JP2002117048A - Purchase motive manifesting method, and method and device for adaptive article and customer extraction - Google Patents

Abstract

(57) [Summary] [Problem] To extract a product group corresponding to a customer's purchase motivation from a past purchase performance of the customer and provide useful information for the customer's product selection. [MEANS FOR SOLVING THE PROBLEMS] One parameter is a list of similar and diverse products, and the other parameter is a list of motivation candidate groups that can be a motivation for determining purchase of a specific type of product from a group of products of that type. An edge map storage unit 602 storing a group of motivations for determining purchase of the product type for each type; a purchase product list storage unit 612 storing a purchase product list for each customer; a purchase product list for each customer Means 604 for accumulating the purchase motivation corresponding to the product type included in the table and calculating the customer-edge correspondence map
And means 604 for searching for a purchase motivation group of an unpurchased product type group of a specific customer using the motivation group included in the customer-edge correspondence map of the customer as a key, and included in the customer-edge correspondence map of the customer. Extract product types that share purchase motivation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing method for processing past purchase result information and converting the information into information useful for future sales and production activities, and an information processing apparatus therefor.

[0002]

2. Description of the Related Art In the case of products of various types, for example, books, music CDs, computer game software, passenger cars, and the like, customers select and purchase a specific type from a group of various types of similar products. There are so many product types, and customers often have difficulty choosing the type that suits them. Distributors often have difficulty in determining the purchase type and purchase quantity. Product suppliers often have difficulty predicting the type of product required by their customers and have difficulty determining production quantities. It is also difficult to determine development goals. By analyzing past purchase results, various marketing techniques for the future have been developed. The technology that seems to be the most advanced at present is to use the past purchase information of “customer-purchased product type” as “customer attributes (age, occupation,
(Gender, annual income, etc.) "as a key to obtain data of" customer attribute--group of product types frequently purchased by the customer of the attribute ", and the customer corresponds to the attribute to which the customer belongs. Presenting the product type group and providing auxiliary information for product selection, and providing the store with attributes that prefer a specific product type and providing auxiliary information for selecting customers to develop sales promotion activities To the product supplier, providing data of "customer attributes-groups of products purchased frequently by customers of the attributes-past purchase quantities" to provide auxiliary information for determining production quantities and development goals. be able to.

[0003]

However, the conventional purchasing information analysis technology only analyzes using the attributes of customers as keys, and does not go into individual characteristics of customers. Therefore, the reliability of the information obtained by the analysis is low. For example, the customer believes the recommended product type prepared based on the analysis result of “customer attribute—a product type group frequently purchased by the customer of the attribute”. When a customer makes a purchase, the customer is often unsatisfied with the product type.

[0004] The present invention provides a technique that enables an analysis that reflects the individuality of each customer, rather than classifying customers by attributes. For example, consider computer game software. At present, there are about 1 million purchasers and about 1000 kinds of products. The past purchase performance per person is about 5 pieces,
There are ten thousand "customer-purchased product type" data. It is easy to say that analyzing this while reflecting the individuality of each person is difficult, but I do not know what to do.

[0005] When the customer A purchases the product ABC and the customer A purchases the product DEF, and there is 1 million purchase history information, the customer A is useful while reflecting the difference in personality between the customer A and the customer A. Information, for example, a customer who has purchased the product ABC recommends a product that matches the personality of the customer, or a store can provide a new product G based on the personality of the customer group of the store.
I don't know how to get useful information to determine the amount to buy.

[0006]

According to the present invention, the above-mentioned information, that is, the customer A purchases the product ABC and the customer A purchases the product DEF from the past sales performance information indicating that the personality of the customer is Transform into useful information, reflecting differences. For this purpose, the present invention uses the edge theory. The edge theory was developed by the present inventors,
This can be roughly explained as follows. (1) The number of factors that a customer weighs in the process of selecting a specific type from similar products having many types, for example, computer game software, books, passenger cars, etc., is finite, and a product sales expert has a factor group. Can be listed. For example,
In the case of computer game software, purchasing from "highly rated in magazines, etc."; purchasing from "stimulated instinct"; purchasing from "low price"; purchasing from "reliable maker" In some cases, product sales professionals with many products and many customers,
Factors to be considered in the process of determining the purchase type by the customer can be listed almost without omission. For example, in the case of computer game software, there are 57 factors, and it has been confirmed that almost all customers select a specific type by considering the 57 factors. In the edge theory, a group of factors that the customer considers in the process of selecting a specific type is referred to as a group of purchase motivation candidates (edge group). (2) Each product type has several purchase motivations in the group of candidate purchase motivation, and when the list of candidate purchase motivation is obtained, the expert of the product sale will give the individual product The motivation group that determines purchasing can be specified. For example, for product A, there are a customer who purchases from “highly evaluated in magazines and the like” and a customer who purchases because of “stimulated instinct to instinct”. Can be accurately specified that there are few customers to purchase from. In the edge theory, a database that stores a group of motivations for deciding purchase of a product type in association with the product type is called an edge map. (3) The above information is provided, for example, by an experienced physician who has a lot of knowledge necessary to identify the cause from the patient's symptoms, and prepares an expert database to create a database. This is almost equivalent to a doctor who has little experience in making a diagnosis at the same level as a doctor who has much experience, and is equivalent to an expert database for product sales. (4) Once the edge map is constructed, various information processing is possible. For example, a customer who has purchased a product ABC in the past has a personality that prefers a product that is “highly evaluated in magazines or the like” or a product that “stimulates the fighting instinct” from an edge map, and recommends a product that has a purchase motivation. Then, while it can be expected that the degree of satisfaction is high, it can be understood that the person has a personality that does not emphasize “price” and “manufacturer”, and the degree of satisfaction is low even when selecting a product from that viewpoint. Alternatively, the new product X has purchase motivation p, q, r, etc., and the new product X has customers C, D
Has a strong satisfaction. In the present invention, the edge theory is embodied in the following method or apparatus to generate useful information.

