input. For "most recent", multiple options can be prepared, such as 1 week, 10 days, 2 weeks, 3 weeks, 1 month, etc. Any other filtering condition can be input in the optional input field. If no setting is made for the above setting items, the setting item will not be a filtering condition.

[0042] By setting such narrowing conditions, you can limit your search to influencers such as those listed below. Influencers who have posted more than a certain number of times recently, for example, more than 10 times in the last month Influencers with a certain number of followers, over 50,000 Influencers who have accounts on multiple services, for example, a YouTube account, an X account, and an Instagram account Influencers whose posts related to keyword 501 account for a certain percentage (e.g., 80%) or more of the number of posts by the influencer

[0043] When a keyword is entered in the input field 501 and a narrowing down condition is set, and then the search button 503 is operated, in S403, the input information is transmitted from the client terminal 102 to the server 101. In S404, the server 101 processes the received input information. In this case, since a keyword for searching for influencers has been received, the server 101 searches the SNS user information database 105 based on the received keyword and narrowing down condition. For example, when a keyword related to "beauty" is specified and female is specified as the attribute, information on registered female influencers related to beauty is extracted.

[0044] The server 101 creates a list of designated candidates by listing the information of the extracted influencers. If the number of designated candidate influencers included in the list exceeds a certain number, the list may include only the designated candidates with the top number of followers, most recent posts, or most recent comments, for example.

[0045] In the next step S405, the list of designated candidates is transmitted to the client terminal 102 as a processing result. When the client terminal 102 receives the list of designated candidates, in step S406, the client terminal 102 updates the display contents of the search screen 500. Specifically, the list of designated candidates is displayed in display area 504 of designated candidates shown in Fig. 5A, resulting in the state shown in Fig. 5C. When there are many candidates, a scroll bar 504A is displayed in display area 504, allowing the display area to be scrolled.

[0046] In the display state shown in FIG. 5C, the client terminal 102 accepts an operation input from the client. When the keyword is re-input, the above-mentioned processes from S402 to S406 are repeated. On the other hand, when a designated candidate is selected from the designated candidate list displayed in the display area 504, the designated influencer is displayed in the display area 505. In addition, detailed information about each of the designated candidates may be acquired from the server 101. For example, when one of the candidates displayed as the designated candidate is selected and an instruction to display detailed information is input (for example, by right-clicking the mouse of the input unit 203 and then selecting detailed information display, or by double-tapping on the touch panel), the detailed information is transmitted from the server 101 and displayed on the screen. This process can also be included in the above-mentioned processes from S402 to S406.

[0047] When the designation of the first influencer is completed according to the above process and the search button 506 for similar influencers is selected in S407, the client terminal 102 transmits search information to the server 101. The search button 506 can be grayed out until the designated influencer is displayed in the display area 505. The search information includes information identifying the designated influencer. Note that, although one influencer is designated in FIG. 5C, multiple influencers may be designated.

[0048] When the server 101 receives a search instruction for similar influencers from the client terminal 102, the server 101 executes a search process in S409. Details of the search process will be described later with reference to Fig. 4B. When the search process is completed, the server 101 transmits the search results to the client terminal 102 in S410.

[0049] When the client terminal 102 receives the search results, it displays them in a list in S411. FIG. 5D shows an example of the display of the search results. In FIG. 5D, the second influencer determined to be similar to the first influencer specified by the client is displayed in each area 521 on the inspection result screen 520, and the account name, the number of views, the number of likes, the number of comments, etc. are also displayed. The "account name" indicates the name of the account used by the second influencer. The "number of views" indicates the total number of views in a certain period of time (for example, it may be one week, two weeks, or one month) or in a certain number of posts of the last time (for example, 5 posts, 10 posts, etc.). Alternatively, it may be the average number of views. The "number of likes" indicates the number of high ratings given to the posts of the influencer, and may be the total or average number for the period corresponding to the number of views or the number of posts. "Number of comments" indicates the number of comments made on the influencer's post, and may be a total or average value for a period corresponding to the number of views or for the number of posts.

[0050] The screen 520 includes a control 522 for displaying detailed information about each influencer. When displaying the search results, influencers determined to be similar are displayed on the screen in descending order of similarity. By operating the control 522, the client can check detailed information about each of the displayed influencers.

