Method and apparatus for advertising

Abstract

An advertising method and apparatus are disclosed. An advertising method according to one embodiment may include the steps of obtaining a list of users corresponding to a channel and a list of message advertising materials corresponding to the channel, obtaining an indicator regarding the response of each user to each message advertising material from a message advertising model, and determining an advertising target including an advertising target user and a message advertising material corresponding to the advertising target user.

Description

Technology Field

[0001] The following embodiments relate to advertising methods and devices. Background Technology

[0002] Advertising is a paid information delivery activity conducted through media by an identifiable advertiser with the goal of influencing customer attitudes or behaviors and ultimately inducing them to purchase their products. Advertising aims to have customers perform actions valuable to the advertiser's business, such as purchasing; conversion refers to a customer performing the target action after interacting with an advertisement (e.g., clicking a text ad or watching a video ad). The target action can be defined by the advertiser and may include, for example, purchasing or signing up for membership. Retargeting is an advertising technique used in the marketing stage to induce conversion by running an advertisement once again targeting the audience that was initially advertised to, or those who have encountered, an ad. The problem to be solved

[0004] Through the following embodiments, advertising technology for efficiently executing message advertising materials linked to display advertising materials can be provided.

[0005] However, technical challenges are not limited to the technical challenges described above, and other technical challenges may exist. means of solving the problem

[0006] An advertising method performed by a server providing an advertisement according to one embodiment comprises: a step of obtaining a user list corresponding to a channel and a message advertising material list corresponding to the channel; a step of obtaining an indicator regarding the response of each user to each message advertising material from a message advertising model based on a feature of each user included in the user list and a feature of each message advertising material included in the message advertising material list; and a step of determining an advertising target user and a message advertising material corresponding to the advertising target user based on the indicator regarding the response and an action feature of each user regarding a display advertising material corresponding to the channel.

[0007] The step of determining the advertising target user and the message advertising material corresponding to the advertising target user may include: determining a candidate list containing message advertising materials corresponding to each user based on an indicator regarding the response; filtering the candidate list based on the action feature of each user; and determining the advertising target user and the message advertising material corresponding to the advertising target user based on the filtered candidate list.

[0008] The step of determining the advertising target user and the message advertising material corresponding to the advertising target user may include: determining priority conditions for the combination of each user and each message advertising material based on the action features of each user; determining the ranking of the combination of each user and each message advertising material based on the priority conditions and indicators regarding the response; and determining the advertising target user and the message advertising material corresponding to the advertising target user based on the determined ranking.

[0009] The above priority condition may include at least one of a condition regarding the number of times the message ad creative is exposed, a condition regarding the number of times the display ad creative is exposed, and a condition regarding the number of times the display ad creative is clicked.

[0010] The above priority condition may further include conditions regarding indicators regarding the above reaction.

[0011] The above advertising method may further include the step of providing a message advertisement containing a message advertising material corresponding to the advertising target user to the terminal of the advertising target user through the chat room of the channel.

[0012] The above message advertisement may include message ad creatives and interfacing objects for conversion actions.

[0013] The above message ad model may include a neural network trained to estimate an indicator regarding a user's response to a message ad material based on a first response log for a message ad material corresponding to the channel and a second response log for a display ad material mapped to the message ad material.

[0014] The above action feature may include an action feature for a display ad creative mapped to each message ad creative included in the above message ad creative list.

[0016] A method for training a message advertising model performed on a server according to one embodiment includes: a step of obtaining a first response log for a message advertising material corresponding to a channel and a second response log for a display advertising material mapped to the message advertising material; a step of obtaining training data based on the first response log, the second response log, features of the message advertising material and features of the display advertising material; and a step of training a message advertising model that estimates an indicator regarding a user's response to the message advertising material based on the training data.

[0017] The step of acquiring the above training data may include: acquiring a first user's feature and an action feature of the first user based on the first response log; acquiring a second user's feature and an action feature of the second user based on the second response log; and acquiring the above training data based on the first user's feature and the second user's feature, the first user's action feature and the second user's action feature.

[0018] The step of acquiring the features of the second user and the action features of the second user may include: a step of determining whether a user included in the second reaction log is a registered user of the channel; a step of selecting the features of the third user and the action features of the third user determined to be a registered user of the channel from the features of the second user and the action features of the second user; and a step of acquiring the training data based on the features of the first user and the features of the third user, the action features of the first user and the action features of the third user.

[0020] A server according to one embodiment includes a processor that obtains a user list corresponding to a channel and a message ad material list corresponding to the channel, obtains an indicator regarding the response of each user to each message ad material from a message ad model based on a feature of each user included in the user list and a feature of each message ad material included in the message ad material list, and determines an ad target including an ad target user and a message ad material corresponding to the ad target user based on the indicator regarding the response and an action feature of each user regarding a display ad material corresponding to the channel.

[0021] The processor can determine the advertising target user and the message advertising material corresponding to the advertising target user by determining a candidate list including message advertising material corresponding to each user based on an indicator regarding the response, filtering the candidate list based on the action feature of each user, and determining the advertising target user and the message advertising material corresponding to the advertising target user based on the filtered candidate list.

