KR102215635B1 - Method of providing marketing data and marketing data platform performing the same - Google Patents

Abstract

In a method for providing marketing information, a marketing information storage receives and stores purchase information from a purchase information providing system, a marketing information analysis system generates a marketing identifier indicating customer purchase propensity based on the purchase information stored in the marketing information storage, the marketing information storage receives and stores a marketing result from a marketing information using system that corresponds to the marketing identifier, and the marketing information analysis system performs machine learning for marketing recommendation information generation based on the marketing result stored in the marketing information storage. In addition, by the method, data integrity can be ensured through the blockchain-based marketing information storage, system security can be enhanced, information asymmetry can be eliminated, and it is possible to encourage participation of purchase information and marketing information using systems.

Description

Marketing information provision method and marketing information platform that performs it {METHOD OF PROVIDING MARKETING DATA AND MARKETING DATA PLATFORM PERFORMING THE SAME}

The present invention relates to a method for providing marketing information, and more particularly, to a method for providing marketing information and a marketing information platform for performing the same.

As Internet use using mobile devices increases, companies frequently collect personal activity details in an online environment, and directly utilize the collected data or provide it to a third party for compensation. In order to protect personal information, the Personal Information Protection Act recognizes collecting and transacting personal information, but only permits the distribution of unidentifiable information (pseudonym information or anonymous information). Accordingly, technology is being developed to process a vast amount of data (mobile application logs, communication logs, transaction records, etc.) collected without specifying an identity, analyze the meaning, and use it for marketing. However, since such technology is difficult to identify an identity due to a limitation of information, there is a limitation in performing target marketing aimed at a small number of target buyers.

An object of the present invention is to provide a method of providing marketing information that is used in marketing activities by collecting purchase information without exposing personal information.

Another object of the present invention is to provide a marketing information platform that performs the marketing information providing method.

However, the object of the present invention is not limited to the above objects, and may be variously extended without departing from the spirit and scope of the present invention.

In order to achieve one object of the present invention, a marketing information platform according to embodiments of the present invention includes a marketing information store for receiving and storing purchase information from one or more purchase information providing systems and the purchase information stored in the marketing information storage. It may include a marketing information analysis system that generates a marketing identifier indicating a purchase tendency of a customer.

According to an embodiment, the marketing identifier is stored in the marketing information storage, and the marketing information storage receives marketing result information including marketing execution environment and marketing result information for a customer corresponding to the marketing identifier. Store, and the information analysis system may perform machine learning to generate marketing recommendation information based on the marketing result information.

According to an embodiment, the information analysis system may perform the machine learning using a deep neural network including an input layer to which the marketing identifier is input and an output layer to which the marketing recommendation information is output.

According to an embodiment, the marketing identifier may be accumulated and stored in a time series for each individual unique key corresponding to each customer in a linked list method.

According to an embodiment, the marketing information storage may be configured as a node of a block chain that stores the purchase information as a first transaction and stores the marketing identifier as a second transaction.

In order to achieve another object of the present invention, a method for providing marketing information according to embodiments of the present invention includes the steps of: a marketing information storage receiving and storing purchase information from a purchase information providing system, and a marketing information analysis system in the marketing information storage. Generating a marketing identifier indicating a purchase tendency of a customer based on the stored purchase information, receiving and storing marketing result information corresponding to the marketing identifier from a marketing information use system by the marketing information storage system, and the marketing information The analysis system may include performing machine learning for generating marketing recommendation information based on the marketing result information stored in the marketing information storage.

According to an embodiment, the information analysis system may perform the machine learning using a deep neural network including an input layer to which the marketing identifier is input and an output layer to which the marketing recommendation information is output.

According to an embodiment, the marketing information storage may be configured as a node of a block chain that stores the purchase information as a first transaction and stores the marketing identifier as a second transaction.

According to an embodiment, the marketing identifier may be accumulated and stored in a time series for each individual unique key corresponding to each customer in a linked list method.

The personal information management method according to the embodiments of the present invention can eliminate information asymmetry and increase the reliability of information through a block chain-based marketing information storage. Accordingly, it is possible to induce participation of a plurality of purchase information systems to collect purchase information, and to induce participation of a plurality of marketing information use systems to collect marketing result information.

In addition, the personal information management method includes information for performing marketing based on the collected purchase information, but generates a marketing identifier, which is pseudonym information that cannot be personally identified, and the generated marketing identifier is used in a number of marketing information use systems. Can be.

