Advertising management systems and programs

Abstract

We provide an advertising management system and program that enables highly accurate ad delivery tailored to users' lifestyles and individual circumstances. The advertising management system according to the present invention displays advertisements tailored to the user for specific products that are the target of advertising. The advertising management system is characterized by comprising: a customer segment data storage unit that stores information about the user's customer segment; a data integration unit that integrates at least external data about the user to generate integrated data about the user; a customer segment estimation unit that analyzes the integrated data and estimates the customer segment category of the user based on the customer segment data stored in the customer segment data storage unit; an advertising presentation unit that dynamically optimizes and presents the content, display format, or display timing of advertisements based on advertising scenarios associated with the estimated customer segment category; and a feedback unit that monitors the user's reaction to the advertisement presented by the advertising presentation unit and updates the contents of the customer segment data storage unit based on the reaction obtained.

Description

Technical Field

[0001] The present invention relates to an advertisement management system and a program that estimate a lifestyle behavior category based on at least external data related to a user and optionally use operation data, and optimize an advertisement effect.

Background Art

[0002] In recent years, behavior targeting technologies for optimizing advertisements based on user behavior on websites have been widely used. For example, Patent Document 1 discloses a technology for acquiring a user's behavior history such as browsing and clicking, and presenting an advertisement based on a score calculated from the analysis result. Further, Patent Document 2 discloses a product management system that continuously improves the development and operation of a product based on a business strategy through requirement specification generation, automatic code generation, log analysis, questionnaire processing, causal relationship visualization (DAG, CLD generation), etc. Also, Patent Document 3 discloses a business strategy evaluation technology that evaluates the success possibility of a business strategy by a machine learning model using external environments and marketing elements such as PESTLE analysis, 3C analysis, and 4P analysis as inputs, and presents parameters to be improved.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, while the behavioral targeting system described in Patent Document 1 enables the presentation of advertisements tailored to user interests, it is limited to website behavior logs, and therefore does not support advanced advertising optimization that comprehensively utilizes diverse external data. Furthermore, the product management system described in Patent Document 2 is useful in automating product development and operational improvements based on business strategies, but it does not have a dynamic optimization mechanism that targets ad delivery itself. In addition, the business strategy evaluation technology described in Patent Document 3 numerically evaluates the likelihood of success of a business strategy and suggests directions for improvement, but it does not directly lead to ad presentation control that takes into account the behavior of individual users or the external environment. Therefore, beyond evaluation at the business strategy level and management at the product development level, there is still a need for technology that integrates diverse external data and user operation data at the advertising operation site and dynamically optimizes ad content, format, and timing in real time.

[0005] This invention was made to solve the above problems and aims to provide an advertising management system and method that enables advertising delivery adapted with high accuracy to the user's lifestyle and individual circumstances. [Means for solving the problem]

[0006] To solve the above problems, the advertising management system according to the present invention displays advertisements tailored to the user for specific products that are the target of advertising. The advertising management system is characterized by comprising: a customer segment data storage unit that stores information about the user's customer segment; a data integration unit that integrates at least external data about the user to generate integrated data about the user; a customer segment estimation unit that analyzes the integrated data and estimates the customer segment category of the user based on the customer segment data stored in the customer segment data storage unit; an advertisement presentation unit that dynamically optimizes and presents the content, display format, or display timing of an advertisement based on an advertising scenario associated with the estimated customer segment category; and a feedback unit that monitors the user's reaction to the advertisement presented by the advertisement presentation unit and updates the contents of the customer segment data storage unit based on the reaction obtained.

[0007] In this invention, the data integration unit may match and combine operation data from the user terminal with external data related to the user based on an individual identifier.

[0008] In this invention, operation data may be acquired based on video recording or real-time logs on the user terminal.

[0009] In this invention, the external data relating to the user may include at least one of health data, household data, and financial / guarantee asset data.

[0010] In this invention, the data integration unit may further integrate seasonal trend data to create integrated data.

[0011] In the present invention, the customer segment data storage unit stores at least one of health data, household data, and asset data as customer information for each individual customer, and classifies the customer segment by reducing the dimensionality of the customer information and clustering it.

[0012] In this invention, the customer segment data storage unit may store advertising scenarios for customers belonging to each customer segment.