[0022] The processor can determine the advertising target user and the message advertising material corresponding to the advertising target user by determining the priority condition of the combination of each user and each message advertising material based on the action feature of each user, determining the ranking of the combination of each user and each message advertising material based on the priority condition and the indicator regarding the response, and determining the advertising target user and the message advertising material corresponding to the advertising target user based on the determined ranking.

[0023] The above processor can provide a message advertisement including a message advertisement material corresponding to the advertising target user to the terminal of the advertising target user through the chat room of the above channel.

[0024] The above message ad model may include a neural network trained to estimate an indicator regarding a user's response to a message ad material based on a first response log for a message ad material corresponding to the channel and a second response log for a display ad material mapped to the message ad material.

[0026] A server for learning a message advertising model according to one embodiment includes a processor that acquires a first response log for a message advertising material corresponding to a channel and a second response log for a display advertising material mapped to the message advertising material, acquires training data based on the first response log, the second response log, features of the message advertising material and features of the display advertising material, and trains a message advertising model that estimates an indicator regarding a user's response to the message advertising material based on the training data.

[0027] In acquiring the training data, the processor may acquire a first user's feature and a first user's action feature based on the first response log, acquire a second user's feature and a second user's action feature based on the second response log, and acquire the training data based on the first user's feature and the second user's feature, the first user's action feature and the second user's action feature. Brief explanation of the drawing

[0029] FIG. 1 is a flowchart of an advertising method according to one embodiment. FIG. 2 is a diagram illustrating a candidate list determined based on an indicator regarding each user's response to each message ad creative according to one embodiment. FIG. 3 is a drawing for explaining a message advertisement according to one embodiment. FIG. 4 is a diagram illustrating the configuration of an advertising system according to one embodiment. FIG. 5 is a diagram illustrating a learning method for a message advertisement model according to one embodiment. FIG. 6 is a flowchart of the filtering operation of the second reaction log according to one embodiment. FIG. 7 is an example diagram of a server configuration according to one embodiment. Specific details for implementing the invention

[0030] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, actual implementations are not limited to the specific embodiments disclosed, and the scope of this specification includes modifications, equivalents, or substitutions included in the technical concept described by the embodiments.

[0031] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0032] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or coupled with that other component, or that there may be other components in between.

[0033] The singular expression includes the plural expression unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to specify the existence of the described features, numbers, steps, actions, components, parts, or combinations thereof, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0034] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this specification.

[0035] Hereinafter, embodiments will be described in detail with reference to the attached drawings. In the description with reference to the attached drawings, identical components are given the same reference numeral regardless of the drawing number, and redundant descriptions thereof will be omitted.

[0037] FIG. 1 is a flowchart of an advertising method according to one embodiment.

[0038] An advertising method according to one embodiment may be performed by a server that provides an advertisement to a terminal. For example, the server may include a server that provides an instant messaging service. For example, the server may be linked with an instant messaging service or a server that provides an instant messaging service. The instant messaging service may include a service that enables multiple users to perform real-time chat by transmitting instant messages, such as text messages, voice messages, and media files, in real time through a network such as a wireless internet or a wireless communication network, and related services.

[0039] The server may be linked with an application (hereinafter referred to as "app") or a web related to an instant messaging service executed on a terminal for a user interface (UI), function, operation, or service, etc. In the following, the app related to the instant messaging service or the web related to the instant messaging service may be referred to as an instant messenger.

[0040] Users can create a user account by signing up for an instant messaging service through an instant messenger. Users can use the instant messaging service through a device associated with the account registered to the service. A device associated with an account registered to the instant messaging service may refer to a device logged in with the account registered to the instant messaging service.

[0041] An account for using an instant messaging service according to one embodiment may include various types of accounts, such as a personal account, a corporate account, or a service account. A personal account may be an account for a general user, a corporate account may be an account for a specific company, and a service account may be an account for a specific service. A corporate account or a service account may be referred to as a channel.

[0042] Referring to FIG. 1, an advertising method according to one embodiment may include the step (110) of obtaining a user list and a message advertising material list corresponding to a channel. For example, the channel may be an advertiser's account or a service account provided by an advertiser. Based on the channel, a user list including one or more user accounts corresponding to the recipients of advertisements regarding the channel may be determined. In the following, a user may mean a natural person user or a user account.

[0043] According to one embodiment, a user list may include one or more users who have a certain relationship with the channel. For example, the user list may include users who have registered the channel as friends. For example, the user list may include users who subscribe to the channel. According to one embodiment, the user list may include users who do not have a certain relationship (e.g., friends) with the channel.

[0044] A list of message ad materials may include one or more ad materials provided in the form of a message. An ad material is advertising content provided to a user's terminal and may include, for example, at least one of text data, image data, video data, and audio data. A message ad material may be provided to a user's terminal in the form of a message transmitted through a chat room. A list of message ad materials corresponding to a channel may include one or more message ad materials registered corresponding to the channel. A message ad material corresponding to a channel may include an ad material for advertising at least one of an advertiser corresponding to the channel, a product sold by the advertiser, and a service provided by the advertiser.