This device has an edge map storage means 100 as shown in FIG.
In the edge map 102 stored here, one parameter is used as a similar and various product list. In the case of FIG. 1, product types A, B,... Are shown on the horizontal axis as parameters. Thick arrows in the figure indicate that the same type of data continues. Another parameter of the edge map 102 is a list of motivation candidate groups that can be a motivation for determining purchase of a specific type of product from a group of such products. In the case of FIG. 1, the vertical axis indicates the purchase motivation candidate groups a, b,. Here, for example, “a” is “highly evaluated in magazines and the like”, “b” is “stimulated instinct”, “c” is “low price”, and “d” is “high reliability of the manufacturer”. is there. The edge map 102 stores, for each product type, a group of motivations for determining purchase of the product type. In the case of FIG. 1, the product type A is sometimes purchased because it is a and sometimes it is purchased because it is g, but it is illustrated that it is not purchased with the motivation from b to f. The apparatus according to claim 1 has means 104 for searching the storage contents 102 of the edge map storage means 100 using a purchase motive of a specific type of product or a specific purchase motive 106 as a key. Reference numeral 108 in FIG. 1 exemplifies a product group sharing a purchase motive (a and g in this case) with the product type A. In this case, a type F sharing a and g, a type DE sharing a, and a G sharing g are extracted. In addition, for type F, type A sharing two purchasing motivations cf among the purchasing motivations (abcfg in this case)
It illustrates that the CE was extracted. 110, the product type ADEF having a is extracted for the purchase motive a,
This illustrates a case where a product CF is extracted for a purchase motivation cf (AND condition).

[0008] According to the present apparatus, it is possible to extract a product type group having a similar purchase motive from a variety of similar products. When the apparatus is operated, it may be found that completely unexpected product types actually have similar purchase motives. For example, two computer game softwares with different genres, “Golf for Everyone 2” and “Dance Dance Revolution” are “highly rated in magazines, etc.”, “recommended by friends”, “because they are a reliable maker”, “play against someone” It turns out that they share many motivations to buy because they can do it, because they are familiar with the game material, and because they are likely to be good. In many cases, many customers actually purchase the similar product group found in this way. In addition, it has been confirmed that a customer who purchases one and recommends an unpurchased product to a customer who has not purchased the other has a high degree of satisfaction. In other words, it has been confirmed that the products have similar purchase motives, in other words, similar product types based on the personality of the purchasing customer.

When similar products are extracted in this way,
It can recommend product types with a high degree of satisfaction to customers, provide useful information to dealers to accurately determine purchase quantities, and determine the production volume to product suppliers without excess or shortage. Can provide useful information to help set development goals and provide useful information to those engaged in promotional activities to extract similar data from past promotional activities. Yes, very effective.

(Method of Claim 2) This method is conceptually shown in FIG. 2 and is executed by the apparatus of claim 1. In this method, a step 200 of constructing the edge map 102 and a step 204 of searching the contents of the edge map 102 by using a purchase motive of a specific type of product or a specific purchase motive as a key are executed to purchase from among a variety of similar products. The product type group 208 having similar motivation is extracted. The conditions for defining the edge map are the same as in the case of the first aspect of the present invention.

(Embodiment 3) As shown in FIG. 3, this apparatus also has an edge map storage means 3 which stores an edge map 302, similarly to the apparatus shown in FIG.
With 00. This device has a storage means 303 for an emotional personality map 305 that stores a specific purchase motivation in a group of purchase motivation candidates corresponding to a personality at the time of purchase decision. This differs from the device of claim 1. For example, when purchasing computer game software, there may be a person or a person who decides to make a purchase as a fan of the appearing character, or a person who decides to make a purchase by trying out game skills. In this case, in order to extract similar products from a group of product types that are often purchased as fans,
Similar disagreement in purchasing motivation when testing game skills is the main issue is not very problematic. Therefore, in the present invention, there is provided a storage means 303 for the emotional personality map 305 which stores a specific purchase motive in the purchase motive candidate group in accordance with the personality at the time of purchase decision. Note that the same person may make a purchase decision as a fan, or the game skill may come to the forefront to make a purchase decision, and there are a plurality of personalities within the same person.

In this device, the edge map storage means 30
Means 304 for retrieving the stored content 302 of 0 using a purchase motive corresponding to a specific personality as a key.
A group of product types having a purchase motivation adapted to a specific personality (for example, a personality that determines purchasing as a fan, or a personality that comes out to try out game skills to determine purchasing) is extracted from similar and diverse products. For this reason, a similar product when purchasing as a fan or a product type adapted to the personality is extracted, and more useful information is obtained.

(Method of Claim 4) This method is conceptually shown in FIG. 4 and is executed by the apparatus of claim 3. In this method, a step 400 of constructing an edge map 302 and an emotional personality map 3 storing a specific purchasing motivation in a group of purchasing motivation candidates corresponding to a personality at the time of purchasing decision.
05, and a step 404 of searching the contents of the edge map 302 using the purchase motivation corresponding to the specific personality as a key to execute a product having a purchase motivation adapted to the specific personality from among a variety of similar products. The type group 408 is extracted.

(Apparatus of Claim 5) This apparatus is an apparatus for eliciting a group of purchasing motivation on which a customer relies upon selecting a product, and as shown in FIG.
And a purchased merchandise list storage unit 512 that stores a purchased merchandise list for each customer, and a unit 514 that accumulates purchase motivation corresponding to a product type included in the purchased merchandise list for each customer.
And outputs a purchase motivation group list 516 for each customer. According to this device, for each customer, a group of purchasing motivations on which the customer relies upon when selecting a product is clarified based on past purchase results and output.

(Apparatus of Claim 6) This apparatus is an apparatus for extracting and presenting a product type which is expected to have a high degree of satisfaction to the customer who is uncertain about product selection. As shown in FIG.
And a purchased product list storage unit 612 storing a purchased product list for each customer, and a purchase-motivation corresponding to the product type included in the purchased product list for each customer, to calculate a customer-edge correspondence map 616. The means 614 and the purchase motivation group of the unpurchased product type group of the specific customer are determined by the customer's customer-
A means 604 for searching using a group of purchase motivations included in the edge correspondence map 616 as a key is used to extract a product (conforming product 618 of the customer) sharing the purchase motivation included in the customer-edge correspondence map 616 of the customer.

