JP2001023047A - Integrated customer management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an integrated customer management system which can detect a latent purchaser with considerably high probability by generating sales promotion data based on a customer attribute, a customer satisfaction degree and his/her purchase history. SOLUTION: A business man 30 has a management computer 35, and a purchase history data bases 41, a customer attribute data base 42 and a customer satisfaction degree data base 43 are installed in a hard disk 40 connected to the computer. A telemarketing server 36 is connected to the management computer 35 through a network. The telemarketing server 36 hears the customer attribute of a purchase 10 and the customer satisfaction degree of a purchased commodity from the purchaser 10 through a telephone set 12 and the like and stores them in the customer attribute data base 42 and the customer satisfaction degree data base 43 in the business man 30. Sales promotion data are generated based on the data. Thus, highly precise sales promotion data can be generated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an integrated customer management system for creating new sales promotion data based on past sales data, and more particularly to an integrated customer management system capable of finding potential buyers.

[0002]

2. Description of the Related Art A customer information collecting system that is interested in the present invention is disclosed, for example, in Japanese Patent Application Laid-Open No. Hei 10-187320. According to the publication, there is disclosed a customer information collection system that collects information such as material requests and requests from customers, analyzes the data, and uses the data for subsequent development and sales of advertisements.

[0003]

A conventional customer information collecting system has been configured as described above. Therefore, it is possible to know the past purchase history and data of the product etc. that the user is currently interested in. However, when the user is interested in a certain product, it simply collects the information of interest of the user. I did not have the idea of going forward and applying it to a new sales strategy.

[0004] The present invention has been made to solve the above-described problems, and has as its object to provide an integrated customer management system capable of finding potential purchasers with a considerably high probability.

[0005]

SUMMARY OF THE INVENTION An integrated customer management system according to the present invention stores at least two of a user's purchase history of a product, a user's purchase product satisfaction, and a user's attribute. Means and at least two of said customer attributes, customer satisfaction and purchase history,
Means for creating promotional data.

[0006] Sales promotion data is created based on at least two of user attributes, customer satisfaction, and purchase history, so that sales promotion data can be created with considerably high accuracy.

Preferably, the means for creating sales promotion data includes means for dividing a user into a plurality of groups based on purchase history data, and analysis of customer attributes or customer satisfaction for each of the divided groups. Is stored.

[0008] Since customer attributes or customer satisfaction are analyzed for each group divided based on purchase history data,
You can easily obtain the necessary sales promotion data.

Preferably, the means for creating sales promotion data includes means for dividing a user into a plurality of groups based on customer attribute data, means for analyzing customer satisfaction or purchase history for each divided group, Storage means for storing the analysis result.

Since the purchase history data or the customer satisfaction is analyzed for each group divided based on the customer attribute data, necessary sales promotion data can be easily obtained.

Preferably, the means for creating sales promotion data includes means for creating sales measure data based on customer satisfaction and purchase history data or customer attributes, and the means for creating sales measure data includes a predetermined measure. Means for detecting the relationship between customer satisfaction and importance for each item, and means for listing necessary measures based on the detected relationship.

Since the relationship between the customer satisfaction and the importance for each predetermined item is detected based on the customer satisfaction and the purchase history data or the customer attribute, necessary measures are listed based on the detected relationship. , You can easily know the measures required for sales promotion.

[0013] More preferably, the means for creating sales promotion data includes means for specifying a product that is highly likely to be newly purchased by the user.

[0014] More preferably, the means for creating sales promotion data includes means for specifying, based on the data of the user, an attribute having a high possibility of finding new customers.

As a result, potential buyers can be found with a considerably high probability. More preferably, the means for creating promotion data includes creating promotion data for a plurality of different companies, and using another company's promotion data using the promotion data of one of the plurality of different companies. Create data.

[0016] Sales promotion data is created for a plurality of different companies, and the sales promotion data of one company is used as sales promotion data of another company. Highly efficient sales promotion data can be provided.

[0017] In still another aspect of the present invention, the integrated customer management system includes a first storage means for storing a user's purchase history of goods, and a second storage means for storing customer satisfaction of the purchase goods of the user. A third storage unit for storing a user attribute of the user; a customer attribute stored in each storage unit;
Means for creating sales promotion data for the user based on the customer satisfaction and the purchase history.

Since the sales promotion data is created based on the three attributes of the user, the customer satisfaction, and the purchase history, more accurate sales promotion data can be created.

[0019]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the overall configuration of a network system to which the integrated customer management system according to the present invention is applied. The network system 100
The system includes a plurality of companies 20 connected to a network 101 such as the Internet, and a company 30 having an integrated customer management system according to the present invention. The purchaser 10 purchases a product of the company 20.

The purchaser 10 has, for example, a member's card 11 for the company 20 in advance, and means such as a bar code, a magnetic tape, an IC card, and the like for identifying the purchaser 10 are provided. If the member's card 11 has not been issued in advance, such information may be separately introduced by other means.

In the company 20, the sale of the goods is
This is done using a terminal. And that data is company 2
The management computer 3 of the business entity 30 is stored in the management computer 21 provided in the management computer 30 via the Internet.
Sent to 5.

