KR102380750B1 - Method for predicting potential customer and system thereof - Google Patents

Abstract

In accordance with one embodiment of the present invention, disclosed is a method of predicting characteristic information of potential customers. The method includes the following steps of: collecting customer data about personal information of a plurality of customers and customer-only application use information of the customers, campaign data about past campaigns conducted for at least some of the customers, and app statistic data of a customer group collected by statistically processing logs of the customer-only application; calculating a customer profile, a campaign profile and an app statistic profile based on preset heuristic rules in regard to the customer data, the campaign data and the app statistic data; applying the calculated customer, campaign and app statistic profiles to characteristic information about a new campaign to select a first next customer group including effective customers for the new campaign, from among the plurality of customers; and excepting some of the customers included in the selected first next customer group in accordance with exception conditions set based on marketing rules, to select a second next customer group including at least one of the customers included in the first next customer group. Therefore, the present invention is capable of effectively predicting characteristic information of potential customers.

Description

Method for predicting potential customer and system thereof

The present invention relates to a method and system for predicting potential customers, and more particularly, to a method and system for implementing the method for efficiently predicting potential customers optimized for a newly planned campaign for marketing. am.

Due to the COVID-19 outbreak around the world in 2019, non-face-to-face digitalization is accelerating while emphasizing untact in all areas of society. Accordingly, companies are researching ways to effectively select the next customer based on the digital experiences left by the existing customers from various angles.

Here, the next customer can be understood as a potential customer who has the potential to use the company's products or services in the future, although they have not yet provided sales while using the company's products or services. As trends are changing, future customers should be analyzed and predicted through different criteria from existing customers who contributed to the company's sales.

1 is a diagram schematically showing a known next customer selection process.

First, a company stores raw data about customers collected in the step of providing products or services to customers in a database.

The data team that manages the company's data can refine the basic data stored in the database into data that is easy for humans to analyze (processed data, pre-processing data) and provide it to the marketing team. If the marketing team determines that the processed data received from the data team is not suitable for determining future marketing directions, the marketing team may request the data team to revise the processed data.

In addition, if there is a recent major revision of the marketing strategy (or marketing rule), the marketing team may request the data team to revise the processing method of basic data in order to reflect the revision.

For convenience of explanation in FIG. 1 , although it is illustrated that the request for correction of processed data is performed once, in reality, opinions are exchanged several times between the data team and the marketing team, resulting in a waste of time and money. This waste phenomenon is because there is a big gap in expertise between the data team and the marketing team.

All companies have guidelines for effectively controlling various decisions necessary to conduct business. As an example, the business policy means that the above guidelines are expressed in sentences and documented, and the marketing rule specifies specific conditions, environments, actions, etc. in detail so that decisions can be made clearly in accordance with these business policies. it means done

In the conventional decision-making system shown in FIG. 1, it was difficult to quickly accommodate changes in the times (changes in marketing rules). Even if a customer analysis system to effectively analyze customer needs has been built in advance, the marketing rules that are the basis for the operation of the system exist only in hard-coded form in the application by the conventionally well-known programming method, The marketing team without technical knowledge could apply the changed marketing rules to the system only through collaboration with the data team.

More specifically, if the conventional system is sequentially described, the marketing rule developer develops a new marketing rule or establishes a plan for partially modifying the existing marketing rule. Subsequently, the marketing rule developer provides the application developer with correction data containing information on which rule should be modified in the central server constituting the business system. The application developer performs coding based on the correction data, matches the coded result through debugging and simulation, and after modifying the application of the system built on the central server, the marketing rule developer is notified of the modification.

As above, when the marketing rule developer of the marketing team and the IT manager of the data team go through a complex communication process, it is inefficient because it consumes a lot of time and money. There is a need for a system that can not only input, but can also effectively maintain marketing rules that have already been implemented.

1. Republic of Korea Patent Registration No. 10-1034353 (published on October 7, 2010) 2. Republic of Korea Patent Registration No. 10-2013768 (published on September 16, 2019)

The technical problem to be solved by the present invention is to provide a method for predicting a next customer through a digital experience such as data of an existing customer and a system for implementing the method.