As described in claim 5, the customer-edge correspondence map 616 organizes a group of purchasing motivations on which the customer depends upon selecting a product. This is based on past sales records and is objective. This apparatus has a means 604 for searching for a purchase motivation group of an unpurchased product type group of a specific customer using the motivation group included in the customer-edge correspondence map 616 of the customer as a key. A product type having a purchase motive similar to the purchase motive of the purchased product type is extracted. The product type extracted in this way is a product expected to have a high degree of satisfaction for the customer. In fact, it has been confirmed that the products extracted in this way can obtain high satisfaction.

(Method of Claim 7) This method is conceptually shown in FIG. 7 and is executed by the apparatus of claim 6. In this method, a step 700 of constructing an edge map 602, a step 712 of collecting a purchased product list for each customer, and accumulating the purchase motivation corresponding to the product type included in the purchased product list for each customer to obtain a customer-edge A step 714 of constructing a correspondence map 616 and a step 704 of searching for a purchase motivation group of an unpurchased product type group of a specific customer using the purchase motivation group included in the customer-edge correspondence map 616 of the customer as a key are executed. By this method, the product types sharing the purchase motive included in the customer-edge correspondence map 616 of the customer are extracted.

(Apparatus of Claim 8) This apparatus is an apparatus for extracting a product type which is expected to have a high degree of satisfaction for a customer who is uncertain about product selection. In this respect, it corresponds to the device of claim 6. However, the apparatus according to claim 8 corresponds to the fact that there are a plurality of personalities among the same person, and extracts a product type expected to have a high degree of satisfaction when evaluated from a specific personality of the customer. I do.

As shown in FIG. 8, the apparatus includes an edge map storage unit 800, a purchased product list storage unit 812 storing a purchased product list for each customer, and a product included in the purchased product list for each customer. Means 814 for accumulating the purchase motivation corresponding to the type to construct a customer-edge correspondence map 816, and an emotional personality map storing the specific purchase motivation in the purchase motivation candidate group corresponding to the personality at the time of the purchase decision. The storage unit 803 and the emotional personality map storage unit 803 associate the purchase motivation group of the unpurchased product type group of the specific customer with the specific personality of the purchase motivation group included in the customer-edge correspondence map 816 of the customer. Means 804 for searching using stored purchase motivation group as a key
, And a product type 820 suitable for the purchase motive corresponding to the specific personality of the customer is extracted. According to this device, when evaluated based on a specific personality of a customer, a product type with a high degree of satisfaction is extracted.

(Method of Claim 9) This method is conceptually shown in FIG. 9 and is executed by the apparatus of claim 8. In this method, a step 900 for constructing an edge map, a step 912 for collecting a list of purchased products for each customer, and a step of accumulating purchase motivations corresponding to the product types included in the purchased product list for each customer to obtain a customer-edge correspondence. Step 9 of building map 816
14, a step 903 of constructing an emotional personality map storing the specific purchase motivation in the purchase motivation candidate group corresponding to the personality at the time of the purchase decision, and the purchase motivation group of the unpurchased product type group of the specific customer. And executing a step 904 of searching for the purchase motivation group stored in the emotional personality map as a key in association with the specific personality of the motivation group included in the customer-edge correspondence map 816 of the customer, and A product type 920 suitable for the purchase motivation corresponding to the specific personality is extracted.

(Apparatus of Claim 10) This apparatus is an apparatus for extracting customers who will have a high degree of satisfaction with new products or stocked products. With this device,
Efficient sales promotion activities can be performed.

As shown in FIG. 10, the apparatus includes an edge map storage unit 1000, a purchased product list storage unit 1012 storing a purchased product list for each customer, and products included in the purchased product list for each customer. Means 1014 for accumulating the purchase motivation corresponding to the type to calculate the customer-edge correspondence map 1016; the purchase motivation group of the unpurchased customer group of the specific product type 1022, and the purchase motivation group of the specific product type 1022 as a key. A customer 102 having search means 1004 and matching the purchase motive of the specific product type 1022
4 is extracted. According to this device, a customer who will have a high degree of satisfaction with a specific product selected from new products or stock products is extracted.

(Method of Claim 11) This method is conceptually shown in FIG. 11 and is executed by the apparatus of claim 10. The method includes a step 1100 of constructing an edge map, a step 1112 of collecting a list of purchased products for each customer, and a process of accumulating purchasing motivations corresponding to product types included in the purchased product list for each customer to obtain a customer-edge correspondence. A step 1114 of calculating a map and a step 1104 of searching for a purchase motivation group of an unpurchased customer group of the specific product type 1122 using the purchase motivation group of the specific product type 1122 as a key are executed. As a result, the customer 1 matching the purchase motive of the specific product type 1122
124 is extracted.

(Apparatus apparatus according to claim 12) This apparatus extracts customers who will have a high degree of satisfaction with a new product or a stock product. This apparatus extracts customers who will have a high degree of satisfaction with one personality, in response to the fact that there are multiple personalities within the same person.
With this device, efficient sales promotion activities can be performed.

As shown in FIG. 12, the apparatus includes an edge map storage unit 1200, a purchase product list storage unit 1212 storing a purchase product list for each customer, and a product included in the purchase product list for each customer. Means 1214 for accumulating the purchase motivation corresponding to the type to construct a customer-edge correspondence map 1216, and an emotional personality map storing the specific purchase motivation in the group of purchase motivation candidates corresponding to the personality at the time of purchasing decision. The storage unit 1203 and the purchase motivation group of the unpurchased customer group of the specific product type 1222 are searched by associating the purchase motivation group stored in the emotional personality map storage unit with the personality corresponding to the specific product type 1222 as a key. Means 1214. This apparatus extracts a customer 1224 having a personality that matches the purchase motive of the specific product type 1222.