In the company 20, the member's card 1
1 and the data from the POS terminal 22 to create purchase history data of the customer. The business 30 has a management computer 35 and a hard disk 4 connected thereto.
0 is provided with a purchase history database 41, a customer attribute database 42, and a customer satisfaction (hereinafter sometimes abbreviated as CS) database 43. A telemarketing server 36 is connected to the management computer 35 via a network. The telemarketing server 36 retrieves the customer attributes of the purchaser 10 and the customer satisfaction of the purchased product from the purchaser 10 via, for example, the telephone 12 and the like, and uses the information as the customer attribute database 42 and the customer satisfaction database 43 within the business. Store in

The data on customer attributes and customer satisfaction are as follows:
For example, by inputting a questionnaire attached to the purchased product via the company 20 or directly by the business operator 30 in a batch, the integrated customer management system 31 of the business operator is input.
May be stored at

FIG. 2 shows the company 20 and the business 30 shown in FIG.
FIG. 4 is a block diagram showing the relationship between the above and the present embodiment more clearly.
Referring to FIG. 2, only one target company 20 is shown in FIG. 1. However, in reality, the target company 20 is a company A, a company B, a company C, a company D, as shown in FIG. There can be more than one. On the other hand, a plurality of buyers 10 may be duplicated for each company. Here, the buyer 10 is generally considered to be a customer, and is a concept that includes not only ordinary consumers but also other trading companies. That is, the integrated customer management system according to the present invention is generally used in a form as shown in FIG.

FIG. 3 is a block diagram more specifically showing the contents shown in FIG. Referring to FIG. 3, company A 20a, company B 20b and company C 20c as companies 20 shown in FIG. 2 each have card POS system 22a and means 25a for collecting information by questionnaire and telemarketing from customers for products. have. These pieces of information are respectively stored in the purchase history database 4 constituting the integrated customer management system 31 on the side of the business 30.
1a, the customer attribute database 42a, and the customer satisfaction database 43a.

The same applies to Company B, Company C and Company D. The data collected in this way is analyzed for their mutual relations using a statistical method or a data mining method, and the data of companies A, B and C, which are different industries, are analyzed in a multidimensional manner. The details will be described later.

FIG. 4 is a schematic diagram showing the contents shown in FIG. 3 from a different side. For example, in the case of Company A, the mutual relationship among the purchase history database 41a, customer attribute database 42a, and customer satisfaction database 43a of the company alone is grasped. After that, companies B, C and D
Analyze company data to understand relationships.

Next, a specific data analysis method will be described. 5A to 5E show the customer attribute database 42.
FIG. 3 is a diagram showing an example of the data content of FIG. These data are stored at predetermined addresses in the customer attribute database 42. The customer attribute database includes basic data, demographic data, and psychological attribute data.

FIG. 5A shows basic data for communication, and FIG. 5B shows data represented by numerical attributes in demographic data, such as birth date and anniversary date. Is entered. Based on these data, it is possible to create a birthday special sale list, an entrance / graduation sale list, and a wedding anniversary special sale list.

FIG. 5C shows demographic data.
It has a discrete value attribute and a binary attribute, and is data such as gender, marriage / unmarriage, and occupation.

FIG. 5D shows the psychological attribute data.
This is data having a numerical attribute, for example, data such as the number of times of eating out.

FIG. 5E shows the psychological attribute data.
This is data represented by a discrete value attribute or a binary attribute, and is data such as hobbies, subscription papers, and food preferences. Subscription data can be used, for example, in combination with the purchase history database RFM score described later to find out which newspaper is the best customer for your company, so you can select a medium for mass advertising. Available.

In the psychological attribute data shown in FIG.
Accumulates data representing values that determine purchase / consumption and performs customer classification. For example, questions such as "Do you value the brand at the time of purchase?" Or "Would you like to spend your leisure alone or in large numbers?" The question item comprehensively prepares in advance questions that are considered to represent the values that determine purchase and consumption as described above. From the data representing the values that determine purchase and consumption, customers are classified into several groups having similar values. The method of dividing the group is as follows.

First, several basic elements (latent factors) considered to constitute each question item are extracted by a technique called factor analysis. The method of factor analysis will be described in the database of customer satisfaction. The factor analysis determines a factor score for each factor, which is the standard score for the latent factor for each sample. The distance of the factor score between each sample is obtained on the coordinates, and the closest samples are grouped. This is a statistical method called cluster division, and can also be performed by a data mining method.

Next, each of the groups divided by the values that determine purchase / consumption is defined as one of the demographic attributes.
Divide into several groups that are largely related to purchases and consumption, divided by income, and similarly divided into groups divided by life stage (for example, housewives who are raising children, housewives who are far from children). As a result, the customer is ultimately divided into several groups with three classification methods. Hereinafter, the group thus classified is referred to as an attribute group.

On the other hand, in order to have a common axis for opening and analyzing customer data among a plurality of companies in the same trade area, and to know the position of the analysis company in the trade area, In addition to data analysis, conduct market research on customer attributes. The question items are exactly the same. This method is as follows.

First, from consumers in a commercial area serving as a population,
Samples are selected using statistical techniques. With respect to this sample, a questionnaire for asking the above-mentioned customer attributes is conducted, and psychological attribute data indicating values for deciding purchase / consumption and various demographic data are asked by a questionnaire survey. From there, as described above, the basic values (latent factors) that determine purchase and consumption are extracted and used in statistics and data mining, a method called cluster analysis.
Categorize consumers into several groups. They are further divided into groups according to income and life stage.