Further, the personal information management method may perform machine learning using marketing result information collected by a marketing information use system. Accordingly, target marketing can be performed by predicting marketing effects and generating marketing recommendation information through artificial intelligence.

The marketing information platform according to the embodiments of the present invention may support a plurality of purchase information systems and marketing information use systems to participate and perform efficient marketing without exposing personal information by performing the personal information management method. In addition, the marketing information platform can be configured based on a block chain to ensure data integrity and increase security.

However, the effects of the present invention are not limited to the above effects, and may be variously extended without departing from the spirit and scope of the present invention.

1 is a block diagram illustrating a marketing information platform according to embodiments of the present invention.
FIG. 2 is a diagram illustrating an example of purchase information stored in the marketing information platform of FIG. 1.
3 is a block diagram illustrating an example of a marketing information analysis system included in the marketing information platform of FIG. 1.
4 is a diagram illustrating an example of marketing analysis information stored in the marketing information platform of FIG. 1.
5 to 8 are diagrams illustrating examples of a method of generating and managing a marketing identifier included in the marketing analysis information of FIG. 4.
9 is a diagram illustrating an example of marketing result information stored in the marketing information platform of FIG. 1.
10 is a diagram illustrating an example in which machine learning for generating marketing recommendation information is performed in the marketing information platform of FIG. 1.
11 is a flowchart illustrating a method of providing marketing information according to embodiments of the present invention.

Hereinafter, embodiments of the present invention will be described in more detail with reference to the accompanying drawings. The same or similar reference numerals are used for the same components in the drawings.

1 is a block diagram illustrating a marketing information platform according to embodiments of the present invention.

Referring to FIG. 1, the marketing information platform 10 may receive and store purchase information from the purchase information providing system 100 and generate a marketing identifier based on the purchase information. Here, the purchase information is marketing basic information and includes information in a broad sense related to individual consumption activities. For example, the purchase information may include not only a purchase record of an individual's goods/services, but also a log of a mobile application, a web log, and the like. In addition, the marketing identifier is an identifier representing the purchasing tendency of each customer and may be used as data for performing marketing activities. The marketing identifier will be described in detail with reference to FIGS. 5 to 8. In one embodiment, the marketing information platform 10 may include a marketing information storage 200 and a marketing information analysis system 300.

The marketing information storage 200 may receive and store purchase information from one or more purchase information providing systems 100-1 to 100-n. In one embodiment, the marketing information storage 200 may be a blockchain-based storage composed of a plurality of nodes. For example, the marketing information storage 200 may be configured as a closed blockchain (ie, private blockchain) network composed of predetermined organizations or individuals. When the marketing information storage 200 is implemented based on blockchain technology, the asymmetry of information is resolved to participants such as the purchase information providing system 100 and the marketing information use system 400, and the reliability of information (data integrity is secured ), you can induce platform participation and collect various purchase information.

The marketing information analysis system 300 may generate a marketing identifier indicating a purchase tendency of a customer based on purchase information stored in the marketing information storage 200. That is, the marketing information analysis system 300 may generate a marketing identifier for each individual by analyzing various purchase information stored in the marketing information storage. In order to prevent personal information from being exposed, the marketing information analysis system 300 may receive purchase information in the form of pseudonym information in which an identity cannot be identified and may generate a marketing identifier corresponding to the pseudonym information.

In one embodiment, the marketing identifier generated by the marketing information analysis system 300 is stored in the marketing information storage 200, and the marketing information storage 200 stores marketing result information corresponding to the marketing identifier. ), and the information analysis system 300 may perform machine learning to generate marketing recommendation information based on the marketing result information. That is, the marketing information use system 400 performs marketing activities using the marketing identifier generated by the marketing information analysis system 300, generates marketing result information that is feedback on the marketing activity, and stores the generated marketing result information. It can be stored in the marketing information storage 200. The information analysis system 300 may perform machine learning by using the marketing result information as learning data, and accordingly, may predict a marketing result corresponding to the marketing identifier.

Although FIG. 1 describes that the marketing information storage 200 is configured in a private blockchain method, it is not limited thereto. For example, the marketing information storage 200 may be implemented as a public blockchain method, or may be implemented as a general database management system rather than a blockchain method.

FIG. 2 is a diagram illustrating an example of purchase information stored in the marketing information platform of FIG. 1.