(Method of Claim 13) This method is conceptually shown in FIG. 13 and is executed by the apparatus of claim 12. The method includes a step 1300 of constructing an edge map, a step 1312 of collecting a list of purchased products for each customer, and a process of accumulating purchase motivations corresponding to product types included in the purchased product list for each customer to obtain a customer-edge correspondence. A step 1314 of constructing a map, a step 1303 of constructing an emotional personality map storing a specific purchase motivation in the purchase motivation candidate group corresponding to the personality at the time of purchase decision, and an unpurchased customer of the specific product type 1322 A step 1304 of searching the group of purchase motivation groups by using the purchase motivation group stored in the emotional personality map storage unit as a key in association with a personality matching the specific product type 1322 is executed. By this method, a customer 1324 having a personality matching the purchase motive of the specific product type 1322 is extracted.

[0027]

FIG. 14 shows a system configuration of an apparatus according to the present invention, which is a retail store that sells merchandise (the merchandise is not limited to tangible objects, but includes intangible merchandise, that is, services. Time deposits for various periods are similar products, and time deposits for each period are each product type, and bank stores are retail stores, and many retail stores are bundled. An analysis center, a product supplier, a sales promotion source, and a customer are involved. Each subject has an information processing device represented by a personal computer, a retail store has a retail store information processing device 1402, a product supplier has a product supplier information processing device 1404, and a sales promotion activity source is a sales promotion activity. The original information processing device 1406 is provided, the customer is provided with a customer information processing device (which may be an I-mode compatible mobile phone) 1408, and the analysis center is provided with an analysis center information processing device 1412. All information processing apparatuses are connected by a computer network communication network 1410 represented by the Internet, and constitute various apparatuses of the present invention as a whole.

The retail store information processing device 1402 is connected to a store register, and collects information on which customer purchased which product type and sends it to the analysis center information processing device 1412 every time the customer makes a purchase. The analysis center information processing apparatus 1412 aggregates “information about which customer has purchased which type” sent from the retail store and creates a “customer-purchased product type list 1414” for each customer. While most retailers subscribe to the system of FIG. 14 while customers may purchase at different retailers, aggregation at the analytics center provides an almost complete “customer-purchased product list”. 1414 "are collected. The analysis center information processing apparatus 1412 includes two types of expert databases (hereinafter, referred to as D / B) 1416. One is an edge map 1418 and the other is an emotional personality map 1420. Edge map 141
8, as exemplified in FIG. 15B, one parameter is a list of similar and various products, and another parameter is a motive that can be a motive for determining purchase of a specific type of product from a group of such products. As a candidate group list, a two-dimensional map using the two types of parameters stores, for each product type, a purchase motivation group that determines the purchase of the product type.
As illustrated in FIG. 17, the emotional personality map 1420 stores a specific purchase motivation in the purchase motivation candidate group in accordance with the personality at the time of product purchase decision. Edge map 1418 and emotional personality map 142
0 is prepared by sales professionals who have deep product knowledge and customer knowledge in contact with many products and many customers,
This is equivalent to an expert D / B who has arranged the relationship between symptoms and causes by an experienced doctor.

The analysis center information processing device 1412 creates a similar product type group list 1424 based on the purchase motive based on the edge map 1418. Also, the edge map 1418 and the emotional personality map 1420
, A product type group list 1426 matching one personality appearing at the time of purchasing decision is created. In addition, based on the past purchase results of the customer, an analysis list 1428 of the purchase motivation on which the customer depends upon purchasing the product is created. Further, from the customer-specific purchase motivation list 1428, a customer list 1430 suitable for a product or a product list 1432 suitable for a customer is obtained.
Create Further, various kinds of statistical data 1434, for example, statistical data in which the total number of customers who match a certain product are obtained are created. These data are provided to the customer information processing device 1408 and the retail store information processing device 140, as required.
2. The information is sent to the product supply source information processing device 1404 and the sales promotion activity source information processing device 1406. In this case, sufficient consideration is given to the protection of personal information, and information to be transmitted and information not to be transmitted upon request are determined in advance by a lawyer. Great care is taken against data tampering, etc., and the latest encryption technology is used for communication between information processing apparatuses, ensuring high information security.

The apparatus and method of the present invention will be described below based on an example in which computer game software (hereinafter referred to as game software) is sold. FIG. 15 is a diagram illustrating an edge map that is an expert D / B. At present, there are numerous types (over 1000 types) of game software ABC ... in the market, and depending on the type, many or few are purchased.

It is extremely difficult to predict whether a product will or will not be accepted in the marketplace, which plagues retailer purchasers. It is also difficult for manufacturers to determine the production quantity. Many market methods have been developed and, for example, data on “customer attribute—a group of product types that are frequently purchased by customers of that attribute” is obtained, so that customers of that attribute are often purchased by customers of that attribute. By promoting product types, it is possible to efficiently promote sales activities,
In addition to low accuracy, it can not be applied to new products, and in fact, the purchase quantity is determined by the intuition of the purchaser,
In many cases, the production quantity is determined based on the intuition and expectations of the manufacturer's sales staff. Therefore, intuition is often lost, resulting in excessive inventory and shortage. Although manufacturers make great efforts to develop products that can be accepted in the market, they often have to develop products without knowing the conditions for acceptance in the market, and their plans are often disappointed. Further, it is difficult for the customer to select a type that is satisfactory, and often a disappointing product is selected.