Assuming that five groups (1, 2,
3, 4, 5) and 5 groups by income (A, B, C,
D, E), 5 groups (a, b,
When classified into c, d, and e), they are classified into, for example, 1Aa and 2Ac.

When performing cluster analysis based on values in the trade area in advance, a model of this cluster division is modeled by a data mining method, and when an individual company performs grouping based on values. , Using a method called class discrimination in data mining.

Here, data mining refers to finding hidden rules and regularities that humans do not notice by analyzing a large amount of collected raw data by computer.

For example, by performing a correlation analysis, mutually related data items can be found from among many data items (events). You can find the relationship itself, or find another data item that is closely related to a particular data item. In addition, by performing time-series analysis, data A (event)
It is possible to analyze the relationship such as what kind of data (event) occurs after a certain period of time when the occurrence occurs.

FIG. 6 is a diagram showing the contents of the customer satisfaction database 43. These data are also stored at a predetermined address of the customer satisfaction database 43. Referring to FIG. 6, the customer satisfaction database here shows data obtained by a questionnaire. The customer satisfaction database 43 includes a questionnaire implementation date, a customer code of company A as a specific example of a company, overall satisfaction, clerk correspondence, clerk correspondence, clerk knowledge, information provision, price range, assortment, and quality. , Design, functionality, location in the store, location of the store, atmosphere in the store, and the like are each given a five-point scale.
We clarify the relationship between total satisfaction and specific customer satisfaction data. This will be described later, for example
It is calculated from the data of the staff member, the guide member, and so on.

Next, a method for creating and analyzing a customer satisfaction database will be described. In the customer satisfaction database, as described above, the factor items are reduced from the actually obtained data by the factor analysis method. Factor analysis means that when trying to obtain knowledge about psychological constructs such as intelligence and human characteristics, we measured variables that are considered to be closely related to the construct of interest and obtained It is a method that examines the interrelationship between these variables based on measured values.

For example, when the intelligence is measured, there are various test items, and each test item has a smaller number of factors (= variables) constituting the intelligence that cannot be specified before the survey, less than the test items. Think of it as composed of The idea is to extract the factors and aggregate the relationships between the factors (= variables).

In the case of the customer satisfaction survey, assuming that there are 16 items for asking partial customer satisfaction, each item is evaluated by a 5-point method. The standard deviation is calculated from the questionnaire results, and the standard score is calculated from the standard deviation.
The correlation coefficient (matching degree) of each of the six question items is calculated. There are 120 correlation coefficients. The method for obtaining the correlation coefficient is expressed by Equation 1.

[0047]

(Equation 1)

Here, r in the equation (1)_jkIs the jth and kth
Represents the correlation coefficient of the question item of _ijIs j of the ith person
Represents the standard score for the th question item, z_iKIs the i-th person
Represents the standard score of the k-th question item.

In the equation (1), a factor analysis that the standard score of a specific question item (j-th question) of a certain person (i-th person) is the sum of the product of the standard score and the weight of a plurality of factors. By substituting the model equation (2) and the conditional equation defining the average and variance of the factor scores, the equation (3), which is a relational equation for obtaining the factor load, is obtained.

In the equation (2), z _ij is j of the i-th person.
A _i1 represents the standard score of the j-th question item, a _i1 represents the factor loading of the first factor in the j-th question item, and f _i1 represents i
Represents the standard score of the first factor given by the second person.

By solving this relational expression, the factor load a _jk (the degree of influence of each factor on each question item) is obtained. Next, a factor score f _ik (a standard score for each factor given by each respondent) is obtained. Then, this factor score becomes the degree of satisfaction of each factor. When a standard score for each question item is calculated, a standard score based on data of the entire industry is obtained, so that a factor score at an industry level can be obtained.

This standard score has an average of 0 and a variance of 1.
The deviation value t_ik= 10f _ikFlat with +50
A deviation value score with the average being 50 is obtained.

FIG. 7 shows the factor loading matrix A, in which variables 1 to p are taken in the row direction and factors 1 to m are taken in the column direction. The sum of squares of each is shown at the right end and the bottom end. Have been.

As shown in FIG. 7, the sum of squares of the factor loading of each factor becomes a factor contribution, and each factor contribution to the total factor contribution becomes a factor contribution rate. Factor interpretation is performed based on FIG. 7 with the first factor and the second factor in order from the one with the largest factor contribution.

Next, the influence (importance) of each factor on the overall satisfaction is determined by multiple regression analysis. The overall customer satisfaction indicates, for example, in the case of an inn, the customer wants to come back again when checking out and does not want to come again.

The degree of influence of each factor on the overall customer satisfaction is obtained by a multiple regression analysis method using the overall customer satisfaction as an objective variable and each factor as a dependent variable.

Equation (4) is a multiple regression equation.

[0058]

(Equation 2)

Here, y _i represents a standard score of the overall customer satisfaction of the ith person, f _i1 is a standard score given by the ith person to the first factor, and β ₁ , β ₂ , ... is the multiple regression coefficient, which is the degree of influence of each factor.