Referring to FIG. 2, the marketing information storage 200 may be configured as a node of a block chain that stores purchase information as a first transaction. For example, a block chain is a structure in which blocks and blocks are connected in a chain form (linked list), and each block may be composed of first transactions including a block header and purchase information.

The block header may include a previous hash value (previous hash), a nonce (nounce), and a merkle root (merkle root). The hash value of the previous block may be a value converted using the SHA256 hash function after summing all block header information of the previous block. The Merkle root may be a value of the root node of the Merkle tree to verify the integrity of the transaction. For example, the Merkle root may be a value of a root node of a Merkle tree with a hash value of the first transaction as a leaf node. Nonce is data for obtaining a hash value in the process of creating a block, and can be increased by 1 until a hash value that satisfies the condition is found. In addition to this, the block header may further include version information and block generation time.

The first transaction may be purchase information provided by the purchase information providing system 100. In an embodiment, the purchase information is pseudonym information that cannot identify the identity of the purchaser alone, and may represent the purchase details of an individual's goods/services. For example, the first transaction may include a personal unique key, a business type, a service area, a service provider, a use date, a use amount, a sex, an age group, a place of residence, and the like. Here, the personal unique key may be a hash value for a personal identifier (name, date of birth, etc.). The business type, service area, and service provider may be information on the business type, region, and provider of the provided goods/services. The date and time of use and the amount of use may indicate the date and amount of goods/services provided. The sex and age range of residence may indicate information on the purchaser of goods/services.

Although, in FIG. 2, it has been described that the personal unique key corresponds to a hash value for a combination of a name and a date of birth, but the present disclosure is not limited thereto. For example, the personal unique key may be data indicating personal identification information such as a resident registration number, email, and member ID. In addition, although FIG. 2 shows an example of configuring a block chain, the block chain can be configured in various ways.

3 is a block diagram illustrating an example of a marketing information analysis system included in the marketing information platform of FIG. 1.

3, the marketing information analysis system 300 may generate marketing analysis information based on purchase information stored in the marketing information storage 200, or may generate marketing recommendation information through machine learning based on the marketing result information. have. In one embodiment, the marketing information analysis system 300 may include a data pre-processing module 310, a data analysis information generation module 320, a marketing result learning module 330, and a marketing recommendation information generation module 340. .

The data preprocessing module 310 may perform data preprocessing for generating a marketing identifier for purchase information stored in the marketing information storage 200. As the purchase information is generated by the plurality of purchase information providing systems 100-1 to 100-n, the data format may be different, and thus the data may be changed to the same format before analysis. In this case, the data preprocessing module 310 may filter purchase information to exclude some purchase information. For example, purchase information generated in a purchase information system for which reliability is not guaranteed may be excluded, or purchase information meeting a preset condition (eg, occurring in a specific industry or a specific region) may be excluded.

In addition, the data pre-processing module 310 may perform data pre-processing for performing machine learning on the marketing result information stored in the marketing information storage 200. As the marketing result information is generated by the plurality of marketing information using systems 400-1 to 400-n, the data format may be different, and thus the data format may be changed to the same format before being used as machine learning training data. In this case, the data preprocessing module 310 may filter the marketing result information to exclude some marketing result information. For example, excluding marketing result information generated by a marketing information use system that is not guaranteed to be reliable, or marketing result information that meets preset conditions (e.g., specific weather, specific date, specific day of the week, or occurring in a specific region) Can be excluded.

The data analysis information generation module 320 may generate marketing analysis information based on purchase information stored in the marketing information storage 200. The data analysis information generation module 320 may analyze a customer's recent use performance and consumption tendency based on purchase information for each customer to derive a marketing identifier and generate marketing analysis information. A method of generating marketing analysis information will be described in detail with reference to FIGS. 4 to 8.

The marketing result learning module 330 may perform machine learning by using the marketing result information as learning data. In an embodiment, the marketing result learning module 330 may perform machine learning using a deep neural network including an input layer into which marketing identifiers and marketing basic information are input, and an output layer through which marketing recommendation information is output. A method of performing machine learning will be described in detail in FIG. 10.

The marketing recommendation information generation module 340 may generate marketing recommendation information based on the model learned in the marketing result learning module 330. Here, the marketing recommendation information represents information on a business type, item, product, service, etc. that are expected to have high marketing effect. In one embodiment, the marketing recommendation information generation module 340 inputs the marketing identifier and marketing basic information of the customer to derive marketing recommendation information, thereby predicting the results of which marketing will be effective, and providing appropriate products and services for the customer. Can recommend.