The present inventors have repeatedly studied a method for making the above opacity understandable, and have finally arrived at the edge theory. The edge theory has been developed by the present inventors, and can be roughly described as follows. (1) The number of factors that the customer weighs in the process of selecting a specific type from similar products, for example, game software, is finite, and a product sales expert can list the factor group.
For example, in the case of game software, purchasing from “highly evaluated in magazines, etc.”, purchasing from “stimulated instinct”, purchasing from “low price”, purchasing from “reliable maker” In some cases, a product sales expert who is in contact with many products and many customers can almost completely list a group of factors that the customer weighs in the process of determining the purchase type. In the case of computer game software, there are 57 factors, and it has been confirmed that almost all customers select a specific type by considering the 57 factors. In the edge theory, a group of factors that a customer considers in the process of selecting a specific type is referred to as a group of purchase motivation candidates. On the vertical axis in FIG. 15B, a motif candidate group list abc that can be a motive for determining purchase of a specific type of product from a certain type of product group is used as a parameter. Also,
On the horizontal axis, a multi-product list ABC is used as a parameter. (2) Each product type (for example, A) has some purchase decision motives in the purchase decision motive candidate group abc... When the purchase decision motive candidate group list is obtained, The expert can specify a group of motivations for deciding the purchase of each product. For example, for the product A, there are a customer who purchases from the purchase motivation a “highly evaluated in a magazine or the like” and a customer who purchases from the purchase motivation g “because it can play against someone”, and the customer who purchases from the purchase motivation c “low price”. Therefore, it is possible to accurately specify that there are almost no customers who purchase or almost no customers who purchase based on the purchase motive d because "the manufacturer is reliable". In FIG. 15 (B), the horizontal axis represents a variety of product lists ABC.
Is used as a parameter, and the vertical axis represents a purchase motivation candidate list abc... As a parameter. Is stored as an example. A thick arrow in the figure indicates that the same type of information is further continuous. (3) Once the edge map illustrated in FIG. 15B is constructed, various information processing is possible. For example, a customer who has purchased a product ABC in the past has a personality that prefers “highly evaluated in magazines or the like” or “can play against someone” products from the edge map, and is satisfied with recommending products having such purchase motivation. While it can be expected that the degree is high,
It can be seen that there is an individuality that does not emphasize "price" or "manufacturer", and that the degree of satisfaction is low even when selecting a product from that viewpoint.

First, a technique for extracting a product type group having a similar purchase motive from the edge map illustrated in FIG. 15B will be described. By picking up the purchase decision motive for each product by the product sales specialist, FIG.
The edge map of (B) is obtained. This map will expand as new products are introduced. Once this edge map is created, various useful data can be obtained. When extracting a product type group having a similar purchase motive, two methods are available. The first is a method of designating a product and searching for a product having a purchase motive similar to the purchase motive of the product. FIG. 15C illustrates the result of the extraction. For example, the product A has the purchase motive of a and g, and the product F
Has two purchasing motivations (100
%), The product DEG has one of the purchase motives (50% coincidence), and it can be seen that the product BC is not completely similar. Similarly, it can be seen that the merchandise C shares three purchasing motivations with the merchandise F having the purchasing motivation abcffgh (the matching degree is 50%), and the other merchandise has a low matching degree. Double purchase circles in the figure indicate common purchase motivations (edges).

For example, if it is found that the product F has the same purchase motive as the product A, various things become possible. For example, recommending a product F to a customer who has a product A but not a product F can be expected to be satisfied. The retail store can determine the purchase amount of the product F with reference to the past sales results of the product A. Similarly, a sales promotion activity plan for the product F can be made with reference to the degree of effectiveness of the sales promotion activity for the product A in the past. The manufacturer can also determine the production amount of the product F with reference to the past sales results of the product A. Furthermore, by intensively analyzing past hit products, it is also possible to find a combination of purchase motivations common to hit products, and the development goal is clear to the manufacturer.

FIG. 16 shows an example of a search result of an edge map using a purchase motivation as a search key. By referring to this list, a combination of purchase motivations commonly owned by the hit product is clarified.

FIG. 17 shows the emotional personality map 30.
5, 1420 are shown. For example, when purchasing game software, there may be a person or a person who makes a purchase decision as a fan of the appearing character, or a person or a person who decides to make a purchase by trying out game skills. The personality α indicates the purchase motive that is weighed when making a purchase decision as a fan. In this case, the purchase motive acghi is emphasized, and the purchase motive b and the like are not weighed. The personality β indicates the personality when the game skill is brought to the forefront and the purchase is determined. In this case, the purchase motivation abdef is emphasized,
This illustrates that the purchase motivation c and the like are not considered.

FIG. 18 shows an example of a product similar to the product F when viewed from the personality γ (evaluating the purchase motivation cfh).
In this case, it can be seen that the product C has 100% coincidence. As shown in FIG. 15 (C), when product-product is compared without assuming a specific personality, product C is 5 compared to product F.
While the degree of coincidence is 0%, when the personality γ comes to the forefront and a product is selected, the product C is 100%
It turns out that they match. On the other hand, although the merchandise A and E have the merchandise-merchandise coincidence of 33% in FIG. 15, there is a possibility that the merchandise is a conforming merchandise. Is 0%, which indicates that the product is not actually a conforming product.

FIG. 19 illustrates an example of purchased merchandise type data for each customer. FIG. 20 is a diagram in which purchase motivations corresponding to the product types included in the purchase product list for each customer are developed from the edge map in FIG. 15B and the purchase product list for each customer in FIG. FIG. 21 is a diagram in which the purchase motivation of FIG. 20 is accumulated for each customer. For example, customer A
It can be seen that the merchandise is selected based on the purchase motivation abfghi and the merchandise is selected without paying attention to the purchase motivation c. The “purchasing motivation group list by customer” in FIG. 21 is useful by itself, but serves as a basis for various applications as described below.

FIG. 22 shows how to find a customer conforming to a new product X (or a stock product, in the case of a stock product, only those customers who have not yet purchased the product are searched for). Show. (A) shows an edge map of the product X. (B) shows a customer-specific purchase motivation group list, that is, a customer-edge correspondence map.
In (B), a double circle is attached to the edge corresponding to the edge map of the product X. While three of the six purchase motivations important to customer A are satisfied with product X and satisfaction is 50%, four of the five purchase motivations important to customer A are: Is satisfied, and the satisfaction is 80%. Obviously, the product X is more suitable for the customer A than for the customers A and C.

FIG. 23 shows a process of extracting a product which is likely to be satisfied from the products B to F and X to Z which have not yet been purchased by the customer C. (A) shows a customer-edge correspondence map of customer c. (B) shows an edge map of an unpurchased product. In (B), the edge corresponding to the customer-edge correspondence map of the customer C is marked with a double circle. Products F and Y
Indicates that up to five of the six purchase motives of the product match the purchase motives of the customer c, and the degree of conformity is 83
%. Obviously, it can be expected that customer C will have a high degree of satisfaction with product F and product Y.