[0046] An advertising method according to one embodiment may include a step (120) of obtaining an indicator regarding the response of each user to each message advertising material from a message advertising model. The step (120) may include obtaining an indicator regarding the response of each user to each message advertising material from a message advertising model based on the feature of each user included in a user list and the feature of each message advertising material included in a message advertising material list. The indicator regarding the response of each user to each message advertising material is a value indicating the degree of response of each user to each message advertising material, and may include, for example, at least one of a score indicating a response probability and a response possibility. In the following description, the indicator regarding the response is described as an example of a response probability, but the indicator regarding the response is not necessarily limited to a response probability.

[0047] The message ad model is a model that outputs the probability of a response to an arbitrary message ad creative by an arbitrary user, and may include, for example, a learning model. As an example, the message ad model may include a neural network trained to estimate the probability of a user's response to a message ad creative based on a first response log for a message ad creative corresponding to a channel and a second response log for a display ad creative mapped to the message ad creative. The learning method of the message ad model is described in detail below.

[0048] The input data of the message ad model may include user features and message ad creative features, and the output data of the message ad model may include response probabilities corresponding to the input data, that is, response probabilities of the user given as input data for the message ad creative given as input data. User features may include features indicating information stored corresponding to the user (e.g., gender, age, residence, occupation, hobbies, etc.). Features of the message ad creative may include features indicating at least one of information of the target advertised by the message ad creative (e.g., product, service, etc.) (e.g., identifier, type, category, sales channel, descriptive information, etc.) and information of the content included in the message ad creative (e.g., image, video, text, sound effects, color, style, etc.). User features may be stored by mapping to information for specifying the user (e.g., user identification information). Features of the message ad creative may be stored by mapping to information for specifying the message ad creative (e.g., message ad creative identification information). The server can query the user's features corresponding to the user's identification information from the database storing the user's features. The server can query the message ad creative features corresponding to the message ad creative's identification information from the database storing the message ad creative's features.

[0049] The probability of a user's response to a message ad creative is the probability that, when a user is exposed to a message ad creative, the user will perform a conversion action corresponding to the message ad creative, and may include, for example, the probability that the user will click on the message ad creative (e.g., pCTR (predicted click-through rate)), the probability that the user will visit the advertiser's site corresponding to the message ad creative, the probability that the user will purchase a product corresponding to the message ad creative, and the probability that the user will sign up for or participate in a service corresponding to the message ad creative.

[0050] From the message ad model, the response probability for each message ad material included in the message ad material list for each user included in the user list can be obtained. For example, if the user list includes a first user and a second user, and the message ad material list includes a first message ad material and a second message ad material, the response probability of the first user for the first message ad material, the response probability of the first user for the second message ad material, the response probability of the second user for the first message ad material, and the response probability of the second user for the second message ad material can be obtained.

[0052] An advertising method according to one embodiment may include a step (130) of determining an advertising target including an advertising target user and a message advertising material corresponding to the advertising target user. The step (130) may include a step of determining an advertising target user and a message advertising material corresponding to the advertising target user based on a response probability and an action feature of each user regarding a display advertising material corresponding to a channel. The display advertising material is advertising content in a form that is visually conveyed through a screen, and may include, for example, an advertising material intended to be displayed in the form of a banner in a part area of ​​a website or app screen.

[0053] According to one embodiment, a display ad creative corresponding to a channel can be mapped to a message ad creative corresponding to a channel. Message ad creatives included in a list of message ad creatives can be mapped to one or more display ad creatives. In other words, message ad creatives and display ad creatives can have a 1:1 or 1:n mapping relationship. The mapped message ad creatives and display ad creatives may correspond to related ad creatives. For example, the mapped message ad creatives and display ad creatives may include content for advertising the same product or service. For example, at least a portion of the content (e.g., text, images, videos) of the mapped message ad creatives and display ad creatives may be identical or similar.

[0054] An action feature according to one embodiment may include an action feature for a display ad creative mapped to each message ad creative included in a message ad creative list. The user's action feature for a display ad creative may include action data performed by the user regarding the display ad creative. The action data is data regarding an action performed on the ad creative, and may include, for example, information indicating the type of action and information regarding the time at which the action was performed. The type of action may include, for example, at least one of an action of viewing the ad creative, an action of playing the ad creative, and an action of clicking the ad creative. When a user performs an action on a display ad creative, an action feature indicating that the user performed the corresponding action on the display ad creative may be created and stored. For example, when a user views a display ad creative, an action feature indicating that the user viewed the display ad creative may be created and stored. For example, if a user clicks on a display ad creative, an action feature indicating that the user clicked on the display ad creative can be generated and saved.

[0055] A step (130) according to one embodiment may include a step of determining a candidate list containing message advertising materials corresponding to each user based on response probability, a step of filtering the candidate list based on each user's action feature, and a step of determining an advertising target user and a message advertising material corresponding to the advertising target user based on the filtered candidate list.