[0051] The screen 520 also includes check boxes 523, and the user can check the influencers to be consulted and requested for quotation. The check boxes 523 are divided into two types: one is assigned to each influencer, and the other is for selecting all influencers at once. The bottom of the screen 520 displays a button 524 for making a consultation and quotation request, and a back button 525. When the button 524 for making a consultation and quotation request is operated, a request for making a quotation for the cost of making an advertisement request to the influencer whose check box is checked can be sent to the server 101. When the button 524 for making a consultation and quotation request is operated in S417, it is considered that the request for consultation and quotation has been accepted, and the client terminal 102 sends to the server 101 information on the request for quotation, including information on the influencer whose check box has been checked in S418. If the user wishes to re-execute the search, the user can select the back button 525 to return to the screen of FIG. 5A or FIG. 5C.

[0052] When the request information is transmitted, the server 101 creates a quotation in S419, and transmits the quotation to the client terminal 102 in S420. The client terminal 102 displays the received quotation on the display unit 206 in S421.

[0053] In addition, the display of the search results in FIG. 5D may include a list of multiple influencers. For example, influencers determined to be similar may be grouped according to the magnitude of similarity and displayed for each group. The groups may be divided into three groups, for example, high, medium, and low similarity groups. For the group with the highest similarity, the individual influencers included therein may be displayed by default. Also, the checkboxes 523 may be made checkable on a group-by-group basis.

[0054] Next, the details of the search process in S409 in Fig. 4A will be described with reference to Fig. 4B. First, in S409-1, information on followers of a first influencer, who is an influencer designated by a client, is acquired from the SNS user information database 105. The information on followers includes, for example, the following information. Note that the SNS service (YouTube, Instagram, X, etc.) from which the information is acquired does not matter, but the information is acquired for each SNS service. Number of followers Information of users who commented on the first influencer's post - Information on users who liked the first influencer's post

[0055] Here, the posts from which user information is obtained refer to at least one of the following posts, for example: a predetermined number of posts, posts made within a predetermined period from the present (e.g., one week, ten days, two weeks, three weeks, one month, or more), a predetermined number of posts with a predetermined number of comments per post (e.g., 100, 200, 500, or more), posts with a predetermined number of comments per post made within a predetermined period from the present, a predetermined number of posts with a predetermined number of ratings per post (e.g., 100, 200, 500, or more), and posts with a predetermined number of ratings per post made within a predetermined period from the present. Here, a designation of an SNS service can be accepted from the client, and in that case, user information is obtained regarding posts on the designated SNS service. For example, if a narrowing down condition is designated in FIG. 5B, the condition can be referenced.

[0056] In the next step S409-2, influencers to be compared are selected. The selection criteria here may be, for example, at least one of influencers with the same attributes (e.g., gender, age) or field of expertise as the first influencer, influencers with the same or more followers, and influencers with a certain number of recent views or more, or all of these. If a narrowing-down condition is specified in FIG. 5B, the condition may be referenced.

[0057] In the following S409-3, for the influencer to be compared acquired in S409-2, information on followers is acquired from the SNS user information database 105. The information on followers is the same as that described in S409-1.

[0058] In the next S409-4, the information on the followers acquired in S409-1 is compared with the information on the followers acquired in S409-3 to calculate the follower similarity, and the influencers to be compared are sorted in order of similarity to be determined as the second influencer. When comparing information on followers, for example, when comparing comment users, a comparison is made for a common SNS service. For example, when targeting users who posted comments on YouTube videos posted by the first influencer, the comparison target influencer is also compared with users who posted comments on YouTube videos posted by the comparison target influencer. This is because even if comments posted on different SNS services are compared, the comment users do not necessarily use a common account name, and the number of comments also differs depending on the nature of the SNS service.

[0059] The similarity can be calculated by calculating a confidence interval of a population ratio for follower information. Specifically, the similarity between the followers of the first influencer and the followers of the comparison target influencer is calculated by calculating a confidence interval of a population ratio for the overlap rate of comment users who posted comments on each post of the first influencer and the comparison target influencer. The overlap rate of comment users refers to the proportion of comment users who comment on the posts of the first influencer who also comment on the posts of the comparison target influencer.