The method according to an embodiment of the present invention for solving the above technical problem, the personal information of a plurality of customers and customer data for the use information of the customer-only application of the customers from the database, conducted for at least some of the customers Collecting app statistical data of a customer group collected by statistical processing of campaign data for past campaigns and logs of the customer-only application; calculating a customer profile, a campaign profile, and an app statistics profile based on a preset heuristic rule using the customer data, the campaign data, and the app statistical data; By applying the calculated customer profile, campaign profile, and app statistics profile to characteristic information of the new campaign, the first next customer group including customers effective for the new campaign among the plurality of customers is selected to do; And by excluding some of the customers included in the selected first-term customer group according to the exclusion conditions set based on the marketing rule, selecting a second-term customer group including at least one of the customers included in the first-term customer group includes steps.

In the method, the step of selecting the second next customer group comprises the step of outputting at least one or more characteristic information of customers included in the first customer group through supervised training for a machine learning model. may include more.

In the above method, the characteristic information for the new campaign may be information input as the most recent among campaign information input by an authorized user.

In the above method, the marketing rule may be information updated by an authorized user.

In the above method, the characteristic information for the new campaign and the marketing rule may be information updated by an authorized user.

The method includes receiving a campaign result of conducting a campaign for customers included in the second next customer group, reflecting the received campaign result, and including at least one customer included in the second next customer group It may further include; selecting the third customer group.

The method may further include outputting a combination of characteristic information of customers included in the third next customer group through supervised learning of the machine learning model.

A system according to another embodiment of the present invention for solving the above technical problem, customer data for a plurality of customers' personal information and customer-only application usage information of the customers from a database, targeting at least some of the customers a raw data collection unit for collecting statistical data of a customer group collected by statistical processing of campaign data for the past campaign and logs of the customer-only application; a profile calculation unit for calculating a customer profile, a campaign profile, and an app statistics profile based on a preset heuristic rule for the customer data, the campaign data, and the app statistical data; By applying the calculated customer profile, campaign profile, and app statistics profile to characteristic information of the new campaign, the first next customer group including customers effective for the new campaign among the plurality of customers is selected the first group selection unit; And by excluding some of the customers included in the selected first-term customer group according to the exclusion conditions set based on the marketing rule, selecting a second-term customer group including at least one of the customers included in the first-term customer group Includes the second group selection unit.

In the system, the second group selection unit may output at least one or more characteristic information of customers included in the first customer group through supervised training for a machine learning model.

In the system, the characteristic information on the new campaign may be information input as the most recent among campaign information input by an authorized user.

In the system, the marketing rule may be information updated by an authorized user.

In the system, the characteristic information for the new campaign and the marketing rule may be information updated by an authorized user.

The system receives a campaign result of conducting a campaign for customers included in the second next customer group, reflects the received campaign result, and includes at least one customer included in the second next customer group It may further include; a third group selection unit for selecting the third customer group.

The system may further include a characteristic information output unit for outputting a combination of characteristic information of customers included in the third next customer group through supervised learning of the machine learning model.

An embodiment of the present invention may provide a computer-readable recording medium storing a program for executing the method.

According to the present invention, based on the digital experience of the existing customer, it is possible to effectively predict the characteristic information of the potential customer.

In addition, according to the present invention, effective customer prediction is possible through the system simply by inputting updated marketing information into the system by the marketing rule developer.

1 is a diagram schematically showing a known next customer selection process.
Fig. 2 is a schematic representation of the process of the present invention compared to the existing process described in Fig. 1;
3 is a block diagram illustrating an example of a system for predicting characteristic information of potential customers according to the present invention.
4 and 5 are diagrams for explaining characteristic information of customers included in the first customer group output by the second group selection unit.
6 is a diagram conceptually illustrating a cyclical characteristic of the system described in FIG. 3 .
7 is a view for explaining the overall operation of the system of the present invention.
8 is a flowchart illustrating an example of a method in which the system of FIG. 3 operates.

Since the present invention can apply various transformations and can have various embodiments, specific embodiments are illustrated in the drawings and described in detail in the detailed description. Effects and features of the present invention, and a method of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various forms.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, and when described with reference to the drawings, the same or corresponding components are given the same reference numerals, and the overlapping description thereof will be omitted. .

In the following embodiments, terms such as first, second, etc. are used for the purpose of distinguishing one component from other components without limiting the meaning.