[0056] For example, the candidate list may include combinations of each user included in the user list and message ad creatives that have a high response probability for that user. For example, when the response probability for each message ad creative for each user is as indicated in the table (210) shown in FIG. 2, the candidate list (220) may include a combination of User 1 and message ad creative 1, which has the highest response probability for User 1, and a combination of User 2 and message ad creative 2, which has the highest response probability for User 2. The candidate list (220) of FIG. 2 illustrates a combination of each user and one message ad creative that has the highest response probability for each user. Unlike what is shown in the candidate list (220) of FIG. 2, a combination of each user and multiple message ad creatives that have a high response probability for each user may be included in the candidate list. Alternatively, a combination of each user and message ad creatives where the response probability for each user is greater than or equal to a threshold probability may be included in the candidate list. The candidate list (220) may include user identification information and message ad creative identification information. The candidate list (220) may include information regarding response probabilities. For example, the candidate list (220) may further include information regarding a response probability value of 0.88 for User 1's message ad creative 1.

[0057] The candidate list can be filtered based on each user's action features regarding display ad creatives. As described above, display ad creatives can be mapped to message ad creatives. For example, among the combinations of users and message ad creatives included in the candidate list, the list can be filtered to retain only those combinations where the user has been exposed to the display ad creative mapped to the message ad creative n times (where n is an arbitrary natural number) or more. In other words, assuming the combination of User 1 and Message Ad Creative 1 is included in the candidate list, if User 1 has been exposed to the display ad creative mapped to Message Ad Creative 1 n times or more, it may remain in the candidate list, and otherwise, it may be excluded from the candidate list. For example, among the combinations of users and message ad creatives included in the candidate list, the list can be filtered to retain only those combinations where the user has a history of clicking the display ad creative mapped to the message ad creative. In other words, assuming that a combination of User 1 and Message Ad Creative 1 is included in the candidate list, if User 1 has a history of clicking on a display ad creative mapped to Message Ad Creative 1, it may remain in the candidate list, and otherwise, it may be excluded from the candidate list. For example, among the combinations of User and Message Ad Creative included in the candidate list, only combinations in which the User has been exposed to the Message Ad Creative less than n times (n is an arbitrary natural number) may be filtered to remain. In other words, assuming that a combination of User 1 and Message Ad Creative 1 is included in the candidate list, if User 1 has been exposed to Message Ad Creative 1 less than n times, it may remain in the candidate list, and otherwise, it may be excluded from the candidate list.

[0058] Based on a filtered candidate list, one or more combinations of users and message ad creatives may be determined as advertising targets. The users in the combinations determined as advertising targets may be referred to as the advertising target users, and the message ad creatives in the combinations determined as advertising targets may be referred to as the message ad creatives corresponding to the advertising target users or the advertising target message ad creatives. For example, a predetermined number of combinations of users and message ad creatives randomly extracted from the filtered candidate list may be determined as advertising targets. For example, the top n combinations of users and message ad creatives with high response probabilities (where n is an arbitrary natural number) from the filtered candidate list may be determined as advertising targets. For example, all combinations of users and message ad creatives included in the filtered candidate list may be determined as advertising targets.

[0060] A step (130) according to one embodiment may include a step of determining priority conditions for a combination of each user and each message ad creative based on an action feature of each user, a step of determining a ranking of a combination of each user and each message ad creative based on priority conditions and response probability, and a step of determining an ad target user and a message ad creative corresponding to the ad target user based on the determined ranking.

[0061] The ranking of combinations of user and message ad creatives that satisfy priority conditions can be determined as a higher value than the ranking of combinations of user and message ad creatives that do not satisfy priority conditions. The ranking among combinations of user and message ad creatives that satisfy priority conditions can be determined in descending order of response probability. Following the ranking of combinations of user and message ad creatives that satisfy priority conditions, the ranking of combinations of user and message ad creatives that do not satisfy priority conditions can be determined in descending order of response probability.

[0062] According to one embodiment, the priority condition may include a condition regarding the number of times a message ad creative is exposed. For example, the priority condition may include a condition in which a user is exposed to the message ad creative less than n times (where n is any natural number). In this case, among the combinations of user and message ad creative, the ranking of the combination in which the user is exposed to the message ad creative less than n times (where n is any natural number) may be determined to be higher than the ranking of the combination in which this is not the case. In other words, assuming that the combination of user and message ad creative includes a combination of User 1 and Message Ad Creative 1, if User 1 is exposed to Message Ad Creative 1 less than n times, the ranking of the combination of User 1 and Message Ad Creative 1 may be determined to be higher than the ranking of the combination in which this is not the case.

[0063] According to one embodiment, the priority condition may include a condition regarding the time during which the message ad creative was exposed. For example, the priority condition may include a condition that the user has not been exposed to the message ad creative for more than the recent n hours (where n is an arbitrary real number). In this case, among the combinations of user and message ad creative, the ranking of the combination in which the user has not been exposed to the message ad creative for more than the recent n hours (where n is an arbitrary real number) may be determined to be a higher value than the ranking of the combination in which the user has not been exposed. In other words, assuming that the combination of user and message ad creative includes the combination of User 1 and Message Ad Creative 1, if User 1 has not been exposed to Message Ad Creative 1 for the recent n hours, the ranking of the combination of User 1 and Message Ad Creative 1 may be determined to be a higher value than the ranking of the combination in which the user has not been exposed. For example, the longer the time during which User 1 has not been exposed to Message Ad Creative 1, the higher the ranking of the combination of User 1 and Message Ad Creative 1 may be determined.