[0013] In this invention, the advertising scenario may specify at least one of the ad size, color tone, contrast level, and ad type.

[0014] In this invention, the customer segment data storage unit may store a score indicating the value of each customer segment as an advertising target.

[0015] In this invention, the feedback unit may update the score of each customer segment in the customer segment data storage unit based on an evaluation index for advertising effectiveness.

[0016] In the present invention, the feedback unit may analyze reaction data for advertisement presentations accumulated over a predetermined period, identify axes important for positioning strategies in customer layer data based on the reaction data, and update by weighting scores for a plurality of customer layer classifications based on the axes.

[0017] In the present invention, the customer layer estimation unit may apply a dimensionality reduction algorithm to the integrated data, reduce the dimension of the feature amount, and then collate with the customer layer classifications clustered by the customer layer data storage unit to estimate the customer layer classification of the user.

[0018] A program according to another example of the present invention causes a computer to function as any one of the above advertisement management systems.

Advantages of the Invention

[0019] According to the present invention, it becomes possible to display advertisements according to the actual actions and situations of users, and it is possible to achieve a higher advertisement effect (for example, CVR, CTR, LTV, etc.) than in the prior art. Further, by introducing dimensionality reduction technology, complex operation data can be intuitively understood and visualized, enabling effective user classification and personalized advertisement distribution.

Brief Description of the Drawings

[0020] [Figure 1] It is a diagram showing an advertisement management system according to an embodiment of the present invention together with a user terminal and an external data source. [Figure 2] It is a diagram showing the configuration of an advertisement management system according to an embodiment of the present invention. [Figure 3] It is a functional block diagram of an advertisement management system according to an embodiment of the present invention. [Figure 4] It is an example in which the user clustering result after dimensionality reduction in the customer layer data storage unit is visualized. [Figure 5] It is a flowchart showing a series of processing procedures by an advertisement management system 1 according to an embodiment of the present invention.

Modes for Carrying Out the Invention

[0021] Hereinafter, the advertisement management system 1 according to an embodiment of the present invention will be described with reference to the drawings. Note that the embodiment described below is an example for specifically explaining the present invention, and the present invention is not limited to the embodiment. The advertisement management system 1 performs advertisement display tailored to the user who views the advertisement in order to improve the advertisement effect for a specific product to be advertised.

[0022] 〔Configuration of the System〕 FIG. 1 is a schematic diagram showing an advertisement management system 1 according to an embodiment of the present invention together with a user terminal 2 and an external data source 3 connected to the advertisement management system 1 via a network NW. The advertisement management system 1 integrates the operations of the user using the user terminal 2 and various information provided from the external data source 3, and realizes advertisement management for improving the advertisement effect.

[0023] In addition to health data, household budget data, and financial and guaranteed asset data, the external data source 3 can include various data that affect the effect of advertisement presentation, such as location information data, weather data, regional event information, purchase history data, analysis results of posts on social media, traffic congestion information, and search trend data. These external data may be obtained from open data published by third-party data providers, partner companies, or public institutions. For example, by using weather data, advertisements for indoor-related products can be emphasized during rainy days, and by using event information, advertisements for related products can be intensively distributed during the period when a regional festival is held. Flexible advertisement optimization based on external data becomes possible. Also, seasonal trend data is an example of the data obtained from the external data source 3 and contributes to formulating an advertisement strategy according to seasonal variations and social events. For example, by combining purchase history data and seasonal trends, it is possible to predict in advance the period when the demand for seasonal products increases and perform intensive advertisement distribution during that period. Also, by linking the analysis results of social media and location information data, it is possible to realize advertisement publication in response to products or events that are currently popular in the region.

[0024] [Hardware configuration of advertising management system 1] Figure 2 is a schematic diagram showing the hardware configuration of the advertising management system 1. The advertising management system 1 is implemented, for example, as a computer. Specifically, the advertising management system 1 comprises a processor 101, RAM 102, HDD 103, graphics processing unit 104, input interface 105, and network interface 106. Although Figure 2 shows an example where the advertising management system 1 is implemented as a so-called standalone type on a single computer, the advertising management system 1 can also be implemented in a manner in which multiple computers (for example, one server computer and multiple client computers connected via a LAN) cooperate with each other via a network line such as a LAN. It may also be configured to perform distributed processing using multiple virtual machines or containers on the cloud. A configuration in which part of the processing is performed on edge terminals and privacy information is processed within the terminal is also possible.

