Advertisement management device, advertisement management method, and advertisement management program

Abstract

Generate new ad assets optimized for performance. [Solution] The advertising management system acquires performance information for advertisements of specified products from advertising distribution media, as well as actual customer purchase behavior information from the customer relationship management server. Furthermore, it extracts messages such as catchphrases from eye-catching images on landing pages through image analysis. The system compares and analyzes the relationship between performance information and purchase behavior information, and extracts advertising materials where performance is skewed, such as when media metrics are high but actual business results are not. Then, based on the extracted advertising materials, acquired messages, and purchase behavior information, it generates new advertising materials with optimized performance by providing appropriate prompts to the generation AI, etc.

Description

[Technical Field]

[0001] The present invention relates to an advertisement management apparatus, an advertisement management method, and an advertisement management program. [Background Art]

[0002] In recent years, distribution of digital advertisements using the web and SNS (Social Networking Service) has been widely performed. For such advertisement distribution, technologies for generating and improving advertisement materials based on past advertisement performance data have been proposed.

[0003] For example, Patent Document 1 discloses a system that collects past advertisement performance data and extracts specific improvement points for advertisement creatives based on results of AI analysis. Further, Patent Document 2 discloses an image generation apparatus that acquires input information such as a landing page address, generates a catch copy and an image generation prompt using a large language model, and automatically generates a banner image from a background image and the catch copy. [Prior Art Documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Unexamined Patent Application Publication No. 2026-36339 [Patent Document 2] Japanese Unexamined Patent Application Publication No. 2026-70850 [Summary of the Invention] [Problems to be Solved by the Invention]

[0005] However, while conventional technologies as described above allow for improvements based on performance data such as click-through rates on advertising media, they have the challenge of making banner images that take into account actual customer purchasing behavior. Furthermore, although catchphrases included in images on landing pages to which users are directed from banner images influence customer purchasing behavior, banner image generation has not adequately considered these factors.

[0006] Therefore, while measures to improve the performance of advertising materials may increase click-through rates and view rates, they do not necessarily translate into actual customer purchasing behavior, and have not led to improvements in advertising materials and banners that achieve sufficient performance. [Means for solving the problem]

[0007] An advertising management device in one aspect of the present invention includes: an outcome information acquisition unit that acquires outcome information of advertising materials from an advertising distribution medium; a purchase behavior information acquisition unit that acquires customer purchase behavior information about products from a customer relationship information management server; a message acquisition unit that accesses a website introducing products and extracts messages from images on the website by image analysis; a performance evaluation unit that analyzes the relationship between outcome information and purchase behavior information and evaluates the performance of advertising materials; and an advertising material generation unit that generates new advertising materials with optimized performance based on the advertising materials, messages, outcome information, and purchase behavior information for advertising materials that have been evaluated as having low performance. [Effects of the Invention]

[0008] According to the present invention, it is possible to generate advertising materials with high advertising effectiveness. [Brief explanation of the drawing]

[0009] [Figure 1] This is a block diagram showing the configuration of an advertising management system. [Figure 2]This is a block diagram showing the configuration of the advertising management device. [Figure 3] This is a chat screen with the advertising management device displayed on the user's terminal. [Figure 4] This is a flowchart showing the processing operation in the advertising management device. [Modes for carrying out the invention]

[0010] (Example 1) Hereafter, identical or equivalent components, members, processes, and signals shown in each drawing will be denoted by the same reference numerals, and redundant explanations will be omitted as appropriate. In addition, some components that are not explanatoryly important will be omitted from each drawing. The following embodiments are not intended to unduly limit the content of the disclosure as described in the claims, and not all components shown in the embodiments are necessarily essential components of the disclosure. Furthermore, each figure is a schematic diagram and is not necessarily a strict illustration.