FIG. 24 illustrates how the personality of the customer is classified from the customer-edge correspondence map of FIG. As shown in FIG. 17, for example, the personality α is acghi
It has been found to rely on the five purchase motivations. FIG.
In the case of 4, from the customer-edge correspondence map, customer A selects a product with personality β in front, and customer C has personality α and personality γ.
Appears on the front and is selecting a product. That is, it is understood that the customer C has a dual personality when selecting the product.

FIG. 25 shows a state in which the personality γ comes to the forefront and a product suitable for the customer who selects the product is extracted. It is understood that the customer F having the personality γ matches the products F, Y, and Z. When this is compared with the result of FIG. 23, it is clear that by introducing a personality, the conforming product is widened (in this case, it is newly found that the product Z is a conforming product).
In addition, it can be seen that the degree of conformity is improved, and conformable products can be extracted more reliably. Conversely, it can be seen that the commodity FYZ is well suited to the personality γ. Thus, in order to extract a customer conforming to the product FYZ, it is understood that a customer known to have the personality γ may be extracted as shown in FIG.

FIG. 26 shows an example in which the edge map is quantified. Until now, the decision was made based on whether or not to rely on purchasing motivation, but this time, the degree of dependence is quantified. This figure is also objectively created by a sales specialist.

FIG. 28 shows the quantified edge map of FIG. 26 and the customer-edge correspondence map obtained from the customer-specific purchased commodity list of FIG. See Figure 2 for details on how to find it.
As shown in FIG. This is simply an averaging process. For example, the customer A has two edges of the product A and the edge of a have two edges of a.
If customer B purchases product B
It is calculated to have a weight of 2.5. In FIG. 28, the value is rounded off and converted to an integer.

FIGS. 30 and 31 show how to extract products suitable for customers from the quantified customer-edge correspondence map. Here, a heavy edge is used in addition to the quantified data. The heavy edge is a purchasing motivation that is of particular importance. For example, if the value of the purchase motive a of the customer A is 1, the value of 1 may be calculated because one of the 10 purchased items happens to have 10 evaluation points in the purchase motive a. All of the purchased 10 products may have a rating of 1 for the purchase motive a. In order to make this difference clear, this embodiment uses a heavy edge.
In the case of FIG. 30, customer A indicates that the purchase motivation b is a heavy edge. The product also has a heavy edge, for example, product B
Most of the customers who purchase the product B rely on the purchase motivation b, so that the product b has the heavy edge b.

As shown in FIG. 31, when extracting a product suitable for a customer, first, a product having a matching heavy edge is preferentially found. In this case, the product BDG matches and the product A
Since the CEFs do not match, first, the former is regarded as a product having a high degree of conformity. Next, a product having a high number of coincident edges is a product having a high degree of matching. Customer-edge correspondence map (A) in FIG.
In the edge map (B), zero is a purchase motivation that does not depend on. Therefore, when comparing the customer-edge correspondence map (A) and the edge map (B) in FIG. 30, the number of columns (the number of edges) both having a value other than zero is counted. For example, it can be seen that the product BG shares four edges with the customer A, and the product D shares three edges with the customer A. When the number of edges to be shared is equal, it is determined that a product having a smaller deviation has a higher degree of conformity. Here, the deviation refers to the sum of deviations in the corresponding column of the matching edge other than the heavy edge for the product,
This is illustrated in FIG. Columns with cross hatches in (C) correspond to columns with heavy edges or irregular edges. No deviation is calculated in these columns.

The logic described above, that is, (1) Goods with matching heavy edges have higher fitness than goods without matching. (2) It is assumed that a product having a large number of matching edges has a high degree of conformity between products whose heavy edges match or products having a mismatching edge. (3) It is assumed that commodities having no difference between them still have a high degree of suitability of commodities having a small deviation in FIG. In the case of FIG. 30, it can be seen from the processing of FIG. 31 that the products having a high degree of conformity are GBDFEAC (AC is the same order). It has been confirmed that the products with high relevance extracted in this manner are actually accepted by customers with high satisfaction.

With substantially the same logic, it is possible to extract customers who match the product. In this case, as shown in FIGS. 32 and 33, first, the customer whose heavy edge coincides is preferentially found. In this case, the customer account matches, and the customer Iueo does not match. First, the former is a customer with a high degree of conformity.
Next, a customer having a high number of matching edges is determined to be a customer having a high degree of matching. In the edge map (A) and the customer-edge correspondence map (B) in FIG. 32, the customer account shares four edges with the product B, the customer Eo shares four edges, and the customer IU shares three edges. It can be seen that the edges are shared. If the number of edges to be shared is equal, a customer with a smaller deviation is determined to have a higher degree of conformity. The deviation is illustrated in FIG.

The above logic, that is, (1) A customer with a coincident heavy edge has a higher matching degree than a customer without a coincident edge. (2) It is assumed that a customer with a large number of coincident edges has a high degree of matching between customers whose heavy edges match or customers who do not match each other. (3) It is assumed that the customers who have no difference between them still have a high degree of conformity with the customers having a small deviation in FIG. In the case of FIG. 32, it can be seen from the processing of FIG. 33 that the customers with the highest degree of matching are Akao Eui in order. It has been confirmed that a customer with high relevance extracted in this way accepts the product with high satisfaction.

FIGS. 34 and 35 show a list of many kinds of similar products called game software by genre.
Here, the genres are classified from A to W. A plurality of products or types belong to each category. Hereinafter, a specific product or type is indicated by a genre symbol and a type number.
For example, the first product of the card battle, Digimon World Digital Card Battle, is indicated by A1.

The upper part of FIG. 36 shows a group of similar goods for which similarity was found by the edge theory with respect to the goods shown at the top. The number of matches indicates the number of common edges in the 72 edge groups. For example, D for the product of A1
This indicates that the 22 items share 26 edges (purchase motives). The figure shows the top 20 similar products with the highest number of matches. However, when there are multiple similar products with the same number of matches, all the products of the same rank are displayed,
There are cases where 20 or more similar products are listed. According to the edge theory, for example, the merchandise A1 belonging to the card battle is similar to the merchandise D22 belonging to the RPG genre and the merchandise H1 belonging to the simulation RPG from a purchase motive. The quantity column is
It shows the number of customers who have purchased the listed similar products among the 112 people who have purchased the product A1. For example, it indicates that 32 out of 112 people who purchased the product A1 have purchased the product D22, and 7 people have purchased the product H1. This data is obtained by collecting the purchase results of a group of 18 game software stores and 1059 members for the past 5 years.