4 is a diagram illustrating an example of marketing analysis information stored in the marketing information platform of FIG. 1.

Referring to FIG. 4, the marketing information storage 200 may be configured as a node of a block chain that stores marketing analysis information as a second transaction. The second transaction may be marketing analysis information generated by analyzing a first transaction (ie, purchase information) by the marketing information analysis system 300. In one embodiment, the marketing analysis information may include a marketing identifier corresponding to a personal key as pseudonym information in which the identity of a purchaser cannot be determined by itself. Here, the marketing identifier may be generated from a customer's recent usage performance viewpoint, a demographic viewpoint, or a consumption tendency viewpoint through purchase information corresponding to the personal unique key. In addition, the marketing analysis information may include a creator ID and a creation date. Since a plurality of operating entities may each operate the marketing information analysis system 300, it is possible to store the creator ID and the creation date that generated the marketing analysis information.

In one embodiment, the marketing identifier may be generated with 13 digits of the same number as the social security number. In the case of interlocking with a legacy system that identifies customers using a resident registration number, it is possible to facilitate interworking with the legacy system by generating a marketing identifier with 13 digits identical to the resident registration number. For example, in a system where it is difficult to store a social security number according to the Personal Information Protection Act, it is possible to perform customer identification and marketing activities by storing a marketing identifier and a personal unique key instead of a social security number.

5 to 8 are diagrams illustrating examples of a method of generating and managing a marketing identifier included in the marketing analysis information of FIG. 4.

5 to 8, the marketing information analysis system 300 may generate a marketing identifier indicating a purchase tendency of a customer based on purchase information stored in the marketing information storage 200. In one embodiment, the marketing identifier may consist of 13 digits. The first 6 digits of the marketing identifier may represent the identifier creation year, identifier production month, and marketing data type, respectively. The identifier generation year may indicate the year in which the marketing identifier was generated, and the identifier generation month may indicate the month in which the marketing identifier was generated. For example, if the identifier creation year is '18' and the identifier creation month is '03', it can be confirmed that the marketing identifier was generated in March 2018. In addition, the marketing data type may indicate the format type of the marketing identifier. Since the marketing identifier can be analyzed and generated in various ways, it is possible to check what meaning of the last 7 digits of the marketing identifier according to the marketing data type.

As shown in FIG. 5, the marketing information analysis system 300 may generate a marketing identifier based on the customer's recent usage performance. When the marketing data type is '01', the marketing identifier is generated based on the customer's recent usage performance, and the last 7 digits are the service use area, type of business, service use days, lost months, consecutive months of ineffectiveness, and months. You can display the average number of usages and the average monthly usage amount. For example, the 7th number of the marketing identifier can indicate the service use area (the product purchase area), and it can be confirmed that the service was used or the product was purchased in Seoul in case of 1, Gyeonggi in case of 2, and Incheon. have. The eighth number of the marketing identifier can indicate the type of business, and if it is 1, it can be confirmed that furniture has been purchased, and if it is 2, it can be confirmed that household appliances have been purchased. The ninth number of the marketing identifier may represent the number of days of service use in the last 6 months, and it can be confirmed that the number of days of service use for the last 6 months is 1, 20 days or more, 2 is 15 days or more, and 3 is 12 days or more. The tenth number of the marketing identifier may represent the number of months lost in the last 3 years, and it can be confirmed that the number of months of loss in the case of 1 was 36 months or more, 34 to 35 months for 2, and 30 to 33 months for 3. The 11th digit of the marketing identifier can indicate the number of consecutive months of non-performing, and it can be confirmed that the service has not been used for 12 months or longer in the case of 1, 8 to 11 months in the case of 2, and 6 to 7 months in the case of 3. The 12th number of the marketing identifier can represent the average number of monthly usage for the last 6 months. In the case of 1, an average of 100 or more per month, in the case of 2, an average of 60 to 99 per month, and in case of 3, an average of 40 to 59 Can be confirmed. The 13th digit of the marketing identifier can represent the average monthly usage amount for the last 6 months. In the case of 1, it corresponds to an average of 5 million won or more per month, in the case of 2, an average of 4-4.99 million won, and in the case of 3, it corresponds to an average of 3 million to 3.99 million won. It can be confirmed that the service was used.