[0011] (System Configuration) The advertising management system in this disclosure comprises an advertising management device 1, an advertising distribution medium 2, a customer relationship information management server 3, a web server 4, and a user terminal 5, all connected via a network (NW). Figure 1 is a block diagram showing the hardware configuration of the advertising management system. Figure 2 is a block diagram showing the functional configuration of the advertising management device 1.

[0012] (Configuration of advertising management device 1) The advertising management device 1 comprises a control unit 101, a storage unit 102, and a communication unit 103. The control unit 101 is composed of a processor, and the control unit 101 realizes each functional unit by reading application programs stored in the storage unit 103 and executing processing.

[0013] The communication unit 102 communicates with other computers via a network NW using wired or wireless communication standards. The network NW consists of a mobile communication network, wireless LAN, other short-range wireless communication, telephone lines, or a combination thereof, and the communication unit 102 is configured with an interface that can input and output signals to the network and communicate. The storage unit 103 is a storage device for permanently or temporarily storing data and programs, and consists of volatile memory, hard disks, SSDs, etc.

[0014] (Composition of advertising distribution medium 2) Ad distribution medium 2 provides web advertising by offering ad copy along with banners via SNS (Social Networking Services) and providing banners to other websites. Ad distribution medium 2 is an external server separate from ad management device 1. Examples of specific ad distribution mediums include various ad distribution platforms such as search engine marketing services and real-name SNS. In this embodiment, ad banners are used as an example to describe the ad materials to be distributed, but the ad materials of this embodiment also include content such as videos and audio.

[0015] (Configuration of Customer Relationship Information Management Server 3) The Customer Relationship Management Server 3 is a so-called CRM (Customer Relationship Management) server that manages customer attributes and purchasing behavior information. Examples of information that can be obtained include actual product purchase history, subscription retention rates, and sales profits.

[0016] (Web server 4 configuration) Web Server 4 is a server that provides web pages (primarily landing pages) that introduce the target products. A landing page is a website that introduces a product, consisting of images such as eye-catching images and text. Eye-catching images are not just simple photographs or illustrations, but also incorporate messages such as catchy slogans that are integrated as part of the image design, playing an important role in product sales.

[0017] (Configuration of User Terminal 5) The user terminal 5 is a terminal used by a user who places advertisements, and is constituted by a smartphone, a personal computer, or the like. The user terminal 5 communicates with the advertisement management device 1 via the network NW and the communication unit 103. Furthermore, the user terminal 5 communicates with the advertisement management device 1 in an interactive manner via a server that provides communication services such as chat and SNS. This allows the user to give instructions to the advertisement management device and confirm outputs through an interactive interface using natural language.

[0018] Next, each functional unit of the control unit 101 will be described. The functional units are composed of a result information acquisition unit 1011, a purchase behavior information acquisition unit 1012, a message acquisition unit 1013, a performance evaluation unit 1014, an advertisement material generation unit 1015, and an interaction unit 1016.

[0019] (Result Information Acquisition Unit 1011) The result information acquisition unit 1011 acquires, for each advertisement material, result information regarding an advertisement for a predetermined product or service from the advertisement distribution medium 2. Specific examples of result information held by the advertisement distribution medium 2 include the following. First, there are indicators related to display and clicks (visibility and guidance). This includes the number of impressions, which is the total number of times the advertisement is displayed to users; the number of clicks, which is the total number of times the advertisement is actually clicked; the click-through rate (CTR), which indicates the ratio of the number of clicks to the number of times the advertisement is displayed; and visibility (such as the visibility rate and the viewing distance for outdoor advertisements).

[0020] Second, there are indicators related to results and specific actions (conversions). This includes the number of conversions (CV), which indicates the number of times specific actions defined as results by the advertiser have occurred, such as product purchases, sample applications, or requests for materials; and the conversion rate (CVR), which indicates the ratio of the number of conversions to the number of impressions.

[0021] Thirdly, there are cost-related metrics. These include acquisition costs (Cost) for acquiring and delivering ad space, cost per mille (CPM), which is the cost per 1,000 impressions, cost per click (CPC), which is the cost per click, and cost per action (CPA), which is the cost per conversion.