[0025] The advertising management system 1 is controlled as a whole by a processor 101. The processor 101 may be a multiprocessor. The processor 101 may be, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), a GPU (Graphics Processing Unit), or a PLD (Programmable Logic Device). Alternatively, the processor 101 may be a combination of two or more elements from among the CPU, MPU, DSP, ASIC, and PLD.

[0026] The RAM (Random Access Memory) 102 is used as the main memory of the advertising management system 1. At least a portion of the OS (Operating System) program and application programs to be executed by the processor 101 are temporarily stored in the RAM 102. Additionally, various data necessary for processing by the processor 101 are stored in the RAM 102.

[0027] The HDD (Hard Disk Drive) 103 is used as an auxiliary storage device for the advertising management system 1. The HDD 103 stores the OS program, application programs, and various data (e.g., operation logs, external database information, trained models, and cached processing result data). Other types of non-volatile storage devices, such as an SSD (Solid State Drive), can also be used as auxiliary storage devices.

[0028] A display device 104a is connected to the graphics processing unit 104. The graphics processing unit 104 displays images on the screen of the display device 104a according to instructions from the processor 101. The display device 104a can be a liquid crystal display or an organic EL (Electro-Luminescence) display, among others.

[0029] An input device 105a is connected to the input interface 105. The input interface 105 transmits signals output from the input device 105a to the processor 101. Examples of input devices 105a include keyboards and pointing devices. Examples of pointing devices include mice, touch panels, tablets, touchpads, and trackballs.

[0030] Administrators and advertising operators use the graphics processing unit 104 and input interface 105 to monitor processing status, check insights, set parameters, etc., on a GUI.

[0031] The network interface 106 enables communication with external devices via the network NW. For example, the network interface 106 is used to connect to a database on the cloud or an external analytics platform, enabling real-time data updates, sharing, and external API integration. Communication via the network interface 106 through the network NW may be wired or wireless. As shown in Figure 1, user terminals 2 and external data sources 3 are connected to the network NW, and the advertising management system 1 can communicate with them via the network NW.

[0032] [Functional blocks of the advertising management system 1] Next, the functions of the advertising management system 1, which is realized by the hardware configuration described above, will be explained. Figure 3 shows a functional block diagram of the advertising management system 1. Specifically, the advertising management system 1 comprises a customer segment data storage unit 10, an operation data acquisition unit 11, a data integration unit 12, a customer segment estimation unit 13, an advertising presentation unit 14, and a feedback unit 15. Each of these functional blocks is realized by the processor 101 in the hardware configuration of the advertising management system 1 described above executing programs stored in the RAM 102 and HDD 103.

[0033] The customer segment data storage unit 10 stores information about the customer segment of the products targeted for advertising. The customer segment data storage unit 10 stores information about each individual customer, associating it with health data, household data, asset data, etc., and performs dimensionality reduction on this large amount of customer information and clustering, i.e., classification into customer segment categories (hereinafter also referred to as clusters). When applying a dimensionality reduction algorithm, in the case of PCA, it is good to select the top principal components with a cumulative contribution rate of 80% or more, and in the case of t-SNE, it is good to set the perplexity value and learning rate appropriately (for example, perplexity value of 30, learning rate of 200, etc.). Classification into customer segment categories is good to be done automatically by a clustering algorithm (for example, k-means, DBSCAN, etc.). In addition to / instead of automatic classification, the administrator may manually classify the customer segments.

[0034] Furthermore, the information recorded in the customer data storage unit 10 undergoes preprocessing prior to clustering. Preprocessing may include, for example, missing value imputation (median, mean, or machine learning estimation), feature normalization (Z-score transformation, minimum / maximum normalization, etc.), and removal of unnecessary features.