As shown in FIG. 6, the marketing information analysis system 300 may generate a marketing identifier based on a customer's recent usage performance and a demographic perspective. When the marketing data type is '02', the marketing identifier is generated based on the customer's recent usage performance and demographic perspective, and the last 7 digits are gender, residential area, household life cycle, sales potential, average monthly usage amount, and week. It can indicate the type of merchant used. For example, the 7th digit of the marketing identifier may indicate the gender, and 1 may be identified as male and 2 may be identified as female. The 8th to 9th digits of the marketing identifier can indicate the residential area, and can be identified as Seoul for 01, Gyeonggi for 02, and Incheon for 03. The tenth number of the marketing identifier can represent the life cycle of the household, and in the case of 1, it can be identified as a university student, and in the case of 2, it can be identified as an employee. The 11th number of the marketing identifier can indicate the sales potential, and it can be identified as exceeding 60 million won for 1, 5,000 to 60 million won for 2, and 40 to 50 million won for 3. The twelfth number of the marketing identifier can represent the average monthly usage amount for the last 6 months. In the case of 1, it is equivalent to an average of 5 million won or more per month, in the case of 2, an average of 4-4.99 million won per month, and in the case of 3, it corresponds to an average of 3 million to 3.99 million won per month. It can be confirmed that the service was used. The 13th number of the marketing identifier may indicate the type of business, and if 1, it can be confirmed that furniture mainly purchased, and if 2, household appliances, etc.

As shown in FIG. 7, the marketing information analysis system 300 may generate a marketing identifier based on a consumer's consumption tendency viewpoint. That is, by analyzing consumption propensity information, a customer's life style and preferred business type may be reflected. When the marketing data type is '03', the marketing identifier is generated based on the viewpoint of consumption propensity, and the last 7 digits may represent gender, needs cluster, customer prism matrix (CPM) index, and purchase pattern. Here, the needs cluster refers to an index representing a customer's life style as information that is profiled by linking and analyzing research information and consumption information. CPM index refers to an index for service use characteristics based on service use details. For example, the 7th digit of the marketing identifier may indicate the gender, and 1 may be identified as male and 2 may be identified as female. The 8th to 9th digits of the marketing identifier may indicate the needs cluster, and in the case of 01, it may indicate the first-year social type, in the case of 02, the child-rearing type, and in the case of 03, the middle-aged male abundance type. The 10th to 11th digits of the marketing identifier can represent the CPM index, and in the case of 01, it can be identified as leisure activity, in the case of 02, education for children, and in the case of 03, it can be identified as culture and art. The 12th to 13th digits of the marketing identifier can indicate purchasing pattern characteristics, and can be identified as vegan (vegan) in case of 01, vegetarian (vegetarian) in case of 02, and diet (diet-related product) in case of 03.

As shown in FIG. 8, the marketing identifier may be accumulated and stored in a time series corresponding to a unique number (ie, a personal unique key) of one customer. Customers need to update their marketing identifiers as their marketing characteristics may change over time. In addition, since the marketing identifier can be generated by various subjects or in various formats, the marketing information use system 400 that performs marketing activities using the marketing identifier further obtains various marketing identifiers corresponding to one private key. Can perform various marketing activities. Furthermore, when the marketing identifier is used as the individual's unique identification information, data mismatch can be prevented by managing the history of the marketing identifier. For example, if the marketing identifier corresponding to '188adc6cae4a5aae4a6b41b93e891b9235a26165' is sequentially changed to '180301-1234542', '190102-0112351', and '200203-1010102' in the personal unique key, it can be managed in the form of a linked list.

Although, in FIGS. 5 to 8, it has been described that the marketing identifier consists of 13 digits, it is not limited thereto. In one example, the marketing identifier may be generated in a variable size according to the marketing data type. In another example, a serial number or the like may be added so that the marketing identifiers do not overlap with each other. In another example, the marketing identifier may include characters in addition to numbers.

9 is a diagram illustrating an example of marketing result information stored in the marketing information platform of FIG. 1.