[0022] Fourth, there are metrics related to user reactions on social media (engagement). This includes the engagement rate, which shows the percentage of users who respond to an advertisement, and specific engagement actions such as the number of "likes" and shares on social media posts.

[0023] (Purchasing behavior information acquisition unit 1012) The purchasing behavior information acquisition unit 1012 acquires customer purchasing behavior information about products from the customer relationship information management server 3. The purchasing behavior information includes actual data that shows the specific purchasing behavior of customers, such as actual product purchase history, subscription continuation rate, conversion rate to paid plans, average purchase price, and sales profit.

[0024] (Message acquisition unit 1013) The message acquisition unit 1013 acquires information from the landing page, which is a website about the product provided by the web server 4. Specifically, in addition to the text information within the landing page, it acquires images from the landing page and extracts messages such as catchphrases contained in those images through image analysis. The image should preferably be an eye-catching image, that is, a large image at the top of the page that attracts the visitor's attention on the website, and generally the text of the catchphrase that the provider wants to convey is integrated into the image design. In this embodiment, the message acquisition unit 1013 processes eye-catching images, but the message acquisition unit 1013 may process any image within the landing page. Furthermore, the website accessed by the message acquisition unit 1013 is not limited to a landing page, as long as it is a website about the product.

[0025] Eye-catching images play an extremely important role in product sales. Users are attracted to the visuals of an advertisement and click on it, and within the first few seconds of opening the landing page, they instantly decide whether to "continue reading or leave." If they feel the same worldview and image as the advertisement the moment they open the page, they will feel at ease and continue reading, but if the image is different, it will be a reason for them to leave. Therefore, since the combination of image design and the catchy slogan incorporated into the image design determines the first impression, capturing the message written in the eye-catching image is important in ensuring consistency with the advertisement and conveying specific benefits that cannot be conveyed by the image alone.

[0026] (Performance Evaluation Department 1014) The performance evaluation unit 1014 compares and analyzes the relationship between the performance information of advertising materials acquired by the performance information acquisition unit 1011 and the purchase behavior information acquired by the purchase behavior information acquisition unit 1012, and extracts advertising materials with low performance from the performance information and purchase behavior information under predetermined conditions. There are two criteria for determining low performance: 1) both the performance information and the purchase behavior information are low (below a predetermined threshold), and 2) there is a difference between the performance information and the purchase behavior information, where the performance information is high but the purchase behavior information is low (hereinafter this state will be called performance discrepancy).

[0027] One concrete example of a performance discrepancy is a situation where the "click-through rate (CTR)" is high, but the actual "purchase rate" and "retention rate" are low. This refers to a situation where, although the CTR is high on the advertising medium and many users are directed to the landing page, the actual purchase rate of the product is extremely low in the real data on the customer relationship information management server 3, or where customers only make an initial trial purchase and do not become regular customers. Specifically, one example is a situation where the click-through rate obtained from the advertising medium is above a predetermined first threshold of 5.0% or more, and many users are directed to the landing page (high performance on the medium), but the actual purchase rate of the product in the real data on the customer relationship information management server 3 is below a predetermined second threshold of less than 0.5%, or where customers only make an initial trial purchase and do not become regular customers.

[0028] Secondly, there is a situation where there are many "conversions (CV)" on the media, but actual "sales profits" are low. This refers to a situation where, although many conversions such as "requests for information" and "free membership registrations" occur in the advertising media reports, when compared with the actual data from the customer relationship information management server 3, the conversion rate to paid plans is low or the average purchase price is low, and it does not contribute to actual business profits. Specifically, one example is a situation where, even though the conversion rate for "requests for information" and "free membership registrations" occurs at a predetermined third threshold of 2.0% or higher in the advertising media reports, when compared with the actual data from the customer relationship information management server 3, the conversion rate to paid plans is less than 5.0%, or the average purchase price per customer is less than 1,000 yen, and it falls below a predetermined fourth threshold, and does not contribute to actual business profits.