Note that, without performing the above-described factor analysis, the degree of influence (importance) of each question item itself on the overall customer satisfaction may be calculated by a statistical method, and used together with the above method.

Next, the purchase history database 41 will be described. FIG. 8 shows the contents of the purchase history database 41. The purchase history database 41 includes a customer code, a purchase date, a product name, a purchase price, a last purchase date, a recent purchase date (R points), a period purchase count, and a purchase frequency (F points).
, Cumulative period, cumulative amount (M points), and RFM total points.

Based on these purchase history databases,
You can create a thank-you direct mailing list after purchase, a follow-up direct mailing list after purchase, and a direct mailing list to promote replacement.

Next, from the purchase history database, R,
A method for obtaining F and M points will be described. RF
The M-analysis was announced in 1963 by the American Direct Mail Marketing Association as an effective model for determining customers. Here, R is Recenc indicating the date of last purchase
This is the first letter of y. It is considered that a customer who has recently purchased is likely to purchase again. Therefore, for example, the R score is given as follows.

Latest purchase date 24 customers under 3 months 24 latest purchase date 12 customers under 3-6 months Latest purchase date 6 customers under 6-12 months Latest purchase date 3 months or more customers 3 points It is an acronym for Frequency that represents the purchase frequency. It is based on the idea that frequent customers are good customers.
For example, four points are given for one purchase during a certain period.

M is an acronym of Monetary Value representing the purchase price. Based on the idea that customers who buy a lot are good customers. For the total purchase amount during a certain period,
A fixed amount is set as one point, and the points are scored.

Of course, the method of determining the score depends on the characteristics of the industry. Also, the weighting method for each RFM point changes in the same manner.

Next, a method of classifying customers according to the purchase history database will be described. The above R, F, and M points are arranged in order from the top according to the total points. For example, a customer with a certain score or more is called a sympa, a group with the next highest score is called a heavy user, and a group with the next highest score is classified as a middle user.

The boundaries of classification are determined based on the distribution status and industry characteristics. For example, in the tourist hotel industry, a visit is made four times a year, and a customer who consumes a specific amount or more is empirically set as a guide that is a heavy user. Hereinafter, the grouping based on the RFM analysis is referred to as an RFM group.

FIG. 9 shows the result of ranking customers by using such a purchase history database. By visualizing in this way, the analysis company reads the management issues such as increasing the number of sympa layers as a priority issue and the need to expand the base mainly on the user segment by cultivating new customers.

FIG. 10 is a diagram showing a conversion table of a unified code and a code of each company in a plurality of companies. This is stored on the hard disk 40 separately from other databases.
As shown in FIG. 10, the conversion table between the unified code and each company code includes a unified customer code, a name, an address, a date of birth,
It consists of customer codes of each company. Using this conversion table, it is possible to grasp the mutual relationship between a plurality of companies having different business types.

Incidentally, the statistical method as a method for analyzing the above three databases is not limited, and
For example, use principal component analysis instead of factor analysis,
Other methods that meet the original analysis purpose may be used.

Next, the analysis procedure of the integrated customer management system will be described. 11 to 24 are flowcharts showing the analysis procedure of the integrated customer management system. Here, first, a case in which the analysis is performed using the purchase history database 41 and the customer attribute database 42 will be described.

Here, in the analysis, each data of the purchase history database 41, the customer attribute database 42 and the customer satisfaction database 43 is read out from the hard disk 40 and is read from the management computer 3.
The program is executed in accordance with the flow chart shown below to obtain a desired correlation. Thereafter, the data is listed and graphed as necessary and output to the CRT or the printer 37 of the management computer 35.

Referring to FIGS. 11 and 12, first, the customer attribute database 42 and the purchase history database 41 are opened (step S11, steps are abbreviated hereinafter).
Next, grouping is performed based on the RFM points as shown in FIG. 8 from the data of the purchase history database 41 (S
12). Next, customer attributes are analyzed for each group (S13).

FIG. 12 is a flowchart showing the customer attribute analysis contents for each group shown in S12 of FIG.
Referring to FIG. 12, in the customer attribute analysis for each group, first, the user is divided into a sympa, a heavy user, a middle user, and a user (S21). For each group (S
22, S25, S28, S31), the customer attributes are analyzed (S23, S26, S29, S32), and the respective customer attributes are grasped (S24, S27, S30, S33).
Then, a response corresponding to each layer is performed (S34).

In addition, by cross-tabulating each group (RFM group) classified by RFM analysis and demographic data, characteristic attributes of each group become clear. For example, the tendency of each group, such as gender, age, annual income, educational background, and subscription paper, is clarified.

A general consumer having an attribute similar to the attribute of the group at the top of the RFM analysis can be presumed to have a high possibility of success in new customer development, and this attribute can identify a prospective customer for new customer development. .

For example, it is considered that it is most efficient to develop a prospective customer having the same attribute as a special customer who is the best customer. Therefore, the attributes (for example, age, income, educational background, It can be understood that it is only necessary to search for prospects that match the subscription paper). This will be described specifically.

