Server, method, and program

Abstract

This server generates candidate advertising messages for a product or service on the basis of information relating to the product or service. The server acquires a plurality of revised advertising messages, which are the result of revisions of the advertising messages revised by each of a plurality of revisers. The server acquires a voting result for each of the plurality of revised advertising messages from each of a plurality of voters.

Description

Server, method, and program

[0001] The present disclosure relates to a server, a method, and a program.

[0002] Conventionally, management devices that generate information used for advertising or sales promotion of products have been known (for example, Literature 1: JP 2023-149229 A). Literature 1 discloses that evaluation items of a product are input to a trained estimation model, and at least one statement output by the estimation model is used as a candidate for a promotional statement for the product.

[0003] Although Literature 1 discloses that the words output from a trained estimation model are used as candidates for product promotional words, it is difficult to directly use the promotional words for the product or service. Therefore, even if the technology disclosed in Literature 1 is used, there is a problem in that it is difficult to obtain an advertising message that can be used for the sales promotion of the product or service.

[0004] The present disclosure has been made in consideration of the above circumstances, and aims to obtain an advertising message that can be used to promote sales of products or services.

[0005] A first aspect of the present disclosure is a server including a candidate generation unit that generates candidate advertising messages for a product or service based on information about the product or service, a correction result acquisition unit that acquires multiple revised advertising messages that are the results of corrections to the candidate advertising messages and that have been corrected by each of multiple correctors, and a voting result acquisition unit that acquires voting results for each of the multiple revised advertising messages and that have been voted by each of multiple voters.

[0006] According to the present disclosure, it is possible to obtain an advertising message that can be used to promote a product or service.

[0007] FIG. 1 is a diagram showing an example of a schematic configuration of an advertising message generation system according to an embodiment of the present invention; FIG. 2 is a diagram for explaining an overview of the embodiment; FIG. 3 is a schematic block diagram of a computer that functions as each device in the system; FIG. 4 is an explanatory diagram for explaining processing executed by the advertising message generation system according to the embodiment of the present invention; and FIG. 5 is a diagram showing an example of information recorded in a blockchain.

[0008] Hereinafter, the embodiments will be described in detail with reference to the drawings.

[0009] <Configuration of Advertising Message Generation System> Figure 1 is a block diagram showing an advertising message generation system 10 according to the present embodiment. As shown in Figure 1, the advertising message generation system 10 according to the present embodiment includes an order terminal 11, a server 12, multiple editing terminals 14A, 14B, 14C, ..., multiple voting terminals 16A, 16B, 16C, ..., and multiple blockchain nodes 18A, 18B, 18C, .... Each device is communicatively connected via a network N such as the Internet. Note that, hereinafter, the multiple editing terminals 14A, 14B, 14C, ... will be simply referred to as editing terminals 14, the multiple voting terminals 16A, 16B, 16C, ... will be simply referred to as voting terminals 16, and the multiple blockchain nodes 18A, 18B, 18C, ... will be simply referred to as blockchain nodes 18.

[0010] In the advertising message generation system 10, advertising messages for products or services are generated. Fig. 2 is a diagram for explaining an overview of this embodiment. As shown in Fig. 2, in this embodiment, when generating advertising messages for products or services, advertising messages are generated and selected by a trained model, which is a generative artificial intelligence (AI), multiple modifiers U1, and multiple voters U2.

[0011] Specifically, as shown in FIG. 2 , candidate advertising messages are generated by inputting information about a product or service into a trained model previously generated by a machine learning algorithm. The trained model of this embodiment is a so-called generative AI that learns from past excellent Japanese advertising messages (catchphrases) and is adjusted for the purpose of developing Japanese creative writing. There are one or more candidate advertising messages. The information about the product or service includes, for example, at least one of information representing the product or service category and information about the consumer attributes of the product or service. The category information may be, for example, information such as food, clothing, and cosmetics, or more detailed classification information. The consumer information may be, for example, information such as men in their twenties and women in their twenties, or more detailed classification information.

[0012] Furthermore, in this embodiment, the following four pieces of information are also input into the trained model as information about the product or service to generate candidates for advertising messages.