[0035] The customer segment data storage unit 10 stores effective advertising scenarios (i.e., information on advertising types that are likely to elicit purchasing behavior from customers in that segment) for each cluster. The advertising scenarios specify, for example, the size of the advertisement, its color tone, the strength of its contrast, and the type of advertisement (e.g., text, still image, or video). The advertising scenarios include display parameters corresponding to the attribute information (age group, gender, visual characteristics, etc.) and preference information (past purchasing trends, interest categories, etc.) of the users constituting the cluster. For example, for the elderly cluster, the font size may be increased and the contrast between the background and text color may be strengthened to improve visibility, while for the younger cluster, the advertisement may be configured to prioritize design, presenting advertisements with highly saturated colors and animation effects. Furthermore, for clusters with specific hobbies and preferences (outdoors, music, fashion, etc.), applying color themes and font styles related to those preferences can enhance the appeal of the advertisement. The advertising scenarios may also be automatically generated or updated by a machine learning model (e.g., gradient boosting, neural network, etc.) based on past advertising presentation history and response data. This allows us to continuously provide advertising scenarios optimized in response to changes in timing and market conditions.

[0036] Furthermore, the customer data storage unit 10 stores a score for each customer cluster. This score represents the value of each cluster as an advertising target (in other words, an indicator of the likelihood that advertising will increase sales of the target product). The score for each customer cluster is used for advertising budget allocation and ad slot selection. For example, 70% of the total advertising budget can be allocated to the clusters with the highest scores, while clusters with scores below a threshold can be excluded from ad delivery. In addition, the advertising strategy can be continuously optimized by relearning the cluster configuration when score fluctuations exceed a certain range.

[0037] Figure 4 is an example of a visualization showing the clustering results after dimensionality reduction of the customer segment data stored in the customer segment data storage unit 10, mapped in three dimensions. This allows for a visual understanding of how user groups are distributed and which clusters are showing a high response to advertisements.

[0038] The operation data acquisition unit 11 extracts user actions, including cursor movement, clicks, taps, scrolling, and time spent on a surface, from real-time or recorded data.

[0039] The data integration unit 12 matches customer information such as external health data, household data, and asset data based on individual identification information. When matching customer information, the data integration unit 12 performs privacy protection processing so that individual users cannot be identified. Specifically, when generating individual identification information, identification information such as user IDs should be converted using a hash function (e.g., SHA-256) or tokenization processing to make it difficult to reverse-engineer.

[0040] The data integration unit 12 further combines customer information with seasonal trend data. The integration process is performed in a data cleanroom environment. For anonymization, in addition to hashing, it is desirable to statistically reduce the risk of identifying specific individuals by using k-anonymization or differential privacy techniques. Access control should be implemented using role-based access control (RBAC) or attribute-based access control (ABAC), and the available attributes should be automatically restricted according to the usage permission conditions for each data provider. The data cleanroom has an access control mechanism that prohibits the export of raw data when linking data with external businesses and advertisers, and outputs only the integrated statistical information and analysis results externally. Alternatively, access rights may be controlled for each external data provider, and a configuration that uses only permitted data attributes may be implemented. This enables data integration that complies with laws and regulations such as the Personal Information Protection Act. Seasonal trend data includes seasonal events (e.g., year-end sales, summer sales), weather information, holiday calendars, etc., and by adding these to the integrated data, they can be used for dynamic optimization of advertising timing and content.

[0041] The data integration unit 12 then further integrates the user's operation data acquired by the operation data acquisition unit 11, estimates the user's lifestyle behavior category, and outputs it as integrated data. In recent years, in internet advertising, the acquisition and use of information that could potentially identify an individual is strictly restricted from the perspective of protecting personal information. However, by integrating customer information in a data cleanroom, the data integration unit 12 can protect personal information while obtaining integrated data that includes lifestyle behavior categories useful for estimating the user's customer base.

[0042] The customer segment estimation unit 13 identifies customer segment clusters closest to the user based on the integrated data processed by the data integration unit and the customer segment data stored in the customer segment data storage unit 10. Specifically, it is preferable to compress the data's dimensionality by applying a dimensionality reduction algorithm such as principal component analysis (PCA), singular value decomposition (SVD), or t-SNE to the integrated data. When applying a dimensionality reduction algorithm, it is preferable to compress the data to the same dimensionality as when clustering customer segments in the customer segment data storage unit 10. Then, among the clusters of each customer segment clustered in the customer segment data storage unit 10, the one closest to the combined data after dimensionality reduction is identified as the customer segment cluster closest to the user. Specifically, it is preferable to determine the cluster to which the integrated data mapped to a low-dimensional space belongs by calculating the distance from the centroid of an existing cluster. Euclidean distance, cosine similarity, or Mahalanobis distance can be used for distance measurement. This makes it possible to identify the customer segment cluster that most closely approximates the user with high accuracy while considering the nonlinearity of the data structure.