In the following examples, the singular expression includes the plural expression unless the context clearly dictates otherwise.

In the following embodiments, terms such as include or have means that the features or components described in the specification are present, and the possibility that one or more other features or components are added is not excluded in advance.

In cases where certain embodiments may be implemented differently, a specific process sequence may be performed differently from the described sequence. For example, two processes described in succession may be performed substantially simultaneously, or may be performed in an order opposite to the order described.

Fig. 2 is a schematic representation of the process of the present invention compared to the existing process described in Fig. 1;

The existing next customer predicting process shown in FIG. 1 is premised on close collaboration between the marketing team and the data team, so a lot of time and money is wasted due to unnecessary procedures or communication errors It has already been explained that this occurs.

In addition, even if communication between the data team and the marketing team has been smooth, it takes a lot of time for the data team to manually reflect the changes requested by the marketing team to the system, and communication between the two teams Regardless, errors may occur in the work process of the data team.

Meanwhile, according to the process of the present invention shown in FIG. 2, after the system is built, frequent communication between the data team and the marketing team can be omitted. Specifically, in the "automated data profiling" performed in the system of the present invention, data is refined according to the rules newly defined by the marketer (marketing rule developer) for data processing and the profile is created, so the time taken by the existing process In addition to minimizing waste and cost, there is no room for unnecessary errors in the work process of the data team. As a result, according to the present invention, more improved results can be obtained by consuming less cost than when using the existing process.

3 is a block diagram illustrating an example of a system for predicting characteristic information of potential customers according to the present invention.

Referring to FIG. 3 , the system 300 according to the present invention includes a raw data collection unit 310 , a profile calculation unit 330 , a first group selection unit 350 , a second group selection unit 370 and a third A group selection unit 390 may be included. Each component included in the system 300 of FIG. 3 may be a module in which at least one processor capable of transmitting/receiving and processing a signal is mounted, or the processor itself. Also, the system 300 of FIG. 3 may be driven in a form included in other hardware devices such as a microprocessor or a general-purpose computer system.

The names of the components constituting the system 300 of FIG. 3 are named to most intuitively express the functions performed by the components, and may be called by different names when the system 300 of FIG. 3 is actually implemented. and is not limited to the name shown in FIG. 3 . In addition, since FIG. 3 is a block diagram showing only essential components in order to clearly explain the system 300 according to the present invention, a memory and communication module required to configure a general electronic device even if not shown in FIG. 3 . , a power supply module, a power control module, etc. may of course be included in the system 300 .

First, the raw data collection unit 310 may collect customer data, campaign data, and app statistical data from a database (not shown).

Customer data on personal information of a plurality of customers and customer-only application usage information of customers, campaign data for past campaigns conducted for at least some of the customers, and logs of the customer-only application are collected by statistical processing App statistical data of customer groups can be collected.

Here, the database may be included in the system 300 or may be connected to the system 300 in a wired or wireless manner to provide raw data.

The customer data refers to data of a concept that generalizes personal information of a plurality of customers and usage information of customer-only applications of the customers.

Campaign data refers to data about a past marketing campaign that was conducted for at least some of the customers stored in the database. For example, if an advertisement for a specific product or service is sent to 10 customers, and sales are generated for 3 customers among them, data for 3 people who generated sales and 7 people who did not generate sales Each may be recorded as campaign data.

App statistical data means data collected by statistically processing logs of customer-only applications, and may be data of all customers (customer groups) stored in the database.

Customer data, campaign data, and app statistical data are data that has not been processed for analysis and processing in the system 300, only basic labeling for the time or order in which the data was collected, and is collectively referred to as raw data. can be

Next, the profile calculating unit 330 calculates a customer profile, a campaign profile, and an app statistics profile based on a predetermined heuristic rule based on the customer data, the campaign data, and the app statistical data, respectively.

Each of the profiles calculated by the profile calculating unit 330 is insufficient to accurately select the next customer (potential customer), but through a process to be described later, they are combined and processed, so that the next customer can be accurately predicted and selected. .

The profile calculating unit 330 may calculate a customer profile from customer data, a campaign profile from campaign data, and an app statistics profile from app statistical data, respectively.