Referring to FIG. 9, the marketing information storage 200 may be configured as a node of a block chain that stores marketing result information as a third transaction. The third transaction may be marketing result information provided by the marketing information use system 400. In an embodiment, the marketing result information may include basic marketing information and marketing recommendation information as pseudonym information for which the identity of a purchaser cannot be determined by itself. For example, the third transaction may include a personal unique key, a marketing performer, a creation date, basic marketing information, marketing recommendation information, and the like. Here, the personal unique key may be a hash value for a personal identifier (name, date of birth, etc.). The marketing performer may use the marketing identifier to indicate the identification ID of the subject performing the marketing. The creation date may indicate the date when marketing result information is generated after performing marketing. The marketing basic information may include information on a marketing execution environment, a marketing execution condition, and a marketing target product/service (ie, marketing execution date, day, weather, online marketing, etc.). The marketing recommendation information is information on a field that is judged to have a high marketing effect based on the result of marketing, and may include information on what kind of marketing is effective for the customer in a situation corresponding to the basic marketing information. . For example, if the marketing recommendation information is'KIDULT', it may indicate that it is effective in marketing for adults who purchase figures, toys, etc. with the atmosphere and emotion of childhood.

Although, in FIG. 9, it has been described that the marketing result information includes marketing recommendation information, which is processed data on the result of marketing, the structure of the marketing result information is not limited thereto. For example, the marketing result information may be composed of raw data that has performed marketing. In this case, the marketing result information may be stored only for the marketing result value (whether or not there was a customer's response to marketing execution). In addition, the marketing result information may further include a marketing identifier corresponding to the personal unique key.

10 is a diagram illustrating an example in which machine learning for generating marketing recommendation information is performed in the marketing information platform of FIG. 1.

Referring to FIG. 10, the information analysis system 300 may perform machine learning by using marketing basic information included in the marketing result information, marketing recommendation information, and a marketing identifier corresponding to a personal key as learning data.

In one embodiment, the information analysis system 300 performs machine learning using a deep neural network including an input layer IL in which marketing identifiers and marketing basic information are input, and an output layer OL in which marketing recommendation information is output. can do. Specifically, the deep neural network may be divided into an input layer (IL), a hidden layer (HL), and an output layer (OL), and the hidden layer may be composed of two or more layers. The input layer IL may play a role of receiving marketing identifiers and marketing basic information as input data from outside the system and transmitting them to the system. That is, the input layer IL serves to input a value of a predictor (input variable) to derive a predicted value (output variable), and may have n input nodes when there are n input values. The hidden layer HL is located between the input layer IL and the output layer OL, and may receive and process an input value, and then calculate a result. The hidden layer HL may receive an input value from an input node, calculate a weighted sum, apply this value to a transition function, and transmit the value to the output layer OL. 'Input node-hidden node' and'hidden node-output node' can be connected to a network having a weight. In one example, the number of hidden layers and hidden nodes may be set differently according to the format of the marketing identifier. In another example, the number of hidden layers and hidden nodes may be determined in proportion to the number of basic marketing information. In this case, a plurality of deep neural networks may be generated according to the format of the marketing identifier or the number of basic marketing information. The output layer OL may output marketing recommendation information as a result value according to an input value and a current system state (such as a hidden layer IL configuration state).

In one embodiment, the deep neural network learning method is

1) initializing the weights of the neural network,

2) inputting the marketing identifier and basic marketing information as input values into a deep neural network and calculating an output value,

3) Computing an error by comparing the result value of the training data (ie, marketing recommendation information) and the output value of the deep neural network,

4) adjusting the weight of the neural network to reduce the error,

5) Repeat steps 2) to 4) for all training data and

6) It may include repeating steps 2) to 5) until the output error of the neural network is sufficiently reduced.

In the case of a deep neural network, in the learning process, the weight can be adjusted to a value that best fits the predicted value using a backpropagation algorithm. The backpropagation algorithm moves the output error of the neural network from the output layer to the hidden layer in front of the input layer in reverse order, and determines the error of the hidden layer nodes based on the error of the output node (that is, the difference between the result of the training data and the output of the neural network). And the weight can be adjusted based on the error. Here, when a certain input node contributes to the error of the output node, the connection weight of the two nodes may be adjusted in proportion to the error between the output of the corresponding input node and the output node.

The information analysis system 300 may generate marketing recommendation information for a marketing target customer through artificial intelligence based on the learned model. The information analysis system 300 predicts the outcome of what kind of marketing will be effective by inputting the marketing identifier and marketing basic information of the customer into the learned deep neural network to derive marketing recommendation information, and recommends appropriate products and services to the customer. can do.

Although, in FIG. 10, when machine learning is performed, it has been described that a deep neural network including an input layer into which marketing identifier and marketing basic information are input and an output layer through which marketing recommendation information is output is used, but the present invention is not limited thereto. For example, only information included in the marketing identifier may be input to the input layer, and the marketing execution result value may be output to the output layer.

11 is a flowchart illustrating a method of providing marketing information according to embodiments of the present invention.