The lower part of FIG. 36 shows purchase results by genre. For example, out of 112 people who purchased the product A1, 77 customers who purchased A21 from other products A2 belonging to the same genre are 77 in total. Note that the upper and lower rows have the same number of products.

When comparing the total number of quantities up and down, it is clear that
It is confirmed that the similar products extracted by the edge theory are purchased more, and the purchasing tendency is more similar than the similar products by genre. For example, it is more likely to recommend a similar product discovered by the edge theory shown in the upper row to a purchaser of the product A1 than to recommend a product belonging to the same genre shown in the lower row. Deafness is confirmed. The same can be pointed out to all genres. For example, a customer who has purchased B1 of a baseball game can receive B2 to B22 just because it is the same baseball game.
Is unlikely to be satisfied (only 70 out of 115 people who actually purchased B1 own B2 to B22), whereas similar products shown in the upper row It can be seen that there is a higher probability of being satisfied when recommending similar products is owned by 121 people.

FIG. 37 et seq. Show the result of extracting the personality by the method of FIG. 17, extracting the product adapted to the personality by the method of FIG. 18, and comparing the actual purchase record.
Obviously, it is confirmed that the probability that the product extracted by the method of FIG. 18 is actually purchased is very high. As a result, it can be seen that if an unpurchased product exists in the product group extracted by the method of FIG. 18, when the product is recommended, the product will almost certainly be satisfied.

The above-described embodiment is merely an example of the present invention, and the technical scope of the present invention is not limited to the embodiment. Those skilled in the art can implement the present invention in various forms within the technical spirit described in the claims.

[0056]

According to the present invention, by utilizing the edge theory, it has succeeded in converting past purchase performance data into accurate and useful information based on the individuality of each customer. It is possible to provide useful information to a product supplier, a sales promotion actor, and the like, and it is possible to suppress excessive use and waste that occurs in the distribution stage from product production.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of an apparatus according to claim 1.

FIG. 2 is an explanatory diagram of the method according to claim 2.

FIG. 3 is an explanatory view of the device according to claim 3;

FIG. 4 is an explanatory diagram of the method according to claim 4.

FIG. 5 is an explanatory view of the device according to claim 5;

FIG. 6 is an explanatory diagram of the device according to claim 6.

FIG. 7 is an explanatory diagram of the method according to claim 7.

FIG. 8 is an explanatory view of the device according to claim 8.

FIG. 9 is an explanatory diagram of the method according to claim 9.

FIG. 10 is an explanatory diagram of the device according to claim 10.

FIG. 11 is an explanatory view of the method according to claim 11;

FIG. 12 is an explanatory view of the device according to claim 12.

FIG. 13 is an explanatory view of the method according to claim 13.

FIG. 14 is a diagram showing a system configuration of an apparatus according to the present invention.

FIG. 15 is an explanatory diagram of an edge map.

FIG. 16 is a diagram exemplifying a result of searching an edge map using a purchase motive as a key;

FIG. 17 is a diagram showing an example of an emotional personality map.

FIG. 18 is a diagram showing a product example similar to the product F when viewed from the personality γ.

FIG. 19 is a diagram illustrating an example of purchased merchandise type data for each customer.

FIG. 20 is a diagram in which purchase motives corresponding to product types included in a purchase product list for each customer are developed.

FIG. 21 is a diagram in which the purchase motives shown in FIG. 20 are accumulated for each customer.

FIG. 22 is a diagram showing a state of finding a customer conforming to the new product X.

FIG. 23 is a diagram illustrating a process of extracting a product that will satisfy the customer C.

FIG. 24 is a view for explaining how to classify the personalities of customers from the customer-edge correspondence map of FIG. 22 (B).

FIG. 25 is a diagram showing a state in which a personality γ comes to the forefront and a product suitable for a customer who selects a product is extracted.

FIG. 26 is a diagram showing an example of quantifying an edge map.

FIG. 27 is a diagram showing a list of purchased products by customer.

FIG. 28 is a diagram showing a customer-edge correspondence map obtained from FIGS. 26 and 27;

FIG. 29 is a diagram showing details of a method of obtaining FIG. 28;

FIG. 30: From the quantified customer-edge correspondence map,
The figure which showed a mode that the goods suitable for a customer are extracted.

FIG. 31: From the quantified customer-edge correspondence map,
The figure which showed a mode that the goods suitable for a customer are extracted.

FIG. 32 is a diagram showing a state of extracting a customer conforming to a product.

FIG. 33 is a diagram showing a state of extracting a customer conforming to a product;

FIG. 34: Part 1 of a product list by genre

FIG. 35: Part 2 of a product list by genre

FIG. 36: Upper row: a group of similar products extracted by the edge theory, and the number of purchases of similar products by the purchaser of the specific product. Bottom: Shows the quantity of products that belong to the same genre.

[FIG. 37] A personality is specified from a customer-edge correspondence map.
Part 1 of a table in which it is arranged whether or not a product corresponding to the specified personality is actually purchased.

FIG. 38: Identify personality from customer-edge correspondence map,
Part 2 of a table in which it is arranged whether or not a product corresponding to the specified personality is actually purchased.

FIG. 39: Identify personality from customer-edge correspondence map,
Part 3 of the table in which the merchandise corresponding to the specified personality is actually purchased or not.

FIG. 40: Identify personality from customer-edge correspondence map,
Part 4 of a table in which it is arranged whether or not a product corresponding to the specified personality is actually purchased.

FIG. 41 Identifies a personality from a customer-edge correspondence map,
Part 5 of a table in which it is arranged whether or not a product corresponding to the specified personality is actually purchased.