[0029] Thirdly, there is a situation where "engagement (likes, shares, etc.)" is high, but this is not accompanied by actual "purchase behavior." This refers to a situation in social media advertising where the advertisement itself becomes a topic of conversation and the engagement rate is very high, but it has no impact whatsoever on actual purchase history. Specifically, one example is a situation where, in social media advertising, the advertisement itself becomes a topic of conversation, and the engagement rate calculated from the number of "likes" and "shares" is very high, exceeding a predetermined threshold of 10.0% or more, but the actual data in the customer relationship information management server 3 shows that the purchase rate of products purchased via that advertisement is less than 0.1%, which is below a predetermined threshold of 6, meaning it has no impact whatsoever on actual purchase history.

[0030] Performance evaluation is performed by the performance evaluation unit 1014 by prompting the generating AI. The generating AI may be one provided in the control unit 101 of the advertising management device 1, or a model provided by an external server may be used via the network.

[0031] When the performance evaluation unit 1014 performs extraction and evaluation using the generating AI, it is preferable that the input prompt includes structured instruction information, including task definition information, task execution support information, and output item definition information. Specifically, the task definition information specifies roles and tasks such as "As a data analyst, compare the performance information and purchase behavior information of advertising media and extract advertising materials where performance deviates." The task execution support information includes specific examples that serve as criteria for judgment (for example, conditions such as "extract those with a CTR of 4.0% or higher but an actual purchase rate of less than 0.5%") and output instructions for the thought process, such as "infer the factors causing the deviation." Furthermore, the output item definition information includes a format specification indicating that the extracted advertising materials, the factors causing the deviation, and the comparison results of the indicators should be output in JSON format or the like. By providing such instruction information, the generating AI can not only perform simple numerical comparisons but also reliably output advanced analysis results, including the fundamental reasons for the deviation, in a predetermined format that can be used in subsequent processing.

[0032] (Advertising material generation unit 1015) The advertising material generation unit 1015 generates optimized new advertising materials for those advertising materials extracted by the performance evaluation unit 1014 that are judged to have low performance in terms of conversion information. Specifically, it generates new advertising materials (banner images, etc.) with optimized performance based on the advertising materials in question, the message of the eye-catching image acquired by the message acquisition unit 1013, and the purchase behavior information.

[0033] The advertising material generation unit 1015 provides the generation AI with structured "skill prompts" as input, which reflect expertise in advertising improvement, rather than just simple generation instructions. The skill prompts include, for example, the following information:

[0034] Firstly, there is task definition information (role specification). For example, this may include instructions such as, "You are a professional digital marketer and creator specializing in conversion improvement. Analyze the provided information and generate a new ad banner tagline and design proposal that maximizes actual purchasing behavior (purchase rate)."

[0035] Secondly, there is the input information. This includes data from existing advertising materials, performance information and purchase behavior information, and messages obtained from eye-catching images. Furthermore, the input information may also include factors of performance deviations of the advertising material output by the performance evaluation unit 1014, providing the generating AI with detailed and specific information that can serve as hints for improvement.

[0036] Thirdly, there are web advertising material generation skills (improvement logic and execution support information). This includes specific improvement rules such as "integrating the core message contained in the LP's eye-catching image into the banner's main copy (ensuring consistency)," "eliminating clickbait and clearly stating the product's specific benefits in a highly visible font to improve purchasing behavior data (stimulating purchase intent)," and "adopting an image character that takes the target audience into consideration."

[0037] Fourth, there is the output format. This may include formatting specifications such as "the main copy should be 20 characters or less, and the output should include the recommended background color and layout configuration."