[0043] The ad presentation unit 14 dynamically presents advertisements tailored to the customer segment. Specifically, the ad presentation unit 14 retrieves advertising scenarios corresponding to customer segment clusters identified by the customer segment estimation unit 13 from the customer segment data storage unit 10, and presents advertisements to the user according to those advertising scenarios.

[0044] The ad presentation unit 14 has a function to dynamically optimize the content, display format, and display timing of the advertisements presented to the user. Specifically, the ad presentation unit 14 acquires an advertising scenario associated with the customer cluster identified by the customer cluster estimation unit 13, and when applying the advertising scenario, it may consider the user's access terminal type (PC, smartphone, tablet, etc.), communication environment (wired / wireless, line speed), access time, location information, past browsing history, and advertising response history.

[0045] For example, it is possible to control the display of low-brightness color schemes for access during nighttime hours and high-brightness color schemes for access during daytime hours. Furthermore, the system may be configured to prioritize lightweight still image ads when the communication speed is slow, and to display video ads in high-speed communication environments. These optimizations may be performed based on predefined rules, or by predicting conditions that will increase ad response rates using machine learning models.

[0046] The feedback unit 15 monitors user reactions to advertisements presented by the advertisement presentation unit 14 and updates the contents of the customer segment data storage unit 10 based on the reactions received. The feedback unit 15 updates the score for each customer segment cluster by combining multiple evaluation metrics acquired after the advertisement presentation.

[0047] Evaluation metrics include, for example, CTR (click-through rate), CVR (conversion rate), LTV (lifetime value), and bounce rate. As an example, score S is calculated using the following formula (1). S = w1 × CTR + w2 × CVR + w3 × LTV - w4 × churn rate ... (1)

[0048] Here, w1 to w4 are weighting coefficients set by the advertiser or system administrator, and can be changed according to the importance placed on advertising effectiveness. Score updates may be performed daily or weekly, or at the end of a specific advertising campaign or when a new advertising scenario is introduced. In this way, the evaluation of customer clusters is updated based on actual advertising performance and reflected in subsequent advertising strategies. The updated score history is saved chronologically, and the system may be configured to automatically trigger a customer cluster retraining process if the score fluctuation over a certain period exceeds a predetermined threshold.

[0049] Furthermore, the feedback unit 15 may analyze response data to ad presentations accumulated over a predetermined period and identify strategically important axes in the customer segment data based on that response data. For example, if the analysis reveals that age strongly influences ad responses, age can be identified as an important axis, and if younger age groups show higher responses, the score can be updated with weights according to age, or if age groups closer to their 40s show higher responses, the score can be updated to be higher the closer one is to that age group. In this way, strategically important axes can be identified from ad responses, and the scores of multiple customer segment groups can be updated with weights based on those axes, thereby enabling highly accurate optimization of ad delivery while aligning it with strategic importance in business strategy.

[0050] As shown in Figure 5, the advertising management system 1 of this embodiment performs a series of processes from acquiring user data to ad delivery, response acquisition, and score updating / feedback in the following order: acquisition of user operation data and external data (step S01), integration processing by the data integration unit 12 (step S02), identification of clusters (customer segmentation) by the customer segment estimation unit 13 (step S03), ad presentation by the ad presentation unit 14 (step S04), and score updating by the feedback unit 15 (step S05). Based on the updated score, the advertising strategy is then appropriately revised.

[0051] According to the embodiments of the present invention described above, by integrating diverse external data related to the user's customer base with operational data acquired as needed, and applying analytical methods such as dimensionality reduction and clustering, it becomes possible to estimate customer base classifications with high accuracy. As a result, it is possible to dynamically present advertising scenarios optimized for each customer base, and advertising effectiveness indicators such as click-through rate (CTR), conversion rate (CVR), and lifetime value (LTV) can be significantly improved compared to conventional technologies.