Referring to FIG. 11, in the marketing information providing method, the marketing information storage 200 receives and stores purchase information from the purchase information providing system 100, and the marketing information analysis system 300 is in the marketing information storage 200. Generating a marketing identifier representing a customer's purchasing tendency based on the stored purchase information, receiving and storing the marketing result corresponding to the marketing identifier from the marketing information use system 400 by the marketing information storage 200, and marketing The information analysis system 300 may include performing machine learning for generating marketing recommendation information based on the marketing result stored in the marketing information storage 200.

Specifically, the marketing information storage 200 may receive and store purchase information from the purchase information providing system 100 (S10). In one embodiment, the marketing information storage 200 may be configured as a node of a blockchain that stores purchase information as a first transaction. Purchasing information is pseudonym information in which the identity of the purchaser cannot be identified by itself, and may represent an individual's purchase of goods/services. The first transaction may include a personal unique key, a business type, a service area, a service provider, a use date, a use amount, a gender, an age group, a place of residence, and the like. However, since the manner in which the marketing information storage 200 stores purchase information has been described in FIGS. 1 and 2, duplicate descriptions will be omitted.

The marketing information analysis system 300 may analyze the purchase information S20 for each customer stored in the marketing information storage 200 to generate a marketing identifier (S30). The marketing information analysis system 300 may derive a marketing identifier and generate marketing analysis information by analyzing a customer's recent usage performance, consumption tendency, etc. based on purchase information for each customer. In one embodiment, the marketing identifier may be generated with 13 digits of the same number as the social security number. However, since the method of generating the marketing identifier by the marketing information analysis system 300 has been described in FIGS. 5 to 7, duplicate descriptions will be omitted.

The marketing information storage 200 may receive and store marketing analysis information including a marketing identifier from the marketing information analysis system 300 (S40). The marketing information storage 200 may be configured as a node of a block chain that stores marketing analysis information as a second transaction. The second transaction may be marketing analysis information generated by analyzing a first transaction (ie, purchase information) by the marketing information analysis system 300. In one embodiment, the marketing analysis information may include a marketing identifier corresponding to a personal key as pseudonym information in which the identity of a purchaser cannot be determined by itself. In one embodiment, the marketing identifier may be accumulated and stored in a time series corresponding to a unique number of one customer. However, since the manner in which the marketing information storage 200 stores the marketing analysis information has been described in FIGS. 4 and 8, duplicate descriptions will be omitted.

The marketing information use system 400 receives a marketing identifier for a marketing target customer from the marketing information storage 200 (S50), and performs marketing by selecting a marketing target customer and a marketing target product using the marketing identifier (S60). can do.

The marketing information storage 200 may receive and store marketing result information (ie, marketing performance feedback), which is a result of marketing execution, from the marketing information use system 400 (S70). The marketing information storage 200 may be composed of a node of a block chain that stores marketing result information as a third transaction. The third transaction may be marketing result information provided by the marketing information use system 400. In an embodiment, the marketing result information may include basic marketing information and marketing recommendation information as pseudonym information for which the identity of a purchaser cannot be determined by itself. However, since the manner in which the marketing information storage 200 stores the marketing result information has been described in FIG. 9, a duplicate description will be omitted.

The marketing information analysis system 300 may perform machine learning for generating marketing recommendation information based on the marketing result stored in the marketing information storage 200 (S90). In one embodiment, the information analysis system 300 may perform machine learning using a deep neural network including an input layer into which marketing identifiers and basic marketing information are input, and an output layer through which marketing recommendation information is output. However, since the method for the marketing information analysis system 300 to perform machine learning has been described in FIG. 10, duplicate descriptions will be omitted.

In addition, the marketing information analysis system 300 may generate marketing recommendation information for a marketing target customer based on the learned model (S100). The information analysis system 300 may generate marketing recommendation information by inputting a customer's marketing identifier and marketing basic information into the learned deep neural network. The generated marketing recommendation information is transmitted to the marketing information use system in various ways, so that the results of which marketing will be effective can be predicted and appropriate products and services can be recommended to customers.

In the above, the method for providing marketing information and the marketing information platform for performing the method according to the embodiments of the present invention have been described with reference to the drawings, but the description is illustrative and in the relevant technical field without departing from the technical idea of the present invention. It may be modified and changed by those with ordinary knowledge. For example, although it has been described that the purchase information providing system and the marketing information using system are separated from each other, the purchase information providing system and the marketing information using system may be the same system, and the types thereof are not limited thereto. In addition, although it has been described above that the marketing identifier consists of numbers, the marketing identifier may include characters and special symbols in addition to numbers.