Table 1 shows an example of a customer profile calculated by the profile calculating unit 330 in a table. Specifically, the profile calculating unit 330 may calculate the customer profile in a manner that hierarchically classifies customer data into large items, small items, etc. according to a classification criterion set therein. Referring to Table 1, it can be seen that the customer profile includes customer basic information, customer transaction information, and customer marketing information as large items, and each large item may further include various small item data.

The sub-items of the customer profile include not only unique personal information (personal information) such as the customer's age, gender, and birthday, but also the terms and conditions information that the customer agreed to before using the application or history information continuously calculated while using the application. can The calculated customer profile as shown in Table 1 can be appropriately labeled with numbers or representative characters and managed.

Table 2 shows an example of a campaign profile calculated by the profile calculating unit 330 in a table. Specifically, the profile calculating unit 330 may calculate the campaign profile in a manner that hierarchically classifies the campaign data into large items, small items, etc. according to a classification criterion set therein. Referring to Table 2, it can be seen that the campaign profile includes campaign information, campaign aggregate information, and campaign target customer type as major items, and each major item may further include various sub-item data.

The campaign profile has the characteristics of total information of customer groups participating in the campaign (marketing). That is, the campaign profile may include representative values collected by a plurality of customers.

Table 3 shows an example of the app statistics profile calculated by the profile calculating unit 330 in a table. Specifically, the profile calculating unit 330 may calculate the app statistics profile in a manner that hierarchically classifies the app statistics data into large items, small items, etc. according to a classification criterion set therein. Referring to Table 3, it can be seen that the app statistics profile includes activity information per app and aggregate information per app as major items, and each major item can further include various sub-item data.

The large items and sub-items described in Tables 1 to 3 are examples of items constituting the profile, and when the actual system 300 is implemented, more or less information than the information in Tables 1 to 3 is large or small. can be included in

As described above, when various profile data is calculated by the profile calculating unit 330, the first group selection unit 350 can select the first next customer group based on the calculated profile data and characteristic information on the new campaign. there is.

The new campaign means a marketing campaign newly introduced, and the characteristic information on the new campaign means detailed information constituting the marketing campaign.

In addition, the first next customer group refers to a set of customers who are expected to have the most positive results when a promotion according to a new campaign is performed among a plurality of customers stored in the database. The first next customer group may include at least one customer. For example, if the number of customers stored in the database is 100, the first next customer group may include 10 filtered out of 100 customers.

The customer profile, campaign profile, and app statistics profile are data calculated by the profile calculation unit 330, and since it is a kind of pre-processing data, it can be processed uniformly by the first group selection unit 350 and , the first group selection unit 350 may select the first next customer group based on past history data and characteristic information constituting the new campaign.

Here, the first next customer group selected by the first group selection unit 350 means the result selected by focusing on the raw data collected by the raw data collection unit 310 and the characteristic information of the new campaign. The present invention includes a feature of selecting the second next customer group by filtering the first next customer group according to a marketing rule additionally defined by the marketer.

Specifically, the second group selection unit 370 may select the second next customer group from which some of the customers included in the first next customer group are excluded according to an exception condition set based on the marketing rule. .

Table 4 shows an example of a marketing rule for selecting the second next customer group as a table. The second group selection unit 370 sets exclusion conditions based on the marketing rules described in Table 4, and according to the set exclusion conditions, selects a second group of customers including at least some of the customers included in the first group of customers. can be selected According to an exemplary embodiment, if no marketing rule is set in the second group selection unit 370 or the exclusion condition is substantially absent due to the rough marketing rule, the customers included in the first next customer group and the second next customer group are the same You may.

The exclusion conditions applied by the second group selection unit 370 and the marketing rules necessary to set the exclusion conditions are information not related to the new campaign. Specifically, the marketing rules of the second group selection unit 370 accurately reflect the changing trends every moment or when selecting a customer group, the function of a safety device to prevent any violation of specific laws (laws) from occurring. can be done The marketing rule and the exclusion condition set based on the marketing rule may be periodically or aperiodically updated by an authorized user such as a marketing rule developer.

In particular, if additional filtering is needed for a specific gender, specific region, specific place, or specific country for a certain period of time due to laws or company circumstances, regardless of the new campaign, the marketing rule developer sets the exclusion conditions based on the marketing rule in the system 300 By updating to , it is possible to control so that only the most appropriate objects are selected.