[0013] (1) In this embodiment, information about the target consumer's usual psychological attributes and behavioral attributes is input to the trained model. For example, information about the psychological and behavioral attributes of consumers who purchase the target product or service, such as what personality traits they have and what values ​​they value (such as valuing family or friends), is input to the trained model. For example, information such as "the target consumer is a housewife who values ​​her family and puts her children first" and "she is concerned about rough skin" is input to the trained model. This makes it possible to generate candidate advertising messages for consumers who have the corresponding psychological and behavioral attributes.

[0014] (2) In this embodiment, the trained model is input with information about what kind of attitude or behavior change will occur in target consumers when they use a target product or service (or brand), as well as consumer tendencies that provide evidence that such change is possible. For example, information such as "my skin becomes smoother" or "I wake up feeling refreshed" is input to the trained model. This makes it possible to generate candidate advertising messages that will resonate with consumers who wish to change their attitude or behavior.

[0015] (3) In this embodiment, information regarding the tone that a product or service (or brand) aims for is input to the trained model. Specifically, information such as whether the target product or service (or brand) is assertive or empathetic, bright or dark, emotional or impersonal, etc. is input to the trained model. This generates candidate advertising messages that are in line with the intentions of the client (specifically, the brand image), which will be described later.

[0016] (4) In this embodiment, information such as the current status of the category to which the product or service belongs, the current position of the product or service, and the customer's future plans in Japan is input into the trained model. This generates candidate advertising messages that correspond to the current status of the product or service and the customer's intentions.

[0017] In the above-mentioned literature 1, emphasis is placed only on the functional value of a product or service. For example, only functional information such as product size and taste is input into the trained estimation model as evaluation items for the product. In contrast, in the present embodiment, by inputting the above-mentioned information into the trained model, candidate advertising messages that emphasize emotional value are generated, thereby generating candidate advertising messages that resonate more with consumers.

[0018] Next, as shown in FIG. 2, each of the multiple modifiers U1 creates a modified advertising message by modifying the candidate advertising message.

[0019] Next, as shown in FIG. 2, each of the plurality of voters U2 determines which of the plurality of revised advertising messages is appropriate for the target product or service, and votes for the plurality of revised advertising messages.

[0020] A final promotional message for the target product or service is then determined based on the voting results for the revised promotional message.

[0021] For example, consider a case where a request (one case) is made to create an advertising message for a single product. In this case, for example, by inputting a prompt such as "Generate five catchphrases for cosmetic product A. Cosmetic product A is a cosmetic product for ____, and the target consumers are women in their 30s to 50s. The target consumers are housewives who care for their families and put their children first. The target consumers are also concerned about rough skin. The consumers can expect to achieve smooth skin by using cosmetic product A. The target tone for cosmetic product A is ____. Cosmetic product A's current market position is ____, but its future outlook is ____," into the trained model, five candidate advertising messages can be generated.

[0022] Next, for example, three modifiers each modify one of the five candidate advertising messages to create 15 modified advertising messages. Then, for example, 100 voters vote for the 15 modified advertising messages. For example, one voter votes for the modified advertising message that they think is best from the 15 modified advertising messages. The modified advertising message with the most votes is adopted as the final advertising message.

[0023] In addition, rewards are awarded in the above process. For example, the person who creates the revised advertising message that receives the most votes is awarded a predetermined amount of cash or points as a reward. Also, for example, the voters who voted for the revised advertising message that received the top three votes are awarded a predetermined amount of cash or points as a reward.

[0024] Various data obtained in the process of creating the final advertising message is stored in the blockchain held by the blockchain node 18. For example, the above-mentioned prompt, candidate advertising messages, revised advertising message, final advertising message, voting results, information on cash or points awarded to the person who made the correction, and information on cash or points awarded to the voter are stored in the blockchain. A detailed explanation is given below.

[0025] As shown in FIG. 1, the server 12 functionally comprises a candidate generating unit 120 , a correction result acquiring unit 122 , a voting result acquiring unit 124 , and an output unit 126 .