[0038] The following three factors are the main reasons why the ad material generation unit 1015 generates superior ad materials with optimized performance:

[0039] The first is the analysis and implementation of "message consistency" between landing pages (LPs) and banner ads. One possible reason for a high click-through rate but low purchase rate is user drop-off due to a mismatch between the message conveyed by the ad banner and the LP. In this embodiment, by incorporating the message obtained from the LP's eye-catching image into banner generation, consistency in the message when the user transitions to the LP is ensured, leading to a reduction in the drop-off rate and an improvement in the purchase rate.

[0040] The second factor is the presentation of a real problem: "performance discrepancy." Instead of simply training the AI ​​on features that are likely to be clicked, by presenting the generating AI with discrepancy data such as "performing well on the platform but poorly in actual purchasing behavior," the AI ​​recognizes this as a problem of "clickbait" and can adjust its course to sincere and specific appeals that resonate with "customers who are genuinely interested in buying."

[0041] The third factor is the ability of the generative AI to "pre-train on advertising performance." Generative AIs such as large-scale language models pre-train on advertising best practices, such as what kind of catchphrases and designs are likely to lead to conversions. As in this embodiment, by providing specific and sophisticated context (skill prompts) that include data on the LP message, performance information, and customer relationship information, the expertise for advertising improvement latent within the AI ​​can be accurately extracted, resulting in output optimized for the target product.

[0042] It should be noted that the data structure of skill prompts and the factors and analytical methods for generating performance-optimized advertising materials described in the paragraph above are merely examples, and this disclosure is not limited to them. For example, skill prompts may be created based on a variety of other factors and analytical results, such as attribute information of the target audience of the advertisement, market trends and competitor data related to the target product, the characteristics of each advertising distribution medium (SNS, search engines, video distribution platforms, etc.), and environmental factors such as season and time of day. In this way, by flexibly combining arbitrary factors and data analysis results according to the target product and distribution situation to construct skill prompts, it becomes possible to generate new advertising materials that are optimized from a more multifaceted and advanced perspective.

[0043] (Dialogue Section 1016) The dialogue unit 1016 receives instructions from the user regarding advertising management via the chat screen 6 displayed on the user terminal 5's display unit and outputs the processing results. Specifically, the dialogue unit 1016 is configured to communicate with an external server providing the chat service via an API (Application Programming Interface). This allows the dialogue unit 1016 to provide input and output of processing content to the chat screen 6, which will be described later, enabling smooth dialogue with the user using natural language and concise messages.

[0044] (Operation of advertising management device 1) The following describes each process of generating advertising materials in the advertising management device 1. Figure 3 is a flowchart showing the operation of the process. Referring to Figure 3, the flow of advertising management processing performed in the advertising management device 1 according to this embodiment will be explained. Although the series of processes in this flowchart are executed by the control unit 101 of the advertising management device 1, for the sake of explanation, the functional unit that is the main component of the processing in each step will be explicitly indicated and explained.

[0045] The performance information acquisition unit 1011 acquires performance information for advertisements of specified products from the advertising distribution media 2 for each advertising material. The performance information acquired includes metrics related to display and clicks, such as the number of impressions and click-through rate (CTR); metrics related to results, such as the number of conversions (CV) and conversion rate (CVR); metrics related to costs and revenues, such as sales profit and cost per click (CPC); and metrics related to user reactions on social media, such as engagement rate and the number of "likes".

[0046] (Step S02) The purchasing behavior information acquisition unit 1012 acquires customer purchasing behavior information for the aforementioned products from the customer relationship information management server 3. The acquired purchasing behavior information includes actual data that directly relates to the customer's business results, such as actual product purchase history, subscription continuation rate, conversion rate to paid plans, average purchase price, and sales profit.

[0047] (Step S03) The message acquisition unit 1013 acquires information from the landing page of the product provided by the web server 4. Specifically, in addition to the text information within the landing page, it acquires an eye-catching image that determines the first impression in product sales, and acquires messages such as catchphrases contained in the eye-catching image through image analysis (OCR analysis, etc.). This extracts specific benefits that cannot be conveyed by the image alone, as well as information to ensure consistency between the advertisement and the landing page.