Through the above process, the system 300 according to the present invention can select a second-generation customer group that includes customers effective for a new campaign and complies with the exclusion conditions according to the marketing rules from among the multiple customers stored in the database. . For example, if the number of customers stored in the database is 100, the number of customers included in the first next customer group may be 10, and the number of customers included in the second next customer group may be 7, which is less than 10.

As an optional embodiment, the second group selection unit 370 may output at least one or more characteristic information of customers included in the first customer group through supervised learning of the machine learning model. The second group selection unit 370 may only perform a function of selecting the second next customer group, and according to this optional embodiment, summarizes characteristic information of customers included in the first next customer group to which the exclusion condition is not applied can also be printed out.

4 and 5 are diagrams for explaining characteristic information of customers included in the first customer group output by the second group selection unit.

The second group selection unit 370 may include a module for building and verifying a model for performing machine learning. In this case, the algorithm used as the machine learning model may be one of the supervised learning algorithms that uses a profile generated through past campaign history as input data and a customer profile as output data (label). For example, the supervised learning algorithm used by the second group selection unit 370 may be any one selected from regression, decision trees, and random forests.

In FIG. 4 , the second group selection unit 370 perfectly learns the model through the past data, so that the characteristic information of the customer who participated in the past campaign is output from the model. Next, as shown in FIG. 5 , the second group selection unit 370 is selected by the first group selection unit 350 based on the learned machine learning model, and customers included in the first next customer group scheduled to participate in the new campaign. It is possible to output the characteristic information of

In this optional embodiment, it is only possible to know how many customers are included in the first next customer group and who they are, and by collectively outputting characteristic information of customers included in the first next customer group, Even if it is not stored as customer information in the database, it can be distinguished from the other embodiments described above in that it can clearly indicate what characteristics of customers that can be included in the first next customer group are.

For example, according to this optional embodiment, the first next customer group to which the new campaign will be applied includes three or more A, B, and C, and A, B, and C have a common origin by the learned model. By being analyzed (output), it can be helpful to determine that the user who received the analysis result is not a customer stored in the database, but that A, B, and C and D are also effective as publicity targets for the new campaign.

Again, FIG. 3 will be described next.

The third group selection unit 390 receives the campaign results of the campaign for customers included in the second next customer group, reflects the received campaign results, and includes at least one customer included in the second next customer group You can select a third-generation customer group.

The third customer group selected by the third group selection unit 390 is data that reflects not only the characteristic information about the new campaign, but also the actual promotional results as feedback data, so it may be more accurate than the information on the second customer group. can

The third group selection unit 390 performs a function of allowing a more accurate next customer prediction result to be obtained through a feedback process, and may be omitted from the system 300 according to an embodiment. A description of the third group selection unit 390 will be described later with reference to FIG. 7 .

As another optional embodiment, the characteristic information on the new campaign may be information updated by an authorized user. Specifically, the system 300 includes an input unit provided in a general electronic system, and characteristic information for a new campaign may be updated by an authorized user accessible to the input unit.

The marketing rule developer updates the changed information in the system 300 without unnecessary consultation with the data team every time a marketing rule is changed or a new campaign is released, so that the system 300 is the next customer of the new campaign ( potential customers) can be accurately predicted.

As another optional embodiment from the above-described embodiment, the system 300 may further include a characteristic information output unit (not shown). The characteristic information output unit (not shown) may perform a function of outputting a combination of characteristic information (customer profile) of customers included in the third group through supervised learning of the machine learning model. Through the characteristic information of the customers output from the characteristic information output unit (not shown), the manager (eg, a marketing rule developer) of the system 300 can identify the characteristics of the customers most optimized for the new campaign.

6 is a diagram conceptually illustrating a cyclical characteristic of the system described in FIG. 3 .

A marketer can access the system 300 as an authorized user and define information about a new campaign, and the system 300 according to the present invention recommends a next customer suitable for a new campaign through a series of processes, or Characteristic information (customer profile) of future customers can be output. In this process, when the actual campaign result is received as feedback data, the system 300 that has learned the feedback data as one parameter may select the third next customer group for the new campaign again.