[0064] According to one embodiment, the priority condition may include a condition regarding the number of times a display ad creative is exposed. For example, the priority condition may include a condition in which a user is exposed to a display ad creative mapped to a message ad creative n times (where n is any natural number) or more. In this case, among the combinations of user and message ad creative, the ranking of the combination in which the user is exposed to the display ad creative mapped to the message ad creative n times (where n is any natural number) or more may be determined to be higher than the ranking of the combination in which this is not the case. In other words, assuming that the combination of user and message ad creative includes a combination of User 1 and Message Ad Creative 1, if User 1 is exposed to the display ad creative mapped to Message Ad Creative 1 n times or more, the ranking of the combination of User 1 and Message Ad Creative 1 may be determined to be higher than the ranking of the combination in which this is not the case.

[0065] According to one embodiment, the priority condition may include a condition regarding the number of times a display ad creative is clicked. For example, the priority condition may include a condition in which a user clicks a display ad creative mapped to a message ad creative. In this case, among the combinations of user and message ad creative, the ranking of a combination in which the user has a history of clicking a display ad creative mapped to a message ad creative may be determined to be higher than the ranking of a combination in which there is no such history. In other words, assuming that the combination of user and message ad creative includes a combination of User 1 and Message Ad Creative 1, if User 1 has a history of clicking a display ad creative mapped to Message Ad Creative 1, the ranking of the combination of User 1 and Message Ad Creative 1 may be determined to be higher than the ranking of a combination in which there is no such history.

[0066] According to one embodiment, the priority condition may include a condition regarding the response probability. For example, the priority condition may include a condition where the response probability is greater than or equal to a threshold probability. In this case, among the combinations of user and message ad creative, the ranking of a combination where the response probability of the user to the message ad creative is greater than or equal to a threshold probability may be determined to be higher than the ranking of a combination where it is not. In other words, assuming that the combination of user and message ad creative includes a combination of User 1 and Message Ad Creative 1, if the response probability of User 1 to Message Ad Creative 1 is greater than or equal to a threshold probability, the ranking of the combination of User 1 and Message Ad Creative 1 may be determined to be higher than the ranking of a combination where it is not.

[0067] Based on the determined ranking, one or more combinations of users and message ad creatives may be selected as ad targets. For example, the top n (n is any natural number) or m% (m is any positive real number) combinations of users and message ad creatives with the highest rankings may be selected as ad targets. For example, combinations satisfying priority conditions may be selected as ad targets. For example, n (n is any natural number) combinations randomly selected from combinations satisfying priority conditions may be selected as ad targets.

[0068] According to one embodiment, the ranking of each user and each message ad creative combination can be determined periodically. For example, the ranking of each user and each message ad creative combination can be determined at one-hour intervals.

[0070] An advertising method according to one embodiment may include the step of providing a message advertisement, comprising a message advertisement material corresponding to the advertising target user, to the terminal of the advertising target user through a channel's chat room. The message advertisement includes a message advertisement material and may be provided in the form of a message through the channel's chat room. The channel's chat room where the message advertisement is provided may correspond to the channel's chat room in which the advertising target user participates.

[0071] According to one embodiment, a message advertisement may include a message advertisement material and an interfacing object for a conversion action. The interfacing object for a conversion action is an interfacing object for inducing a user's conversion action regarding the message advertisement, and may include, for example, at least one of a site link for viewing or purchasing an advertisement product, a button connected to a site for viewing or purchasing an advertisement product, a site link regarding an advertisement service, and a button connected to a site regarding an advertisement service. For example, a conversion action for the advertisement may be performed by an input selecting an interfacing object for a conversion action.

[0072] For example, referring to FIG. 3, a message advertisement (300) may be provided through a channel's chat room. The message advertisement (300) may be displayed as a message sent by the channel. The message advertisement (300) may include a message advertisement material that includes an image (310) of an advertisement product and text (320) of an advertisement phrase. The message advertisement (300) may include interfacing objects (331, 332, 333) for conversion actions. The interfacing objects (331, 332, 333) may include a button (332) that connects to a site for viewing advertisement products, a button (331) that connects to a site for purchasing advertisement products, and a button (333) for downloading a discount coupon to induce the purchase of advertisement products. A site for viewing advertisement products may be displayed on the terminal upon input selecting the button (332). A site for purchasing advertisement products may be displayed on the terminal upon input selecting the button (331). A discount coupon can be downloaded by selecting the button (333).

[0074] FIG. 4 is a diagram illustrating the configuration of an advertising system according to one embodiment.

[0075] Referring to FIG. 4, an advertising system according to one embodiment may include an advertising target determination module (410), an advertising sending module (420), and a message advertising model (430).

[0076] The advertising target determination module (410) and the advertising sending module (420) may be modules of a server that perform the operation of the advertising method described above in FIG. 1. The advertising target determination module (410) and the advertising sending module (420) illustrated in FIG. 4 are merely examples of server configurations, and the server configuration is not limited thereto.

[0077] The message advertisement model (430) may be a model stored in memory inside the server, or a model stored in external memory accessible from the server.

[0078] An advertising target determination module (410) according to one embodiment can output an advertising target list (403) including an advertising target user and a message advertising material corresponding to the advertising target user, based on a user list (401) corresponding to a channel and a message advertising material list (402) corresponding to a channel.