[0026] The server 12, the modification terminal 14, the voting terminal 16, and the blockchain node 18 can be realized, for example, by a computer 70 shown in FIG. 3 . The computer 70 includes a CPU 71, a memory 72 as a temporary storage area, and a non-volatile storage unit 73. The computer 70 also includes an input / output interface (I / F) 74 to which input / output devices (not shown) are connected, and a read / write (R / W) unit 75 that controls reading and writing of data from and to a recording medium. The computer 70 also includes a network I / F 76 that is connected to a network such as the Internet. The CPU 71, the memory 72, the storage unit 73, the input / output I / F 74, the R / W unit 75, and the network I / F 76 are connected to one another via a bus 77.

[0027] The storage unit 73 can be realized by a hard disk drive (HDD), a solid state drive (SSD), a flash memory, etc. The storage unit 73 as a storage medium stores a program for operating the computer 70. The CPU 71 reads the program from the storage unit 73, loads it into the memory 72, and sequentially executes the processes contained in the program.

[0028] <Operation of Advertising Message Generation System 10>

[0029] Next, the operation of the advertising message generation system 10 of this embodiment will be described. The various operations of the advertising message generation system 10 can be realized, for example, by a program that can be installed on a computer. Specifically, the server 12 can perform various operations related to the creation of advertising messages by installing the program in advance. The advertising message generation system 10 executes the sequence shown in FIG. 4.

[0030] In step S100, the order terminal 11 outputs a signal requesting the creation of an advertising message to the server 12. Specifically, an orderer operating the order terminal 11 requests the creation of an advertising message for a product or service for which the orderer wishes to create an advertising message. For example, the orderer operating the order terminal 11 requests the creation of an advertising message for the product or service from the server 12 by inputting various information into a predetermined input form displayed on the display unit (not shown) of the order terminal 11. For example, the input form includes information about the product or service (e.g., at least one of information representing the product or service category, information about the consumer of the product or service, information representing the psychological or behavioral attributes of the target consumers of the product or service, information representing the type of attitude or behavior change that will occur as a result of the consumer using the product or service, information representing the tone that the product or service is aiming for, and information representing the current situation or future outlook for the category to which the product or service belongs).

[0031] In step S102, the candidate generation unit 120 of the server 12 generates candidate advertising messages for the target product or service based on the message creation request output in step S100. Specifically, the candidate generation unit 120 of the server 12 acquires the received request to create a promotional message. Then, information about the target product or service is input to a trained model previously generated by machine learning. The trained model outputs candidate advertising messages for the target product or service. More specifically, the candidate generation unit 120 generates candidate advertising messages for the product or service by inputting a prompt corresponding to information about the product or service into a trained model previously generated for generating promotional messages. For example, the candidate generation unit 120 of the server 12 displays a "Create Promotional Message" button on the display unit (not shown) of the order terminal 11. In response to the purchaser pressing the "Create Promotional Message" button, the candidate generation unit 120 inputs a prompt corresponding to information about the product or service into a trained model previously generated for generating promotional messages, thereby generating candidate advertising messages for the product or service. As mentioned above, the pre-created trained model for generating advertising messages is, for example, a generative AI such as ChatGPT, which learns from excellent Japanese advertising messages (catchphrases) from the past and is pre-tuned for the purpose of developing Japanese creative writing.

[0032] In step S104, the candidate generator 120 of the server 12 presents the advertising message candidates generated in step S102 to the editing terminal 14. For example, the candidate generator 120 of the server 12 presents the advertising message candidates to the editing terminal 14 using a dedicated site (browser). For example, the candidate advertising message candidates are presented only to the editing terminal 14 of pre-registered editers who meet the conditions related to the requested project (for example, the attributes of the target consumer).

[0033] The candidate generation unit 120 may extract a corrector, a candidate corrector, a voter, or a candidate voter from among multiple candidates based on information about a product or service. For example, if a preliminary questionnaire shows that a certain product is liked, people who like that product may be extracted as correctors, candidate correctors, voters, or candidate voters.

[0034] Furthermore, the candidate generation unit 120 may extract a reviser, a candidate reviser, a voter, or a candidate voter, depending on the degree of similarity between attribute information indicating the attributes of each of the multiple candidates and information about the product or service. For example, if the consumer of the product or service is a man in his twenties, the candidate generation unit 120 may extract a reviser, a candidate reviser, a voter, or a candidate voter, who is also a man in his twenties, from the multiple candidates.