7 is a view for explaining the overall operation of the system of the present invention.

The system 300 of FIG. 3 may accurately predict and select the next customer for the new campaign while going through a series of processes from the bottom to the top of FIG. 7 . Hereinafter, it will be described with reference to FIG. 3 .

The raw data collection unit 310 may perform the first process 710 of FIG. 7 . The first process 710 has been described as a process of collecting customer data, campaign data, and app statistical data.

The profile calculating unit 330 may calculate various profiles by collecting raw data, and the function of the profile calculating unit 330 is illustrated in the second process 720 of FIG. 7 .

The functions of the first group selection unit 350 to the third group selection unit 390 are illustrated as a third process 730 in FIG. 7 . According to an embodiment, the second group selection unit 370 operates according to the fourth process 740 of FIG. 7 , and the third group selection unit 390 operates including the fifth process 750 of FIG. 7 . can do. In particular, the fifth process 750 schematically shows the process of receiving the next customer prediction result (campaign result) for the new campaign as feedback data and using it for the new prediction.

A fourth process 740 and a sixth process 760 of FIG. 7 are optional embodiments, and refer to processes through which a marketer can update specific information.

8 is a flowchart illustrating an example of a method in which the system of FIG. 3 operates.

Since the method according to FIG. 8 can be implemented by the system 300 according to FIG. 3 , hereinafter, it will be described with reference to FIG. 3 , and a description duplicated with the content described in FIG. 3 will be omitted.

The raw data collection unit 310 collects customer data, campaign data, app statistical data, and more than three kinds of basic data (S810).

Next, the profile calculating unit 330 calculates each profile data for the basic data (S830).

The first group selection unit 350 may select the first next customer group by using the calculated profile data, and the first next customer group means customer data to which the characteristic information of the new campaign is applied (S850).

The second group selection unit 370 may select the second next customer group from the first next customer group according to the exclusion conditions set based on the marketing rule (S870).

Although omitted in FIG. 8, as an optional embodiment of step S870, the third group selection unit 390 may select the third next customer group using the feedback data has already been described with reference to FIGS. 3 to 7 . there is.

According to the present invention, based on the digital experience of the existing customer, it is possible to effectively predict the characteristic information of the potential customer.

In addition, according to the present invention, effective customer prediction is possible through the system simply by inputting updated marketing information into the system by the marketing rule developer.

The embodiment according to the present invention described above may be implemented in the form of a computer program that can be executed through various components on a computer, and such a computer program may be recorded in a computer-readable medium. In this case, the medium includes a hard disk, a magnetic medium such as a floppy disk and a magnetic tape, an optical recording medium such as CD-ROM and DVD, a magneto-optical medium such as a floppy disk, and a ROM. , RAM, flash memory, and the like, and hardware devices specially configured to store and execute program instructions.

Meanwhile, the computer program may be specially designed and configured for the present invention, or may be known and used by those skilled in the computer software field. Examples of the computer program may include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like.

The specific implementations described in the present invention are only examples, and do not limit the scope of the present invention in any way. For brevity of the specification, descriptions of conventional electronic components, control systems, software, and other functional aspects of the systems may be omitted. In addition, the connection or connection members of lines between the components shown in the drawings illustratively represent functional connections and/or physical or circuit connections, and in an actual device, various functional connections, physical connections that are replaceable or additional may be referred to as connections, or circuit connections. In addition, unless there is a specific reference such as “essential” or “importantly”, it may not be a necessary component for the application of the present invention.

In the specification of the present invention (especially in the claims), the use of the term “above” and similar referential terms may be used in both the singular and the plural. In addition, when a range is described in the present invention, each individual value constituting the range is described in the detailed description of the invention as including the invention to which individual values belonging to the range are applied (unless there is a description to the contrary). same as Finally, the steps constituting the method according to the present invention may be performed in an appropriate order unless the order is explicitly stated or there is no description to the contrary. The present invention is not necessarily limited to the order in which the steps are described. The use of all examples or exemplary terms (eg, etc.) in the present invention is merely for the purpose of describing the present invention in detail, and the scope of the present invention is limited by the examples or exemplary terms unless limited by the claims. it's not going to be In addition, those skilled in the art will recognize that various modifications, combinations, and changes can be made in accordance with design conditions and factors within the scope of the appended claims or their equivalents.