[0048] (Step S04) The performance evaluation unit 1014 compares and analyzes the relationship between the performance information of the advertising materials obtained in step S01 and the purchase behavior information obtained in step S02, and extracts advertising materials with low performance in relation to the performance information and purchase behavior information under predetermined conditions.

[0049] (Step S05) The advertising material generation unit 1015, for advertising materials extracted in step S04 that are judged to have low performance in terms of conversion information, generates new advertising materials with optimized performance using a generation AI or the like, based on the advertising materials, the message of the eye-catching image obtained in step S03, and the purchase behavior information.

[0050] At this point, the advertising material generation unit 1015 provides the generation AI with a structured "skill prompt" that includes task definition information, prerequisite knowledge and input information, web advertising material generation skills including specific improvement rules, and an output format. This ensures that the "message consistency" with the landing page is reflected, the real issue of "performance discrepancies" is presented, and the generation AI's "prior learning regarding advertising performance" is triggered, resulting in the automatic generation of optimized advertising material that leads to actual business results. With these steps, the series of advertising management processes is completed.

[0051] (Interface with user terminal) The processes for generating advertising materials in the advertising management device 1 are as described above. The user gives instructions to the advertising management device 1 for generating advertising materials and checks the results of the processing via the GUI displayed on the user terminal 5's screen. In this embodiment, an example is shown in which the advertising management device 1 and the user terminal 5 perform input and output in an interactive format via a chat application as the GUI.

[0052] The dialogue unit 1016 of the advertising management device 1 is configured to communicate with an external server providing a chat service via an API (Application Programming Interface). Meanwhile, the user interacts with the advertising management device 1 via the chat screen 6 shown in Figure 4.

[0053] (User interface using a chat screen) Figure 4 shows an example of an interactive chat screen 6 displayed on the user terminal 5. The user interacts with the advertising management device 1 through this chat screen 6, performing system operations and giving various instructions.

[0054] As shown in Figure 4, the chat screen 6 includes a message display unit 61, a thread display unit 62, and a message input unit 64. In the message display unit 61, the display of "User 1" indicates a message from the user using the user terminal 5, and the display of "meteo_banner" indicates a message to which the advertising management device 1 responded. The thread display unit 62 shows the dialogue thread between the user and the advertising management device 1 (for example, the "meteo_banner creation" thread).

[0055] The message from User 1 displayed at the top of the message display unit 61 is a request for the creation of a new banner image, and it specifies the URLs of the advertising distribution medium 2 where the advertisement will be placed and the web server 4 that will provide the landing page for the target product. These URLs may be entered by the user each time using the message input unit 64 as shown in the figure, or the URL information may be registered separately in the advertising management device 1 at the time of service contract, etc.

[0056] The dialogue unit 1016 of the advertising management device 1 transmits the progress of processing in response to the user's request message to an external server that provides a chat service, and displays information about optimized and superior advertising materials to the user as appropriate during the processing. In the final message from the message display unit 61, an advertising banner 65 generated by the advertising material generation unit 1015 is presented as an improvement suggestion.

[0057] The dialogue unit 1016 of the advertising management device 1 displays a button 63 (for example, "Approve," "Reconsider," etc.) on the chat screen 6 when user confirmation or approval is required, and requests confirmation from the user. Approval can be performed by pressing the button 63, or by responding by entering text into the message input unit 64. Examples of times when approval is required include when creating a new banner or when uploading an approved banner to the advertising distribution medium 2 and starting distribution.