[0035] In step S106, the multiple editors operate their editing terminals 14 to edit the advertising message presented in step S104, thereby creating an amended advertising message.

[0036] In step S108, the correction result acquisition unit 122 of the server 12 acquires the plurality of corrected advertising messages created in step S106.

[0037] In step S110, the candidate generator 120 of the server 12 presents the revised advertising message obtained in step S108 to the voting terminal 16.

[0038] In step S112, the plurality of voters operate their voting terminals 16 to vote for the plurality of revised advertising messages presented in step S110.

[0039] In step S114, the voting result acquisition unit 124 of the server 12 acquires the voting results obtained in step S112. Based on the voting results, the voting result acquisition unit 124 of the server 12 determines the revised advertising message with the largest number of votes as the final advertising message for the product or service.

[0040] In step S114, the output unit 126 of the server 12 outputs the final advertising message determined in step S114 to the order terminal 11.

[0041] Next, the output unit 126 records information about the product or service, the number of votes received by the person who created the revised advertising message that was adopted as the final advertising message, and information about the product or service for which the final advertising message was created in the blockchain.

[0042] 5 is a diagram showing an example of information recorded in the blockchain held by the blockchain node 18. As shown in FIG. 5, the advertising message information includes information about the product or service, the final advertising message, information about the person who created the amended advertising message adopted as the final advertising message, and the number of votes for the amended message, which are stored in a block in the blockchain.

[0043] By using this information stored in the blockchain, it becomes possible to determine which modifier to ask to make corrections the next time an advertising message for another product or service is created. Furthermore, since the blockchain also records information about the product or service, such as the category or consumer, it becomes possible to determine which modifier to ask to make corrections for a particular product or service. Furthermore, by treating the various data stored in the blockchain as big data, it becomes possible to select an appropriate modifier or voter for each requested project. For example, notifications are sent to tens to hundreds of people who meet the conditions set for the requested project, and modifiers or voters are selected on a first-come, first-served basis.

[0044] Therefore, for example, the next time the server 12 creates a promotional message for a product or service, it may refer to the information recorded in the blockchain to select a modifier, thereby enabling matching between the product or service and the creator of the promotional message.

[0045] As described above, cash or points may be awarded to the modifier according to the number of votes received. In this case, for example, cash or point information is stored in the blockchain as modifier information. The points may be used within the system or may be used outside the system. This is expected to increase the motivation of the modifier, leading to the creation of more appropriate advertising messages. As described above, cash or points may also be awarded to the voter according to the ranking of the revised advertising message for which they voted.

[0046] The output unit 126 may also retrain the trained model based on information about the product or service for which the final promotional message was created and data recorded in the blockchain. As shown in Figure 5, the blockchain records information about the product or service and the final promotional message in association with each other. Therefore, by using this data to retrain the trained model, it is possible to generate better promotional message candidates.

[0047] As described above, the server of the advertising message generation system according to this embodiment generates candidate advertising messages for a product or service based on information about the product or service. The server then acquires multiple revised advertising messages, which are revisions to the candidate advertising messages and are each revised by a plurality of revisers. The server then acquires voting results for each of the multiple revised advertising messages and are each voting results by a plurality of voters. The server then outputs at least one revised advertising message from the multiple revised advertising messages as a final advertising message for the product or service based on the voting results for the multiple revised advertising messages. This allows for the acquisition of advertising messages that can be used to promote the product or service.

[0048] The present disclosure is not limited to the above-described embodiments, and various modifications and applications are possible within the scope of the gist of the present disclosure.

[0049] For example, although the present embodiment has been described with reference to an example in which candidate advertising messages are generated using a trained model, the present invention is not limited to this, and candidate advertising messages may be generated using other methods. For example, information about a product or service may be linked to candidate advertising messages and stored in a database in advance, and when information about the target product or service is received, candidate advertising messages corresponding to the information may be read from the database to generate candidate advertising messages.

[0050] The effects obtained by this embodiment are as follows.

[0051] As described above, the trained model used in this embodiment is a uniquely tuned trained model. Furthermore, this trained model has been adjusted by learning excellent Japanese catchphrases from the past, with the aim of developing creative Japanese writing. Therefore, this embodiment can generate more appropriate advertising message candidates.