[0052] Furthermore, the feedback unit analyzes response data after ad presentation and automatically updates customer cluster scores, enabling continuous and autonomous optimization of the ad delivery strategy. This allows for more efficient ad operations, reduced operating costs, and more effective allocation of the advertising budget.

[0053] Furthermore, by performing privacy protection processing using a data cleanroom in the data integration department, it is possible to achieve high-value advertising targeting while complying with laws and regulations such as the Personal Information Protection Act. This makes it possible to provide an advertising management system that balances legal compliance with improved advertising effectiveness.

[0054] The advertising management system of the present invention is applicable to various media and distribution methods, including internet advertising, in-app advertising, and digital signage advertising, and contributes to the advancement of advertising strategies for a wide range of products and services.

[0055] Although the embodiments described above are examples, the present invention is not limited to these examples. Furthermore, any additions, deletions, or design modifications of components to the aforementioned embodiments, or any combination of features from each embodiment as appropriate by those skilled in the art, are also included within the scope of the present invention, as long as they retain the essence of the present invention. [Explanation of Symbols]

[0056] 1. Advertising Management System 2 User terminals 3. External data sources 10 Customer Segment Data Storage Unit 11 Operation Data Acquisition Unit 12 Data Integration Department 13. Customer Segmentation Department 14. Advertising Section 15 Feedback Section

Claims

1. An advertising management system that displays advertisements tailored to the user for specific products targeted for advertising, A customer segment data storage unit that classifies and stores information about the user's customer segment into customer segment categories, At a minimum, a data integration unit that generates integrated data about a user by matching and combining operation data extracted from video recordings or real-time logs on the user terminal with external data about the user, A customer segment estimation unit analyzes the integrated data and estimates the customer segment classification of the user based on the customer segment data stored in the customer segment data storage unit. An advertising display unit controls advertising budget allocation or ad slot selection based on scores associated with estimated customer segments, acquires advertising scenarios associated with estimated customer segments, and dynamically optimizes the content, display format, or display timing of advertisements according to the user's situation in accordance with those advertising scenarios, thereby presenting optimized advertisements. A feedback unit analyzes user response data accumulated over a predetermined period to advertisements presented by the advertising display unit, identifies axes important for positioning strategy in the customer segment data, and updates the scores of multiple customer segment categories by weighting them based on those axes. Equipped with, The customer segment data storage unit stores effective advertising scenarios and scores for each customer segment category. The advertising management system is characterized in that the aforementioned advertising scenario is updated by a machine learning model based on past advertising presentation history and response data.

2. The advertising management system according to claim 1, characterized in that the data integration unit matches and combines operation data from the user terminal with the external data relating to the user based on an individual identifier.

3. The advertising management system according to claim 1, characterized in that the external data relating to the user includes at least one of health data, household data, and financial / guarantee asset data.

4. The advertising management system according to claim 1, characterized in that the data integration unit further integrates seasonal trend data to obtain the integrated data.

5. The advertising management system according to claim 1, characterized in that the customer segment data storage unit stores at least one of health data, household data, and asset data as customer information for each individual customer, and classifies customer segments by reducing the dimensionality of the customer information and clustering it.

6. The advertising management system according to claim 5, characterized in that the customer segment data storage unit stores advertising scenarios for customers belonging to each customer segment.

7. The advertising management system according to claim 6, characterized in that the advertising scenario specifies at least one of the size of the advertisement, the color tone, the strength of the contrast, and the type of advertisement.

8. The advertising management system according to claim 5, characterized in that the customer segment data storage unit stores a score indicating the value of each customer segment as an advertising target.

9. The advertising management system according to claim 8, characterized in that the feedback unit updates the score of each customer segment in the customer segment data storage unit based on an advertising effectiveness evaluation index.

10. The advertising management system according to claim 5, characterized in that the customer segment estimation unit applies a dimensionality reduction algorithm to the integrated data to reduce the dimensionality of the feature quantities, then compares it with the customer segment classifications clustered by the customer segment data storage unit to estimate the customer segment classification of the user.

11. A program that causes a computer to function as an advertising management system according to any one of claims 1 to 10.

Images