FIG. 13 is a flow chart showing an example of the correspondence of the specially good layer in the correspondence according to each layer shown in S34 of FIG. Referring to FIG. 13, first, the most characteristic attribute data is input to the specially good layer (S41). next,
The data obtained by the market research, the customer data of other companies in the same trade area, and the like are sequentially input (S42). And
Among them, only the data that matches the data of the special customer specialty or the data that matches the predetermined range is listed as new pioneering customer data (Y in S43, S44).
Then, for example, a direct mail is sent using this data (S45).

In this case, if the attribute group and the RFM group are cross-tabulated and analyzed, the customer can be analyzed more deeply than the analysis of the demographic attribute alone. Attribute groups are purchased /
Based on the psychological attribute data that represents the values that determine consumption, it is classified by the factor analysis method and the cluster analysis method described above, and finally classified by income and life stage. This is because it is possible to determine what kind of people each group includes, including psychological aspects.

When this is compared with the attribute group in the trade area obtained by the above-mentioned market research, the following can be understood. Distribution of each attribute group in the trade area (people composition ratio)
Indicates the market potential of the trade area. Therefore, as compared with the distribution status of the attribute groups in the company, it is possible to specify a customer group to be developed in the future by the company in the business area. That is, our RFM
It can be specified that an attribute group that is present in many of the upper groups of the group and has a large market size is an attribute group to be developed in the future.

By analyzing as described above, it is possible to know the attributes of the high-class customers for the company and to know what attributes are targeted by the company to develop new customers.

Next, analysis by the customer satisfaction database 43 and the purchase history database 41 will be described. FIGS. 14 and 15 are flowcharts showing an analysis procedure by the customer satisfaction database 43 and the purchase history database 41. Referring to FIGS. 14 and 15, first, customer satisfaction database 43 and purchase history database 41 are opened (step S51). Next, grouping is performed based on the RFM points from the data of the purchase history database 41 (S52). Next, the customer satisfaction is analyzed for each RFM group (S53).

FIG. 15 is a flowchart showing the customer satisfaction analysis procedure for each RFM group shown in S53 of FIG. Referring to FIG. 15, the processing content is basically the same as the content shown in FIG. In the customer satisfaction analysis for each group, first, the user is divided into a sympa, a heavy user, a middle user, and a user (S61). For each group, the customer satisfaction is analyzed, the customer satisfaction is grasped, and a response corresponding to each layer is performed (S62 to S7).
4).

As for the customer satisfaction, the multiple regression coefficient obtained by the above-mentioned various analyzes represents the importance of each factor with respect to the total customer satisfaction, and the factor score is an evaluation point for each factor. Therefore, if you create a graph with the factor score on the vertical axis and the multiple regression coefficient on the horizontal axis, you need to decide which elements (factors) to improve in order to increase overall customer satisfaction most efficiently. Becomes clear. That is, a factor with high importance (multiple regression coefficient) and low satisfaction (factor score)
This is a priority improvement task.

Therefore, customer satisfaction is analyzed for each group (by RFM group) obtained by the RFM analysis.
Creating the above graph will highlight the priorities for improving overall customer satisfaction for each RFM group.

If it is an important issue to rank the middle user group in each group as a heavy user, if the above-mentioned graph of the middle user group is examined, the products / services can be improved in order to increase customer satisfaction. The direction of how to improve is clear.

FIG. 16 is a diagram showing the relationship between the customer satisfaction and importance of the middle user class obtained as described above.
Referring to FIG. 16, the overall customer satisfaction of the middle user segment is highest in the factor of “cheap price”,
It has been analyzed that the degree of satisfaction is low.

FIG. 17 is a flow chart showing an example of the correspondence to the middle user layer among the correspondence of each layer shown in S74 of FIG. The processing in this case will be described with reference to FIGS. First, the degree of satisfaction and importance for each item are input (S81). Among them, items having high importance and low satisfaction are extracted based on a predetermined reference value (S82, S83). It is listed and the countermeasures are taken (S85).

In the case described above, "low price"
It is analyzed that the factor has the highest weight and the satisfaction level is low, so it becomes clear that it is only necessary to develop a low-priced product for middle users as a countermeasure.

Although the relationship between customer satisfaction and importance is shown as a graph here for easy understanding, in actuality, for example, coordinate data is stored in a storage device such as a hard disk after being calculated by a computer. It is stored as

According to the above analysis, the customers who are good at the company are those who value what factors when purchasing and consuming the products and services of the company, and have high satisfaction with the elements. You can see that you have.

[0093] Further, it is known that a customer group with little purchase is a person who attaches importance to what kind of element when purchasing and consuming the company's products and services, and is not satisfied with the element. be able to. Therefore, in order to increase the purchase of such people, it is necessary to improve goods and services in this regard. In addition, it can be seen that it is only necessary to increase the lineup in consideration of this point. That is, it is possible to know the elements that should be possessed, such as a product.

Next, the analysis by the customer attribute database 42 and the customer satisfaction database 43 will be described. 18 and 19 show the customer attribute database 42.
6 is a flowchart showing an analysis procedure based on the data and a customer satisfaction database 43. Referring to FIGS. 18 and 19, first, the customer attribute database 42 and the customer satisfaction database 43 are opened (step S91). Next, the customers are classified into several attribute groups from the data in the customer attribute database 42 (S92). That is,
Here, customers are classified based on psychological attribute data by a factor analysis method and a cluster analysis method, further classified by income and life stage, and finally classified into several attribute groups. Then, the customer satisfaction is analyzed for each attribute group (S93).