[0052] In addition, this embodiment maintains the transparency of the advertising message creation process and the voting process, which contributes to preventing tampering. In particular, the voting process is performed before the advertising message is published, and this voting process allows for the creation of more appropriate advertising messages. This improves the reliability and security of advertising messages.

[0053] Furthermore, as in this embodiment, the actual target demographic of the product or service is extracted from within the community and participates as editors of the advertising message generated by the trained model, which is generative AI, which is an innovative approach unparalleled in the industry. By directly involving end consumers in the creation process of the advertising message, it becomes possible to create more effective and relatable marketing content.

[0054] Furthermore, in this embodiment, the efficiency and speed provided by the trained model and the creative insights that are unique to humans (and the target of the product or service) provided by the human community make it possible to write more efficiently and in large quantities for each target, while also having a self-cleaning function that detects errors and plagiarism made by the trained model.

[0055] Furthermore, in this embodiment, the use of blockchain can provide significant differentiation from conventional copywriting solutions in that it increases the transparency and reliability of the advertising message creation and evaluation process.

[0056] In addition, the points system keeps modifiers and voters motivated by awarding points that are recorded on the blockchain.

[0057] Furthermore, although the present specification has been described as an embodiment in which the program is pre-installed, the program may also be provided by being stored on a computer-readable recording medium.

[0058] The following supplementary notes are disclosed. (Supplementary Note 1) A server comprising: a candidate generation unit that generates candidates for advertising messages for a product or service based on information about the product or service; a correction result acquisition unit that acquires a plurality of revised advertising messages, which are results of amendments to the candidate advertising messages and which are amended by each of a plurality of modifiers; and a voting result acquisition unit that acquires a voting result for each of the plurality of amended advertising messages and which is a voting result by each of a plurality of voters. (Supplementary Note 2) The server according to claim 1, wherein the voting result acquisition unit determines at least one revised advertising message from among the plurality of amended advertising messages as a final advertising message for the product or service based on the voting results for the plurality of amended advertising messages, and further comprises an output unit that outputs the final advertising message. (Supplementary Note 3) The server according to Supplementary Note 1 or Supplementary Note 2, wherein the information about the product or service includes at least one of: information representing the psychological or behavioral attributes of consumers who are targets of the product or service, information representing what kind of attitude or behavioral change will occur as a result of the consumers using the product or service, information representing the tone that the product or service is aiming for, and information representing the current situation or future prospects of the category to which the product or service belongs. (Supplementary Note 4) The server according to any one of Supplementary Notes 1 to 3, wherein the information about the product or service includes at least one of information representing the category of the product or service and information about consumers of the product or service. (Supplementary Note 5) The server according to any one of Supplementary Notes 1 to 4, wherein the candidate generation unit extracts the reviser, candidates for the reviser, the voter, or candidates for the voter from a plurality of candidates based on the information about the product or service. (Supplementary Note 6) The server described in any one of Supplementary Notes 1 to 5, wherein the candidate generation unit extracts the reviser, candidate revisers, the voter, or candidate voters based on the degree of similarity between attribute information indicating the attributes of each of the plurality of candidates and information related to the product or service.(Supplementary Note 7) The server according to any one of Supplements 1 to 6, wherein the voting result acquisition unit determines the revised advertising message with the largest number of votes as the final advertising message for the product or service based on the voting result. (Supplementary Note 8) The server according to any one of Supplements 1 to 7, wherein the candidate generation unit generates candidates for advertising messages for the product or service by inputting a prompt corresponding to information about the product or service into a pre-trained model for generating advertising messages. (Supplementary Note 9) The server according to any one of Supplements 1 to 8, wherein the number of votes of the modifier who created the revised advertising message adopted as the final advertising message and information about the product or service for which the final advertising message was created are recorded in the blockchain. (Supplementary Note 10) The server according to any one of Supplements 1 to 9, wherein the pre-trained model for generating advertising messages is re-trained based on information about the product or service for which the final advertising message was created and the data recorded in the blockchain. (Supplementary Note 11) A method for a computer to execute the following processes: generating candidate advertising messages for products or services based on information about the products or services, obtaining a plurality of revised advertising messages that are the results of amendments to the candidate advertising messages and that have been amended by each of a plurality of modifiers, and obtaining voting results for each of the plurality of revised advertising messages and that have been voted on by each of a plurality of voters. (Supplementary Note 12) A program for causing a computer to execute the following processes: generating candidate advertising messages for products or services based on information about the products or services, obtaining a plurality of revised advertising messages that are the results of amendments to the candidate advertising messages and that have been amended by each of a plurality of modifiers, and obtaining voting results for each of the plurality of revised advertising messages and that have been voted on by each of a plurality of voters.