[0058] When the user operates button 63 on the chat screen 6, the dialogue unit 1016 of the advertising management device 1 executes the process that was awaiting approval if the user's operation is approval. In the example in Figure 4, the user is asked to approve the advertising banner generated by the advertising management device 1 and to submit it to the advertising distribution medium 2. If the user approves, the control unit 101 of the advertising management device 1 executes the decision on the generated advertising banner and its submission to the advertising distribution medium 2. If the user's operation is to reconsider, the advertising material generation unit 1015 recreates the advertising banner and proposes it to the user again by the dialogue unit 1016.

[0059] Thus, in the system of this embodiment, users can intuitively give instructions to the advertising management device 1 by inputting natural language or concise messages into the chat screen 6, and complete the entire process, from approval processing via button 63 to uploading the created banner to the advertising distribution medium 2. In other words, users can obtain advertising materials optimized based on actual purchasing behavior and significantly improve advertising performance with extremely simple operations.

[0060] The present inventors have described the invention in detail based on embodiments above, but the present invention is not limited to the above embodiments. Furthermore, the above embodiments are described in detail for the purpose of explaining the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the described configurations. Some of the configurations of the above embodiments can be added, deleted, or replaced with other configurations.

[0061] Each of the above configurations, functions, and processing means may be implemented in hardware, in whole or in part, for example, by designing them as integrated circuits. Alternatively, each of the above configurations and functions may be implemented in software by having a processor interpret and execute programs that implement each function. Information such as programs, tables, and files that implement each function can be stored in memory, hard disks, SSDs, or other recording devices, or in recording media such as IC cards, SD cards, or DVDs. [Explanation of Symbols]

[0062] 1. Advertising management device, 2. Advertising distribution medium, 3. Customer relationship information management server, 4. Web server, 5. User terminal, 101. Control unit, 1011. Performance information acquisition unit, 1012. Purchase behavior information acquisition unit, 1013. Message acquisition unit, 1014. Performance evaluation unit, 1015. Advertising material generation unit, 102. Communication unit, 103. Storage unit, 6. Chat screen, 61. Message display unit, 62. Thread display unit, 63. Button, 64. Message input unit, 65. Generated advertising banner

Claims

1. A performance information acquisition unit that obtains performance information for advertising materials from advertising distribution media, A purchasing behavior information acquisition unit that acquires customer purchasing behavior information about products from a customer relationship information management server, A performance evaluation unit extracts advertising materials in which the actual performance value of the results information is higher than the first threshold, and the actual purchase behavior value of users who viewed the advertising material for the product is lower than the second threshold. An advertising material generation unit generates new advertising materials based on extracted advertising materials, performance information, and purchase behavior information. An advertising management device equipped with the following features.

2. A message acquisition unit that accesses a website introducing a product, performs image analysis on the images on the website, and extracts text that is integrally designed within the images. Furthermore, The advertising material generation unit generates new advertising material based on the extracted text. The advertising management device according to claim 1.

3. A dialogue unit that communicates with an external chat service provider server via API and receives instructions from the user in a dialogue format via a chat service executed on the user terminal. Furthermore, The advertising management device according to claim 1 or 2.

4. A performance information acquisition unit that obtains performance information on clicks on advertising materials from advertising distribution media, A message acquisition unit accesses a website introducing the product that is the target of the advertising material, analyzes the images on the website, and extracts text that is integrally designed within the images. An advertising material generation unit generates new advertising materials based on advertising materials, extracted text, and performance information. An advertising management device equipped with the following features.

5. The steps include obtaining click performance information for ad materials from the ad distribution platform, The process involves accessing a website that introduces the product being advertised, performing image analysis on the images on that website, and extracting text that is integrated into the design within the images. The steps include generating new advertising materials based on advertising materials, extracted text, and performance information, A computer-based method for managing advertisements.

6. The steps include obtaining click performance information for ad materials from the ad distribution platform, The process involves accessing a website that introduces the product being advertised, performing image analysis on the images on that website, and extracting text that is integrated into the design within the images. The steps include generating new advertising materials based on advertising materials, extracted text, and performance information, An advertising management program that causes a computer to execute an ad.

Images