FIG. 19 is a flowchart showing the customer satisfaction analysis procedure for each attribute group shown in S93 of FIG. Referring to FIG. 19, in the customer satisfaction analysis for each attribute group, first, it is divided into several attribute groups (S101). Here, for explanation, simply
For example, it is assumed that values are divided into two groups (1, 2), incomes into two groups (A, B), and life stages into two groups (a, b). Then, 1Aa, 1A
The customers are classified into eight attribute groups, b, 1Ba.

Next, the customer satisfaction is analyzed for each attribute group. That is, the customer satisfaction is tabulated for each attribute group, and data serving as a basis for creating a graph as shown in FIG. 16 in which the factor score is on the vertical axis and the multiple regression coefficient is on the horizontal axis is calculated (S103). Then, a response corresponding to each attribute group layer is performed (S104).

Next, referring to FIG. 20, S104 of FIG.
The correspondence corresponding to each group indicated by is described. In the correspondence in each group, first, satisfaction and importance of each item are input (S111). Next, improvement points are listed by importance and satisfaction. More specifically, when items having high importance and low satisfaction are listed (S112-S117), points of improvement are obtained. Items having high importance and high satisfaction are listed (S112).
112-S114) and the appeal point at the time of new development.

In this case, the priorities as improvement points are listed higher as the importance is higher and the satisfaction is lower.

In FIG. 20, the degree of satisfaction is determined after the importance is determined. However, the present invention is not limited to this. The degree of importance is determined after determining the degree of satisfaction first. Needless to say, it is good.

By clarifying the interrelationship between customer attributes and customer satisfaction by such an analysis procedure, when a customer with a certain attribute is newly cultivated, what points need to be improved for products / services? In other words, the needs according to the customer attributes become clear.

Therefore, when improving the purchase amount of a customer having a specific attribute or value, or when seeking to develop a prospective customer having a specific attribute or value, an improvement point of the product / service is improved. In addition, the promotion points for sales promotion and advertising will be clarified. In other words, factors with high importance (multiple regression coefficient) and low satisfaction (factor score) are priority improvement points, and factors with high importance (multiple regression coefficient) and high satisfaction (factor score) are emphasized. Is the point.

Note that, in step S93 of FIG.
The customer satisfaction may be analyzed for each question item of the demographic attribute data before, after, or in parallel with the customer satisfaction analysis for each attribute group. For example, customer satisfaction by gender, customer satisfaction by income category, and customer satisfaction by age are tabulated. Then, a graph in which the vertical axis indicates the factor score and the horizontal axis indicates the multiple regression coefficient for each classification is created. This analysis also reveals improvement points and appeal points for customers with certain attributes.

Next, a case where the purchase history database 41, the customer attribute database 42 and the CS degree database 43 are integrated and analyzed will be described.

FIG. 21 is a flowchart showing a procedure for integrating and analyzing these three databases. FIG.
Referring to FIG. 1, first, the purchase history database 41, the customer attribute database 42, and the CS degree database 43 are opened (S201). Next, the customers are classified into attribute groups by the customer attribute database (S202). To perform RFM grouping (S203).

Next, a customer satisfaction analysis (S204) is performed for each attribute group, a cross tabulation of the RFM group and the attribute group is performed (S205), and a customer satisfaction analysis (S206) is performed for each RFM group.

The above analysis results are listed for each RFM group (S207). Further, a list of measures is made for each attribute group (S208).

FIG. 22 shows the RF shown in S207 of FIG.
It is a flowchart which makes the said analysis result into a list for every M group. Referring to FIG. 22, grouping by RFM group is first divided into a sympa, a heavy user, a middle user, and a user (S212, S218, S22).
4, S230). A customer satisfaction analysis is performed for each group (S213, S219, S225, S231), and RF
Cross-tabulate the M group and the attribute group (S214,
(S220, S226, S232) Further, a customer satisfaction analysis for each attribute group is performed (S215, S221, S232).
227, S233), and list the above analysis results (S216, S222, S228, S234). By converting this list into a story, the following customer profiles can be drawn (S217, S223, S229, S2)
35).

For example, the company's superior customer is a man who is in his forties and has an annual income of about ○ 10,000. In his spare time, he prefers to act as a family or a group and prefers outdoor hobbies.
When purchasing their products and services, they attach great importance to responding to staff and providing information on products, and have a high level of satisfaction with this point.

[0109] On the other hand, the customer group with few purchases is mostly unmarried women in their twenties, and is sensitive to the trend but emphasizes economic efficiency.
When purchasing their own products and services, they are people who value their prices and understand that they are not satisfied with this point (the prices are high).

The profile of the customer in the company can be utilized for the development and improvement of products and services.

A general consumer who has the attribute of a good customer is considered to be a prospective new prospective prospect of the company, but the distribution of the attribute groups in the business area described in the customer attribute database is checked. Thus, people with this attribute can determine if the market size is sufficient or if another attribute should be targeted.

In the case of newly developing a general consumer who has the same attributes as the customer, the appeal points of DM and advertisements are elements that are emphasized by the people and have a high degree of satisfaction, that is, regarding the reception of staff and products. The level of information provision is high.