[0060] Specifically, the total number of followers of the first influencer is Ft, and the total number of followers of the comparison influencer is Fc. When considering the overlap rate between Ft and Fc, if Ft and Fc can be directly compared, the overlap rate can be obtained, but if there are tens of thousands, hundreds of thousands, or millions of followers, the calculation load increases accordingly. Therefore, in this embodiment, the overlap rate of comment users who comment on posts of each influencer is calculated, and the confidence interval of the population ratio is calculated based on this to estimate the overlap rate of the number of followers.

[0061] For example, suppose that the number of followers of the first influencer is 100,000, and the number of followers of the comparison influencer is 80,000. Of these, the number of comment users commenting on the posts of both influencers is 1,000 for the first influencer and 800 for the comparison influencer. When calculating the number of comment users for multiple posts, duplicates of the same person are excluded. Also, suppose that the number of duplicate comment users posting comments is 500. In this case, the sample size is n: 1,000, the sample proportion p: 0.5 (500 / 1,000), and the confidence level is 95%, so ignoring the correction term, the formula for calculating the confidence interval of the population proportion is as follows.

[0062] Lower limit: PRmin = p-1.96 × (p(1-p)) 1 / 2 / (n) 1 / 2 formula 1 Upper limit: PRmax = p + 1.96 × (p(1-p)) 1 / 2 / (n) 1 / 2 formula 2 Applying the above values ​​to this, the lower limit PRmin is calculated to be 47% and the upper limit PRmax is calculated to be 53%. In this case, it is estimated that the followers of the first influencer will match up to 53% of the followers of the comparison influencer.

[0063] The correction term C is expressed as follows for Equation 1 and Equation 2: C={(Nn) / (N-1)} 1 / 2The lower and upper limits are calculated by multiplying Equation 1 and Equation 2 by the value of the correction term. In the correction term, N represents the population size, and n represents the sample size. In the above case, N is 100,000. A population size N of less than 100,000 is called a finite population, and one that is 100,000 or more or cannot be measured is called an infinite population, and the correction term for an infinite population is approximately 1 and can be ignored. In the above example, the population size N was 100,000, so the correction term was ignored.

[0064] Similarly, by performing calculations using the above formulas 1 and 2 for other comparative influencers, the proportion of followers of each of the comparative influencers among the followers of the first influencer can be estimated.

[0065] The upper or lower limit of the confidence interval of the population ratio of each of the influencers to be compared obtained as above is compared, and a predetermined number of influencers to be compared starting from the one with the highest upper limit can be selected as the second influencer. Alternatively, a predetermined number of influencers to be compared starting from the one with the highest lower limit can be selected as the second influencer. In addition, for influencers to be compared whose lower limit or upper limit value is greater than a predetermined lower limit threshold, the one with the highest upper limit or lower limit value may be specified as the second influencer.

[0066] The above describes the case where the first influencer is a single person, but if multiple people are specified, the same process can be performed for each person. In that case, if there are overlapping influencers to be compared that are determined to be the second influencer, the overlaps are eliminated in the processing results. In that case, the overlapping influencers to be compared that are selected may be ranked higher. For example, if three people are specified as the first influencer, the influencer to be compared that is determined to be the second influencer by all three people may be ranked higher than the other influencers to be compared.

[0067] Although the above describes a case where the calculation process is performed based on the first influencer, the same process may be performed based on the second influencer. In this case, the ratio of followers of the first influencer to followers of other influencers is estimated. The higher this ratio, the higher the degree of similarity with the first influencer can be estimated. However, if the number of followers of the influencer to be compared is significantly lower than the number of followers of the first influencer, this similarity may not be useful.

[0068] Note that while Equation 1 and Equation 2 are for a reliability of 95%, they are changed to the following equations if the reliability is 99%. The lower limit is p-2.58×(p(1-p)) 1 / 2 / (n) 1 / 2 formula 3 The upper limit is p+258×(p(1-p)). 1 / 2 / (n) 1 / 2 formula 4 If we use formulas 3 and 4 to estimate the population ratio in the above example, the lower limit is 46% and the upper limit is 54%. Again, the calculation is done ignoring the correction term.