[0079] The advertising target determination module (410) can output an advertising target list (403) based on the response probability for each message advertising material included in the message advertising material list (402) of each user included in the user list (401) obtained from the message advertising model (430).

[0080] An advertisement sending module (420) according to one embodiment can send an advertisement to a terminal of an advertisement target user included in the advertisement target list (403). The advertisement sent to the terminal of the advertisement target user may include a message advertisement material corresponding to the advertisement target user. The advertisement sent to the terminal of the advertisement target user may be provided in the form of a message through a channel's chat room.

[0082] FIG. 5 is a diagram illustrating a learning method for a message advertisement model according to one embodiment.

[0083] Referring to FIG. 5, a learning method for a message advertising model according to one embodiment may include the step of obtaining a first response log (511) for a message advertising material (501) corresponding to a channel and a second response log (512) for a display advertising material (502) mapped to the message advertising material (501).

[0084] The first response log (511) may include user action data regarding an advertisement that includes a message ad material (501) corresponding to a channel. A user who receives an advertisement that includes a message ad material (501) corresponding to a channel may perform an action regarding the advertisement. For example, the user may perform actions such as viewing or clicking the advertisement, accessing a site to view or purchase an ad product, purchasing an ad product, accessing a site regarding an ad service, or using an ad service. When a user who receives an advertisement that includes a message ad material (501) performs an action, the user's identification information and action data performed by the user (e.g., type of action, time of action performed, etc.) may be stored in the first response log (511).

[0085] The second response log (512) may include user action data regarding an advertisement that includes a display ad material (502) corresponding to a channel. The display ad material (502) may correspond to an ad material mapped to a message ad material (501). A user who receives an advertisement that includes a display ad material (502) corresponding to a channel may perform an action regarding the advertisement. For example, the user may perform actions such as viewing or clicking the advertisement, accessing a site to view or purchase an ad product, purchasing an ad product, accessing a site regarding an ad service, or using an ad service. When a user who receives an advertisement that includes a display ad material (502) performs an action, the user's identification information and action data performed by the user (e.g., type of action, time of action performed, etc.) may be stored in the second response log (512).

[0086] A learning method according to one embodiment may include the step of obtaining learning data (531) based on a first response log (511), a second response log (512), features of a message ad creative (501) and features of a display ad creative (502).

[0087] The step of acquiring training data (531) according to one embodiment may include the step of acquiring a first user's feature and an action feature of the first user based on the first response log (511), the step of acquiring a second user's feature and an action feature of the second user based on the second response log (512), and the step of acquiring training data (531) based on the first user's feature and the second user's feature, the first user's action feature and the second user's action feature. The first user is a user who has performed an action on an advertisement including a message advertisement material (501) corresponding to a channel, and may include at least one user recorded in the first response log (511). The second user is a user who has performed an action on an advertisement including a display advertisement material (502) corresponding to a channel, and may include at least one user recorded in the second response log (512). At least a portion of the first user set and the second user set may be identical to each other.

[0088] Based on the identification information of the first user included in the first response log (511), information stored corresponding to the first user (e.g., gender, age, residence, occupation, hobbies, etc.) may be obtained. For example, the feature of the first user may include information stored corresponding to the first user or embedding data of information stored corresponding to the first user. The action feature of the first user may include action data of the first user included in the first response log or embedding data of action data of the first user.

[0089] Based on the identification information of the second user included in the second response log (512), information stored corresponding to the second user (e.g., gender, age, residence, occupation, hobbies, etc.) may be obtained. For example, the feature of the second user may include information stored corresponding to the second user or embedding data of information stored corresponding to the second user. The action feature of the second user may include action data of the second user included in the second response log or embedding data of action data of the second user.

[0090] According to one embodiment, at least some of the data included in the second reaction log (512) may be filtered (520). The filtering (520) of the second reaction log is described in detail below through FIG. 6.

[0091] A learning method according to one embodiment may include a step (530) of training a message advertising model (540) that estimates the probability of a user’s response to a message advertising material (501) based on training data (531). For example, the message advertising model (540) may be trained to determine that a first user included in a first response log (511) is a user with a high probability of response to the message advertising material (501). The message advertising model (540) may be trained to estimate the probability of response to the message advertising material (501) as a higher value the more similar the features of the user are to the first user. The message advertising model (540) may be trained to estimate the probability of response of the first user as a higher value the more similar the features of the first user are to the message advertising material (501). For example, a message ad model (540) can be trained to determine that a second user included in a second response log (512) is a user with a high probability of responding to a message ad creative (501) mapped to a display ad creative (502).

[0093] FIG. 6 is a flowchart of the filtering operation of the second reaction log according to one embodiment.

[0094] Referring to FIG. 6, the server can determine (610) whether a user whose action data is included in the second response log is a registered user of the channel. In the following, among the users whose action data is included in the second response log, a registered user of the channel may be referred to as a third user.

[0095] Based on the second response log, the server can obtain (621) the features of the third user and the action features of the third user, which are determined to be registered users of the channel. The features of the third user and the action features of the third user can be converted (640) into training data for training a message advertising model. The server can store the action features of the third user in an action feature storage (630) corresponding to the identification information of the third user.