300: system
310: raw data collection unit
330: profile calculation unit
350: first group selection unit
370: 2nd group selection unit
390: 3rd group selection unit

Claims (15)

A raw data collection unit, personal information of a plurality of customers and customer data about the usage information of the customer-only application of the customers from the database, campaign data for past campaigns conducted for at least some of the customers, and the log of the customer-only application Collecting app statistical data of the customer group collected by statistical processing;
calculating, by a profile calculation unit, a customer profile, a campaign profile, and an app statistics profile based on a preset heuristic rule for the customer data, the campaign data, and the app statistical data;
The first group selection unit applies the calculated customer profile, campaign profile, and app statistics profile to the characteristic information of the new campaign, and includes customers effective in the new campaign among the plurality of customers. selecting a next customer group); and
The second group selection unit excludes some of the customers included in the selected first next customer group according to the exclusion conditions set based on the marketing rule, which is information not related to the new campaign, and includes it in the first next customer group A method for predicting characteristic information of potential customers, comprising the step of selecting a second next-generation customer group including at least one of the customers who have become customers. According to claim 1,
The step of selecting the second next customer group is
Through supervised training for the machine learning model, the method further comprising the step of outputting at least one or more characteristic information of customers included in the first next customer group, potential customer characteristic information prediction method. According to claim 1,
Characteristic information about the new campaign,
A method of predicting potential customer characteristic information, which is information entered as the most recent among campaign information input by an authorized user. According to claim 1,
The marketing rule is,
A method of predicting characteristics of potential customers, information updated by authorized users. According to claim 1,
The characteristic information for the new campaign and the marketing rules are,
A method of predicting characteristics of potential customers, information updated by authorized users. According to claim 1,
The method is
A third customer group including at least one customer included in the second next customer group by receiving a campaign result of conducting a campaign for customers included in the second next customer group, and reflecting the received campaign result Selecting a; further comprising, the potential customer characteristic information prediction method. 7. The method of claim 6,
The method is
Through supervised learning of the machine learning model, the method for predicting characteristic information of potential customers further comprising outputting a combination of characteristic information of customers included in the third next customer group. A computer-readable recording medium storing a program for executing the method according to any one of claims 1 to 7. Collect personal information of a plurality of customers from the database, customer data on the customer application usage information of the customers, campaign data on past campaigns for at least some of the customers, and logs of the customer application log by statistical processing A raw data collection unit that collects app statistical data of the customer group;
a profile calculation unit for calculating a customer profile, a campaign profile, and an app statistics profile based on a predetermined heuristic rule based on the customer data, the campaign data, and the app statistical data;
By applying the calculated customer profile, campaign profile, and app statistics profile to characteristic information on the new campaign, the first next customer group including customers effective for the new campaign among the plurality of customers is selected the first group selection unit; and
At least one of the customers included in the first next customer group is excluded by excluding some of the customers included in the selected first next customer group according to the exclusion conditions set based on the marketing rule, which is information not related to the new campaign A system for predicting characteristic information of potential customers, including a second group selection unit for selecting a second next-generation customer group that includes. 10. The method of claim 9,
The second group selection unit,
Through supervised training for the machine learning model, the system for predicting characteristic information of potential customers that outputs at least one characteristic information of customers included in the first next customer group. 10. The method of claim 9,
Characteristic information about the new campaign,
A system for predicting characteristics of potential customers, which is information entered as the most recent among campaign information input by an authorized user. 10. The method of claim 9,
The marketing rule is,
A system for predicting potential customer characteristics, information updated by authorized users. 10. The method of claim 9,
The characteristic information for the new campaign and the marketing rules are,
A system for predicting potential customer characteristics, information updated by authorized users. 10. The method of claim 9,
The system is
A third customer group including at least one customer included in the second next customer group by receiving a campaign result of conducting a campaign for customers included in the second next customer group, and reflecting the received campaign result A third group selection unit to select; further comprising, potential customer characteristic information prediction system. 15. The method of claim 14,
The system is
Through supervised learning of the machine learning model, the potential customer characteristic information prediction system further comprising a characteristic information output unit for outputting a combination of characteristic information of customers included in the third next customer group.

Images