[0069] Furthermore, when the sample size n is less than 30, a formula different from the above formulas 1 to 4 can be applied. For example, when the number of comments is small, the following formula may be applied for calculation. If the sample size is n and the sample proportion is p, and n<30, The lower limit is n 2 / (n 1 F n +n 2 ) Equation 5 The upper limit is m 1 F m / (m 1 F m +m 2 ) Equation 6 In addition, n 1 =2n(1-p)+2,n 2 =2np,m 1 =2np+2,m 2 =2n(1-p) F n= the horizontal axis value of the upper probability a / 2% of the F distribution with degrees of freedom n1 and n2, F m = the horizontal axis value of the upper probability a / 2% of the F distribution with degrees of freedom m1 and m2, However, a = 0.05 at 95% confidence level, and a = 0.01 at 99% confidence level. The horizontal axis value F n For example, use the Excel function FINV to calculate FINV(a / 2, n 1 , n 2 ) can be calculated.

[0070] In the above, the similarity of an influencer's followers was estimated based on the overlap rate of comment users who posted comments, but a similar method can also be used to estimate the similarity of an influencer's followers based on the overlap rate of users who gave high ratings to a post by a first influencer.

[0071] In the present embodiment described above, by determining the similarity of followers, it is possible to select influencers similar to the influencer designated by the client. At this time, the processing load can be reduced regardless of the number of followers of the influencer. Furthermore, according to this embodiment, if one influencer is designated, the server 101 selects and presents other influencers with a high degree of commonality in followers without the client having to select them by himself, so that it is possible to improve the accuracy of influencer selection. As a result, the client can easily obtain a comprehensive list of other influencers likely to be suitable for advertising and promotion simply by finding an influencer suitable for the product, and can narrow down the candidates without omission with a limited amount of work.

[0072] In addition, according to the present embodiment, since it is possible to select influencers with a high degree of commonality among followers as the requestees for advertising, it is possible to approach the same followers via a plurality of different influencers to advertise products and services, thereby increasing the effectiveness of advertising.

[0073] <Summary of the embodiment> The above embodiment discloses at least the following information processing device and computer program. (1) A server that communicates with a client terminal, the server comprising one or more processors, a memory, and a program stored in the memory, the program being executed by the one or more processors to cause the server to: Obtaining first user information regarding a first post of a first influencer; Obtaining second user information regarding a second post of an influencer other than the first influencer; Calculating a similarity between the followers of the other influencer and the followers of the first influencer based on the first user information and the second user information; selecting a second influencer that is determined to be similar to the first influencer from among the other influencers according to the degree of similarity; notifying the client terminal of information about the second influencer; A server that runs the command. (2) the first user information is information of a user who posted a comment on the first post; The server according to (1), wherein the second user information is information of a user who posted a comment on the second post. (3) the first user information is information about a user who has given a high rating to the first post; The server according to (1), wherein the second user information is information of a user who gave a high rating to the second post. (4) A server described in any one of (1) to (3), wherein the first post and the second post are at least one of a predetermined number of posts, posts made within a predetermined period from the present, a predetermined number of posts with a predetermined number of comments per post or more, posts made within a predetermined period from the present with a predetermined number of comments per post or more, a predetermined number of posts with a predetermined number of ratings per post or more, and posts made within a predetermined period from the present with a predetermined number of ratings per post or more. (5) The server according to (4), wherein the first user information and the second user information are obtained for each social networking service on which the first post and the second post were made, and the similarity is calculated for each social networking service. (6) The server according to (5), wherein the social networking service on which the first post and the second post were made is a social networking service specified by the client terminal. (7) Calculating the similarity Calculating a first ratio of a number of users common to the first user information and the second user information to a first number of users included in the first user information; The server according to any one of (4) to (6), further comprising: calculating the similarity using a confidence interval of a population proportion based on the first number of users and the first proportion. (8) In the calculation of the similarity, the first number of users is a sample number n, the first ratio is a sample ratio p, and a lower limit and an upper limit of a confidence interval of the population ratio are calculated as follows: Lower limit: p-1.96×(p(1-p)) 1 / 2 / (n) 1 / 2 Upper limit: p+1.96×(p(1-p)) 1 / 2 / (n) 1 / 2 The server according to (7) above, wherein the calculation is based on the above. (9) Selecting the second influencer: The server according to (8), further comprising: selecting the second influencer by comparing at least one of the calculated upper limit value and the calculated lower limit value. (10) The server described in (9), wherein the selection of the second influencer is performed by selecting an influencer for which at least one of the calculated upper limit value and the calculated lower limit value is greater than a predetermined threshold value. (11) The server according to (10), wherein the first influencer is designated in the client terminal and notified to the server. (12) A server described in any one of (1) to (11), wherein the first influencer and the other influencers are registered users of a specified social networking service who disseminate information on the specified social networking service and who have at least one of the number of followers, number of posts, number of views, and number of likes that exceed a specified number. (13) The program, when executed by the one or more processors, causes the server to Accepting a specified request that designates the notified second influencer. The server according to any one of (1) to (12), further comprising: (14) A program according to any one of (1) to (13).