[0096] If the server determines that a user whose action data is included in the second reaction log is not a registered user of the channel, it can acquire the features of the user (622) and convert the features of the user into training data (640).

[0097] In other words, among the users whose action data is included in the second reaction log, the data of a third user corresponding to a registered user of the channel can be filtered. The filtered features of the third user and the action features of the third user can be converted into training data (640), and the action features of the third user can be stored in a separate action feature storage (630).

[0098] For example, masked user data based on whether the user is a registered user of the channel may be used as training data. The user data may include user features and user action features obtained based on the second response log. The masked user data may include information indicating whether the user is a registered user of the channel. For example, data of a user who is not a registered user of the channel may be included in the training data as a value of 0 or null, and data of a user who is a registered user of the channel may be included as is.

[0100] FIG. 7 is an example diagram of a server configuration according to one embodiment.

[0101] Referring to FIG. 7, the server (700) includes a processor (701), memory (703), and a communication module (705). A server (700) according to one embodiment may include a server that performs the advertising method and / or the learning method of the message advertising model described in FIG. 1 to FIG. 6.

[0102] A processor (701) according to one embodiment may perform at least one operation of the advertising method described above through FIGS. 1 to 4. For example, the processor (701) may perform at least one of the following operations: obtaining a user list corresponding to a channel and a message advertising material list corresponding to a channel; obtaining a response probability for each message advertising material of each user from a message advertising model based on the features of each user included in the user list and the features of each message advertising material included in the message advertising material list; and determining an advertising target including an advertising target user and a message advertising material corresponding to the advertising target user based on the response probability and the action features of each user for a display advertising material corresponding to a channel.

[0103] A processor (701) according to one embodiment may perform at least one operation of the learning method of the message advertisement model described above through FIGS. 5 and 6. For example, the processor (701) may perform at least one of the operations of obtaining a first reaction log for a message advertisement material corresponding to a channel and a second reaction log for a display advertisement material mapped to the message advertisement material, obtaining learning data based on the first reaction log, the second reaction log, features of the message advertisement material and features of the display advertisement material, and training a message advertisement model that estimates the probability of a user's reaction to the message advertisement material based on the learning data.

[0104] A memory (703) according to one embodiment may be a volatile memory or a non-volatile memory and may store data regarding the advertising method and / or the learning method of the message advertising model described in FIGS. 1 to 6. For example, the memory (703) may store at least one of data generated during the execution of the advertising method, data required to perform the advertising method, data generated during the learning process of the message advertising model, and data required to perform the learning method of the message advertising model. For example, the memory (703) may store a user list corresponding to a channel and a message advertising material list corresponding to a channel, and may store weight(s) between layers included in the learned message advertising model.

[0105] A communication module (705) according to one embodiment may provide a function for the server (700) to communicate with another electronic device or another server through a network. In other words, the server (700) may be connected to an external device (e.g., a user's terminal, a server, or a network) and exchange data through the communication module (705). For example, the server (700) may transmit and receive data with another server included in the advertising system through the communication module (705). Also, for example, the server (700) may transmit and receive data with an advertiser who requested an advertisement or a user's terminal that is the recipient of the advertisement through the communication module (705).

[0106] According to one embodiment, memory (703) may store a program in which the learning method of the advertising method and / or message advertising model described above through FIGS. 1 to 6 is implemented. A processor (701) may execute the program stored in memory (703) and control the server (700). The code of the program executed by the processor (701) may be stored in memory (703).

[0107] A server (700) according to one embodiment may further include other components not illustrated. For example, the server (700) may further include an input / output interface including an input device and an output device as a means for interfacing with a communication module (705). Also, for example, the server (700) may further include other components such as a transceiver, various sensors, a database, etc.

[0109] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using a general-purpose computer or a special-purpose computer, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. In addition, other processing configurations, such as parallel processors, are also possible.

[0110] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or instruct the processing unit independently or collectively. Software and / or data may be stored on any type of machine, component, physical device, virtual equipment, computer storage medium, or device so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on computer-readable recording media.

[0111] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may store program instructions, data files, data structures, etc., either individually or in combination, and the program instructions recorded on the medium may be those specifically designed and configured for the embodiment or those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

[0112] The hardware device described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.

[0113] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based thereon. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

[0114] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below.