If another attribute is set as a target for new development, a customer satisfaction analysis for each customer attribute is examined, and an element having a low customer satisfaction is improved among elements which are important for people having this attribute. It is necessary to have an assortment of products. This is the product
This is a service improvement point.

On the other hand, when increasing the purchase of the above-mentioned low-purchase customer groups, and by observing the distribution of the attribute groups in the trade area, it is determined that the attribute groups characteristic of the low-purchase customer groups are promising markets. If possible, it turns out that there is a need for an assortment of low-priced products that suits the people of this profile.

Next, a description will be given of a list of measures for each attribute group (S208 in FIG. 21). With reference to FIG. 23, the distribution status of the attribute group of general consumers (each of which exists in the Enter the market size of the attribute group) (S2
21). Next, the customer attribute group distribution status for each RFM group of the analysis company is input (S222). Next, the total customer satisfaction, the satisfaction of each element item (factor), and the importance are sequentially input for each attribute group of the analysis company (S223).

First, an attribute group having a high overall customer satisfaction and a high RFM score is classified as the highest priority customer. Further, a general consumer having the same attribute as this attribute group is a target to focus on new customer development (Y in S224 and S225, S226). Correspondences to these are listed as in S227 and S228.

Next, an attribute group that has a high overall satisfaction but a low RFM score (does not purchase much) is classified as an attribute group that maintains the current status (Y, S224).
N in S225, S229).

Next, an attribute group having a low overall customer satisfaction but a high RFM score is classified as a customer requiring attention (N in S224, Y in S230, S231), and S232.
The countermeasures are listed as follows.

Even if the attribute group has a low overall satisfaction and a low RFM score, if it is an attribute group that has a large number in the trade area (has marketability), it is classified as an attribute group that needs to develop new customers. (N in S224, S230, Y in S233, S234). The countermeasure in that case is S
It is listed up as 235.

Attribute groups that have low overall satisfaction and low RFM scores and do not exist in the commercial area (no marketability) are classified as attribute groups that should be maintained as they are (S224, S230, S230). N in S233, S236).

By listing the analysis results in this manner, the evaluation and the corresponding method for each attribute group become clear.

In the above analysis, factors are extracted by the customer satisfaction analysis. In parallel with this, the influence (importance) of the original question items on the overall customer satisfaction is statistically analyzed. It may be calculated and used by a dynamic method.

In addition, when the customer attribute database 42 and the purchase history database 41 are analyzed by a data mining technique, it is found that people having a specific attribute purchase a product a and then have a probability of%, and then a b product. To buy. In addition, it can be seen that there is a correlation between the product a and a specific hobby. As a result, new sales promotion campaigns can be found.

Further, if the database of the purchase history and customer satisfaction of the same customer is analyzed by the data mining method described above without dividing or assuming in advance, it is found that there are items having purchase history and items having customer satisfaction. Correlations can be found. That is, computer analysis of a large amount of data may reveal hidden regularities between specific items.

Next, an analysis example considering data of a plurality of companies will be described with reference to FIG. First, the analysis method will be described. From FIG. 10, a common customer of company A and company B is extracted from a comparison table of the unified customer code of a plurality of companies and the customer code of each company. For a common customer, the purchase history at Company A and the purchase history at Company B are simultaneously extracted (S3).
01), and analyzed by a data mining technique (S30)
2).

Here, it is determined whether or not there is a parallel purchase pattern between the product A of the company A and the product b of the company B (S30).
3). If there is a parallel purchase pattern, a parallel purchase pattern list is created (S303Y and S304), and if not, another company is examined (S305). Based on the obtained parallel purchase list, a list of customers who have purchased the product a of company A is extracted from the purchase history database 41a, obtained by company B, and promoted the sale of product b. Then, the purchase rate is high.

Further, for common customers, the purchase history of Company A, the purchase history of Company B, and the customer attribute are simultaneously extracted and analyzed by data mining.
Hidden correlations between company buying patterns and specific customer attributes can be discovered. Therefore, if a customer having this attribute is extracted from the customer attribute database of Company A and acquired by Company B and sales promotion is performed, the purchase rate is high.

The group (attribute group) finally classified according to the psychological attribute data, the income and the life stage, which represents the values for determining purchase / consumption,
Company A and Company B are compared, and between the top group of A company's RFM analysis and the top group of B company's RFM analysis,
If there are many identical attribute groups, it can be understood that if the customer data is shared between the companies A and B, the data can be utilized by both companies.

Therefore, suppose that the automobile dealer A company
When the home sales company B, the electric company C, and the record store D publish customer data to each other, the following interrelation may be discovered.

A customer who purchases a six-seater one-box car of Company A has a strong tendency to relocate the house next.

An analysis of the attributes of a customer who buys a 6-seater one-box car and a house at the same time as Company A revealed that the proportion of children who had two or more children was high, and that the proportion of parents living together was high. Then, it can be inferred that as the child grows up, he wants more cars and more housing.

Therefore, if the company B obtains the data of the person who purchased the six-seater one-box car of the company A and guides the two-family house, it is understood that the purchase probability is high.