[0074] [Other embodiments] The invention is not limited to the above-described embodiment, and various modifications and variations are possible within the scope of the gist of the invention. Therefore, in order to clarify the scope of the present invention, the following claims are attached. The information processing device according to the present invention can also be realized by a computer program that causes one or more computers to function as the information processing device. The computer program can be provided / distributed by being recorded on a computer-readable recording medium or through an electric communication line. [Explanation of symbols]

[0075] 10: system, 101: server, 102: client terminal, 103: SNS server, 104: network

Claims

1. A server in communication with a client terminal, the server comprising one or more processors, a memory, and a program stored in the memory, the program being executed by the one or more processors to cause the server to: Obtaining first user information regarding a first post of a first influencer; Obtaining second user information regarding a second post of an influencer other than the first influencer; Calculating a similarity between the followers of the other influencer and the followers of the first influencer based on the first user information and the second user information; selecting a second influencer determined to be similar to the first influencer from among the other influencers according to the degree of similarity; notifying the client terminal of information about the second influencer; Run the command, Calculating the similarity Calculating a first ratio of a number of users common to the first user information and the second user information to a number of first users included in the first user information; calculating the similarity using a confidence interval of a population ratio based on the first number of users and the first ratio; The first number of users is the sample number n, the first ratio is the sample ratio p, a coefficient according to the reliability according to the normal distribution is k, and the reliability coefficient is the lower limit and upper limit of the reliability interval of the population ratio, Lower limit: p-k x (p(1-p)) 1 / 2 / (n) 1 / 2 Upper limit: p + k × (p(1-p)) 1 / 2 / (n) 1 / 2 Calculated based on The selecting of the second influencer includes selecting the second influencer by comparing at least one of the calculated upper limit value and the lower limit value.

2. the first user information is information of a user who posted a comment on the first post, The server according to claim 1 , wherein the second user information is information of a user who posted a comment on the second post.

3. the first user information is information of a user who has given a high rating to the first post, The server according to claim 1 , wherein the second user information is information of a user who has given a high rating to the second post.

4. The server of claim 1, wherein the first post and the second post are at least one of a predetermined number of posts, posts made within a predetermined period from the present, a predetermined number of posts with a predetermined number of comments per post or more, posts made within a predetermined period from the present with a predetermined number of comments per post or more, a predetermined number of posts with a predetermined number of ratings per post or more, and posts made within a predetermined period from the present with a predetermined number of ratings per post or more.

5. The server according to claim 4 , wherein the first user information and the second user information are obtained for each social networking service in which the first post and the second post were made, and the similarity is calculated for each social networking service.

6. The server according to claim 5 , wherein the social networking service on which the first post and the second post were made is a social networking service specified by the client terminal.

7. The server according to claim 1 , wherein the selection of the second influencer is performed by selecting an influencer for which at least one of the calculated upper limit value and the calculated lower limit value is greater than a predetermined threshold value.

8. The server of claim 7 , wherein the first influencer is designated in the client terminal and notified to the server.

9. The server of claim 1, wherein the first influencer and the other influencers are registered users of a specified social networking service who disseminate information on the specified social networking service and who have at least one of the number of followers, number of posts, number of views, and number of likes exceed a specified number.

10. The program, when executed by the one or more processors, causes the server to: accepting a predetermined request specifying the notified second influencer; The server of claim 1 , further comprising:

11. The program according to any one of claims 1 to 10.

Images