[0059] The disclosure of Japanese Patent Application No. 2024-000672, filed on January 5, 2024, is incorporated herein by reference in its entirety. All documents, patent applications, and technical standards mentioned herein are incorporated herein by reference to the same extent as if each individual document, patent application, and technical standard was specifically and individually indicated to be incorporated by reference.

Claims

1. A server comprising: a candidate generation unit that generates candidates for an advertising message for a product or service based on information about the product or service; a correction result acquisition unit that acquires correction results for the candidates for the advertising message and a plurality of corrected advertising messages corrected by each of a plurality of corrector; a voting result acquisition unit that acquires voting results for each of the plurality of corrected advertising messages by each of a plurality of voters, and determines a final advertising message for the product or service based on the voting results for the plurality of corrected advertising messages; and an output unit that outputs the final advertising message.

2. The server according to claim 1, wherein the voting result acquisition unit determines at least one or more of the plurality of corrected advertising messages as the final advertising message for the product or service based on the voting results for the plurality of corrected advertising messages.

3. The information about the product or service includes at least one of information representing a psychological attribute or a behavioral attribute of a consumer targeted by the product or service, information representing what attitude change or behavior change will occur when the consumer uses the product or service, information representing a tone aimed at by the product or service, and information representing the current situation or future prospect of a category to which the product or service belongs. The server according to claim 1 or 2.

4. The information about the product or service includes at least one of information representing a category of the product or service and information about a consumer of the product or service. The server according to claim 1 or 2.

5. The candidate generation unit extracts the corrector, candidates for the corrector, the voter, or candidates for the voter from among a plurality of candidates based on information about the product or service. The server according to claim 1 or 2.

6. The candidate generation unit extracts the corrector, candidates for the corrector, the voter, or candidates for the voter according to the degree of similarity between attribute information indicating attributes of each of the plurality of candidates and the information about the product or service. The server according to claim 5.

7. The server according to claim 1 or claim 2, wherein the vote result acquisition unit determines, based on the vote result, the modified advertising message with the largest number of votes as the final advertising message for the product or service.

8. The server according to claim 1 or claim 2, wherein the candidate generation unit generates candidates for the advertising message of the product or service by inputting a prompt corresponding to information on the product or service into a pre-created learned model for generating the advertising message.

9. The server according to claim 2, wherein the number of votes of the modifier who created the modified advertising message adopted as the final advertising message and the information on the product or service for which the final advertising message was created are recorded on the blockchain.

10. The server according to claim 8, wherein the pre-created learned model for generating the advertising message is re-learned based on the information on the product or service for which the final advertising message was created and the data recorded on the blockchain.

11. A method for a computer to execute a process of generating candidates for the advertising message of a product or service based on information on the product or service, obtaining a plurality of modified advertising messages that are the modification results for the candidates of the advertising message and are modified by each of a plurality of modifiers, obtaining the vote results for each of the plurality of modified advertising messages and the vote results by each of a plurality of voters, determining the final advertising message of the product or service based on the vote results for the plurality of modified advertising messages, and outputting the final advertising message.

12. A program for causing a computer to execute a process of generating candidates for advertising messages for products or services based on information about the products or services, obtaining a plurality of corrected advertising messages that are correction results for the candidates for the advertising messages and are corrected by each of a plurality of correctors, obtaining voting results for each of the plurality of corrected advertising messages and the voting results by each of a plurality of voters, determining a final advertising message for the product or service based on the voting results for the plurality of corrected advertising messages, and outputting the final advertising message.

Images