Further, a person who purchases a house is likely to purchase a home electric appliance on that occasion. Therefore, the company C obtains data of the person who purchases the house of the company B and guides the home electric appliances. It is also possible to guide home appliances that match the customer attributes.

Further, when an audio device is purchased among the home appliances of the company C, it is understood that the person is a music lover. Therefore, if the record shop D obtains the customer data, it is highly likely that the information will be purchased if the information of the CD or the information of the concert is provided.

Further, the customer attributes of Company A and Company B are extracted, and rearranged by the attribute related to the life stage from among them. The life stage is a concept that divides a person's life into several stages. For example, if sorting is performed in accordance with the age of a child, joint marketing of products closely related to the life stage can be developed among a plurality of companies. As an example,
The obstetrics and gynecology customer data can then be used as customer data for the diaper industry, which can then be used as baby store data.

The embodiments disclosed this time are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing an overall configuration of a network system to which an integrated customer management system according to the present invention is applied.

FIG. 2 is a block diagram when customer information from a plurality of companies is considered.

FIG. 3 is a block diagram when data of a plurality of companies is considered.

FIG. 4 is a schematic diagram in a case where a mutual relationship is grasped using data of a plurality of companies.

FIG. 5 is a diagram showing the contents of a customer attribute database.

FIG. 6 is a diagram showing the contents of a customer satisfaction database.

FIG. 7 is a diagram used for factor interpretation in a customer satisfaction database.

FIG. 8 is a diagram showing the contents of a purchase history database.

FIG. 9 is a diagram showing a state in which each customer is grouped by RFM score.

FIG. 10 is a diagram showing a conversion table between a unified code and each company code.

FIG. 11 is a flowchart showing an analysis procedure using a purchase history database and a customer attribute database.

FIG. 12 is a flowchart illustrating a customer attribute analysis procedure for each RFM group.

FIG. 13 is a flowchart showing a procedure of corresponding contents of a specially-excellent layer.

FIG. 14 is a flowchart illustrating an analysis procedure using a customer satisfaction database and a purchase history database.

FIG. 15 is a flowchart showing a customer satisfaction analysis procedure for each RFM group.

FIG. 16 is a diagram showing a relationship between customer satisfaction and importance of a middle user group.

FIG. 17 is a flowchart showing a procedure for responding to each layer in the customer satisfaction analysis.

FIG. 18 is a flowchart showing an analysis procedure using a customer attribute database and a customer satisfaction database.

FIG. 19 is a flowchart showing a procedure for analyzing customer satisfaction for each attribute group.

FIG. 20 is a flowchart showing a procedure for determining correspondence for each group.

FIG. 21 is a flowchart showing a procedure for performing analysis using three databases.

FIG. 22 is a flowchart showing a list of analysis results for each RFM group.

FIG. 23 is a flowchart showing a procedure for creating a list of measures for each attribute group.

FIG. 24 is a diagram showing a procedure for analyzing databases of a plurality of companies.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Buyer 11 Member's card 20 Company 21 Management computer 22 Card POS 30 Business operator 31 Integrated customer management system 35 Management computer 36 Telemarketing server 37 Printer 40 Hard disk 41 Purchase history database 42 Customer attribute database 43 Customer satisfaction database 100 Network system 101 Network

Claims (8)

[Claims] A storage unit configured to store at least two of a user's product purchase history, a user attribute of the user, and customer satisfaction of the user's purchased product; An integrated customer management system, comprising: means for creating sales promotion data for a user based on at least two of the user's customer attributes and customer satisfaction of the purchased product of the user. 2. The means for creating the sales promotion data includes: means for dividing a user into a plurality of groups based on the purchase history data; and analyzing the customer attribute or the customer satisfaction for each of the divided groups. The integrated customer management system according to claim 1, further comprising: a storage unit configured to store the analysis result. 3. The means for creating the sales promotion data includes: means for dividing a user into a plurality of groups based on the customer attribute data; and analysis of the customer satisfaction or the purchase history for each of the divided groups. The integrated customer management system according to claim 1, further comprising: a storage unit configured to store the analysis result. 4. The means for creating the sales promotion data includes means for creating sales measure data based on the customer satisfaction and the purchase history data or the customer attribute, and means for creating the sales measure data. The integration according to claim 1, further comprising: means for detecting a relationship between the customer satisfaction and importance for each predetermined item; and means for listing necessary measures based on the detected relationship. Type customer management system. 5. The means for creating the sales promotion data comprises:
2. The integrated customer management system according to claim 1, further comprising means for specifying a product that is highly likely to be newly purchased by the user. 6. The means for creating the sales promotion data comprises:
2. The integrated customer management system according to claim 1, further comprising a unit that specifies an attribute having a high possibility of cultivating a new customer based on the data of the user. 3. 7. The means for creating the sales promotion data comprises:
The method according to claim 1, wherein the promotion data is created for a plurality of different companies, and the promotion data of another company is created using the promotion data of one of the plurality of different companies. Integrated customer management system. 8. A first storage unit for storing a user's product purchase history, a second storage unit for storing customer satisfaction of the user's purchased products, and a third storage unit for storing the user's customer attributes. An integrated customer management system including means for creating sales promotion data for the user based on customer attributes, customer satisfaction, and purchase history stored in each of the storage means.