The present invention can be applied to any system that stores personal activity information and performs marketing activities. For example, the present invention can be applied to an e-commerce platform, a social network platform, a telecommunication company platform, a distribution company platform, a logistics company platform, and the like.

Although the above has been described with reference to embodiments of the present invention, those of ordinary skill in the art will variously modify and change the present invention within the scope not departing from the spirit and scope of the present invention described in the following claims. You will understand that you can.

10: Marketing Information Platform
100, 100-1 to 100-n: purchase information providing system
200: Marketing Information Store
300: marketing information analysis system
310: data preprocessing module
320: Marketing analysis information generation module
330: Marketing Results Learning Module
340: Marketing recommendation information generation module
400, 400-1 to 400-n: marketing information utilization system

Claims (9)

A marketing information storage storing purchase information received from a plurality of purchase information providing systems and marketing result information received from a plurality of marketing information using systems; And
And a marketing information analysis system for generating a marketing identifier indicating a purchase tendency of a customer based on the purchase information stored in the marketing information storage,
The marketing information analysis system includes a data preprocessing module, a data analysis information generation module, a marketing result learning module, and a marketing recommendation information generation module,
The data pre-processing module performs data pre-processing for generating the marketing identifier for the purchase information,
The data pre-processing excludes purchase information generated by a purchase information providing system whose reliability is not guaranteed among the purchase information providing systems,
The data analysis information generation module generates marketing analysis information based on purchase information on which the data pre-processing has been performed,
The marketing analysis information is pseudonym information in which the identity of the purchaser cannot be independently identified, and includes the marketing identifier corresponding to the personal unique key,
The marketing result learning module performs machine learning by using the marketing result information as learning data,
The marketing result information includes marketing execution result information for a customer corresponding to a marketing execution environment and the marketing identifier,
The marketing recommendation information generation module generates marketing recommendation information based on the model learned in the marketing result learning module,
The marketing result learning module performs the machine learning using a deep neural network including an input layer into which the marketing identifier and the marketing execution environment are input, and an output layer through which the marketing recommendation information is output,
The marketing information store is a marketing information platform comprising a node of a blockchain that stores the purchase information as a first transaction, stores the marketing identifier as a second transaction, and stores the marketing result information as a third transaction. . delete delete The marketing information platform according to claim 1, wherein the marketing identifier is accumulated and stored in a time series for each individual unique key corresponding to each customer in a linked list method. delete Receiving and storing purchase information from a plurality of purchase information providing systems by the marketing information storage;
Generating, by a marketing information analysis system, a marketing identifier indicating a purchase tendency of a customer based on the purchase information stored in the marketing information storage;
Receiving and storing, by the marketing information storage system, marketing result information corresponding to the marketing identifier from a plurality of marketing information using systems; And
And performing, by the marketing information analysis system, machine learning for generating marketing recommendation information based on the marketing result information stored in the marketing information storage,
The marketing information analysis system includes a data preprocessing module, a data analysis information generation module, a marketing result learning module, and a marketing recommendation information generation module,
The data pre-processing module performs data pre-processing for generating the marketing identifier for the purchase information,
The data pre-processing excludes purchase information generated by a purchase information providing system whose reliability is not guaranteed among the purchase information providing systems,
The data analysis information generation module generates marketing analysis information based on purchase information on which the data preprocessing has been performed,
The marketing analysis information is pseudonym information in which the identity of the purchaser alone cannot be identified, and includes the marketing identifier corresponding to the personal unique key,
The marketing result learning module performs machine learning by using the marketing result information as learning data,
The marketing result information includes marketing execution result information for a customer corresponding to a marketing execution environment and the marketing identifier,
The marketing recommendation information generation module generates marketing recommendation information based on the model learned in the marketing result learning module,
The marketing result learning module performs the machine learning using a deep neural network including an input layer into which the marketing identifier and the marketing execution environment are input, and an output layer through which the marketing recommendation information is output,
The marketing information storage is provided with a block chain node that stores the purchase information as a first transaction, stores the marketing identifier as a second transaction, and stores the marketing result information as a third transaction. Way. delete delete The method of claim 6, wherein the marketing identifier is accumulated and stored in a time series for each individual unique key corresponding to each customer in a linked list method.

Images