Claims

Claim 1 An advertising method performed by a server providing an advertisement, comprising: a step of obtaining a user list corresponding to a channel and a message ad material list corresponding to the channel; a step of obtaining an indicator regarding the response of each user to each message ad material from a message ad model based on a feature of each user included in the user list and a feature of each message ad material included in the message ad material list; and a step of determining an advertising target user and a message ad material corresponding to the advertising target user based on the indicator regarding the response and an action feature of each user regarding a display ad material corresponding to the channel, wherein the action feature includes action data performed by each user with respect to an advertisement including a display ad material mapped to each message ad material included in the message ad material list. Claim 2 An advertising method according to claim 1, wherein the step of determining the advertising target user and the message advertising material corresponding to the advertising target user comprises: a step of determining a candidate list including a message advertising material corresponding to each user based on an indicator regarding the response; a step of filtering the candidate list based on an action feature of each user; and a step of determining the advertising target user and the message advertising material corresponding to the advertising target user based on the filtered candidate list. Claim 3 An advertising method according to claim 1, wherein the step of determining the advertising target user and the message advertising material corresponding to the advertising target user comprises: a step of determining priority conditions for the combination of each user and each message advertising material based on the action features of each user; a step of determining the ranking of the combination of each user and each message advertising material based on the priority conditions and an indicator regarding the response; and a step of determining the advertising target user and the message advertising material corresponding to the advertising target user based on the determined ranking. Claim 4 An advertising method according to paragraph 3, wherein the priority condition comprises at least one of a condition regarding the number of times the display advertising material is exposed and a condition regarding the number of times the display advertising material is clicked. Claim 5 An advertising method according to paragraph 4, wherein the above priority condition further includes a condition regarding an indicator regarding the above response. Claim 6 An advertising method according to claim 1, further comprising the step of providing a message advertisement including a message advertisement material corresponding to the advertising target user to the terminal of the advertising target user through the chat room of the channel. Claim 7 An advertising method according to claim 1, wherein the message advertisement includes a message advertisement material and an interfacing object for a conversion action. Claim 8 An advertising method according to claim 1, wherein the message advertising model comprises a neural network trained to estimate an indicator regarding a user's response to a message advertising material based on a first response log for a message advertising material corresponding to the channel and a second response log for a display advertising material mapped to the message advertising material. Claim 9 delete Claim 10 A method for training a message advertising model performed on a server, comprising: a step of obtaining a first reaction log for a message advertising material corresponding to a channel and a second reaction log for a display advertising material mapped to the message advertising material; a step of obtaining training data based on the first reaction log, the second reaction log, features of the message advertising material and features of the display advertising material; and a step of training a message advertising model that estimates an indicator regarding a user's reaction to the message advertising material based on the training data, wherein the first reaction log includes action data of the user performed on an advertisement including the message advertising material corresponding to the channel, and the second reaction log includes action data of the user performed on an advertisement including the display advertising material. Claim 11 In claim 10, the step of acquiring the training data comprises: acquiring a first user's feature and a first user's action feature based on the first response log; acquiring a second user's feature and a second user's action feature based on the second response log; and acquiring the training data based on the first user's feature and the second user's feature, the first user's action feature and the second user's action feature. Claim 12 In claim 11, the step of acquiring the features of the second user and the action features of the second user comprises: a step of determining whether a user included in the second reaction log is a registered user of the channel; a step of selecting the features of the third user and the action features of the third user determined to be a registered user of the channel from the features of the second user and the action features of the second user; and a step of acquiring the training data based on the features of the first user and the features of the third user, the action features of the first user and the action features of the third user. Claim 13 A computer program stored on a medium in combination with hardware to execute the method of any one of claims 1 through 8 and claims 10 through 12. Claim 14 A server comprising a processor that obtains a user list corresponding to a channel and a message ad material list corresponding to the channel, obtains an indicator regarding the response of each user to each message ad material from a message ad model based on the feature of each user included in the user list and the feature of each message ad material included in the message ad material list, and determines an ad target including an ad target user and a message ad material corresponding to the ad target user based on the indicator regarding the response and the action feature of each user regarding a display ad material corresponding to the channel, wherein the action feature includes action data performed by each user for an ad including a display ad material mapped to each message ad material included in the message ad material list. Claim 15 In claim 14, the processor determines the advertising target user and the message advertising material corresponding to the advertising target user by determining a candidate list including message advertising material corresponding to each user based on an indicator regarding the response, filtering the candidate list based on an action feature of each user, and determining the advertising target user and the message advertising material corresponding to the advertising target user based on the filtered candidate list. Claim 16 In claim 14, the processor determines the advertising target user and the message advertising material corresponding to the advertising target user by determining the priority conditions of the combination of each user and each message advertising material based on the action features of each user, determining the ranking of the combination of each user and each message advertising material based on the priority conditions and the indicators regarding the response, and determining the advertising target user and the message advertising material corresponding to the advertising target user based on the determined ranking. Claim 17 In paragraph 14, the processor is a server that provides a message advertisement including a message advertisement material corresponding to the advertised user to the terminal of the advertised user through the chat room of the channel. Claim 18 In claim 14, the message ad model comprises a neural network trained to estimate an indicator regarding a user's response to a message ad material based on a first response log for a message ad material corresponding to the channel and a second response log for a display ad material mapped to the message ad material. Claim 19 A server for training a message advertising model comprises a processor that acquires a first reaction log for a message advertising material corresponding to a channel and a second reaction log for a display advertising material mapped to the message advertising material, acquires training data based on the first reaction log, the second reaction log, features of the message advertising material and features of the display advertising material, and trains a message advertising model that estimates an indicator regarding a user's reaction to the message advertising material based on the training data, wherein the first reaction log includes action data of the user performed on an advertisement including the message advertising material corresponding to the channel, and the second reaction log includes action data of the user performed on an advertisement including the display advertising material. Claim 20 In claim 19, the processor, in acquiring the training data, acquires a first user feature and an action feature of the first user based on the first reaction log, acquires a second user feature and an action feature of the second user based on the second reaction log, and acquires the training data based on the first user feature and the second user feature, the first user action feature and the second user action feature.

Images