[Explanation of symbols]

 1402: Retail store information processing device 1404: Product supplier information processing device 1406: Sales promotion source information processing device 1408: Customer information processing device 1410: Communication network 1412: Analysis center information processing device 1414: Customer-purchased product type list 1416: Expert Database 1418: Edge Map 1420: Emotional Personality Map 1424: Similar Product Type Group List 1426: Personality Adaptive Product Type Group List 1428: Purchasing Motivation Analysis List by Customer 1430: Product-Conforming Customer List 1432: Customer-Conforming Product List 1434: Statistical data

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Hisakazu Hirabayashi 3-4 Hayatocho, Showa-ku, Nagoya-shi, Aichi F-term in D-tank Co., Ltd. 5B049 BB11 CC00 EE05 5B075 NS10

Claims (13)

[Claims] 1. One parameter is a list of a variety of similar products, and the other parameter is a list of motivation candidates that can be a motive for determining purchase of a specific type of product from a group of products of the same type. Edge map storage means for storing a group of motivations for deciding the purchase of a product type, and means for retrieving the stored contents of the edge map storage means using a specific type of purchase motive or a specific purchase motive as a key, A device that extracts a group of product types with similar purchase motives from a variety of similar products. 2. One parameter is a list of a variety of similar products, and the other parameter is a list of motivation candidates that can be a motive for determining purchase of a specific type of product from a group of products of the same type. A step of constructing an edge map storing a group of motivations for deciding the purchase of a product type; and a step of searching for the contents of the edge map using the purchase motivation of a specific type of product or the specific purchase motivation as a key. A method of extracting a group of product types with similar purchase motives from a variety of similar products. 3. One parameter is a list of a variety of similar products, and the other parameter is a list of motivation candidates that can be a motivation for deciding to purchase a specific type of product from a group of products of the same type. Edge map storage means for storing a group of motives for determining purchase of a product type; and emotional personality map storage means for storing a specific purchase motive in a group of purchase motive candidates corresponding to a personality at the time of purchase determination. And a means for retrieving the stored contents of the edge map storage means using a purchase motivation corresponding to a specific personality as a key, and extracting a product type group having a purchase motivation adapted to the specific personality from a variety of similar products. . 4. One parameter is a list of similar and various products, and the other parameter is a list of motivation candidate groups that can be a motive for determining purchase of a specific type of product from a group of products of the same type. A step of constructing an edge map storing a group of motivations for deciding the purchase of a product type; and an emotional personality map storing a specific purchasing motivation in the group of candidates for a purchasing motivation according to the personality at the time of the purchasing decision. And a process of searching the contents of the edge map using the purchase motive corresponding to the specific personality as a key to extract a group of product types that have a purchase motive that adapts to the specific personality from a variety of similar products how to. 5. One parameter is a list of similar and diverse products, and the other parameter is a list of motivation candidate groups that can be a motive for determining purchase of a specific type of product from a group of products of that type. An edge map storage unit that stores a group of motivations for deciding the purchase of a product type, a purchase product list storage unit that stores a purchase product list for each customer, and a product type included in the purchase product list for each customer. A device that has a means of accumulating the corresponding purchase motivation and outputs a list of purchase motivation groups by customer. 6. One parameter is a list of a variety of similar products, and the other parameter is a list of a group of motivation candidates that can be a motivation for determining purchase of a specific type of product from a group of products of the same type. An edge map storage unit that stores a group of motivations for deciding the purchase of a product type, a purchase product list storage unit that stores a purchase product list for each customer, and a product type included in the purchase product list for each customer. Means for calculating a customer-edge correspondence map by accumulating the corresponding purchase motives, and a purchase motive group of the unpurchased product type group of the specific customer as a key with the purchase motive group included in the customer-edge correspondence map of the customer as a key A device that has means for searching and extracts product types that share the purchase motivation contained in the customer-edge correspondence map of the customer. 7. One parameter is a list of similar and diverse products, and the other parameter is a list of motivation candidate groups that can be a motivation for deciding to purchase a specific type of product from a group of products of that type. A step of constructing an edge map storing a group of motivations for deciding the purchase of a product type; a step of collecting a list of purchased products for each customer; and a purchasing motive corresponding to a product type included in the purchased product list for each customer. And building a customer-edge correspondence map by accumulating the data, and searching for a purchase motivation group of the unpurchased product type group of the specific customer using the purchase motivation group included in the customer-edge correspondence map of the customer as a key. A method for extracting a product type that shares a purchase motivation included in the customer-edge correspondence map of the customer. 8. One parameter is a list of a variety of similar products, and the other parameter is a list of a group of motivation candidates that can be a motivation for determining purchase of a specific type of product from a group of products of the same type. An edge map storage unit that stores a group of motivations for deciding the purchase of a product type, a purchase product list storage unit that stores a purchase product list for each customer, and a product type included in the purchase product list for each customer. Means for constructing a customer-edge correspondence map by accumulating the corresponding purchase motives; and emotional personality map storage means for storing a specific purchase motive in the purchase motive candidate group corresponding to the personality at the time of the purchase decision. The emotions associated with the purchase motivation group of the unpurchased product type group of the specific customer are associated with the specific personality of the purchase motivation group included in the customer-edge correspondence map of the customer. Having a means for retrieving the purchase incentive group stored in the Le personality map storage means as a key, extracts the product type having compatible purchase motivation to a particular personality of the customer equipment. 9. One parameter is a list of similar and diverse products, and the other parameter is a list of motivation candidates that can be a motive for determining purchase of a specific type of product from a group of products of that type. A step of constructing an edge map storing a group of motivations for deciding the purchase of a product type; a step of collecting a list of purchased products for each customer; and a purchasing motive corresponding to a product type included in the purchased product list for each customer. Accumulating the data and constructing a customer-edge correspondence map; constructing an emotional personality map that stores a specific purchasing motivation in the purchasing motivation candidate group in accordance with the personality at the time of purchasing decision; The emotional personality map is obtained by associating the purchase motivation group of the customer's unpurchased product type group with the specific personality of the purchase motivation group included in the customer's customer-edge correspondence map. Run the step of searching the purchase motivation group stored in the flop as a key, a method of extracting a product type having a matching purchase motivation to a particular personality that customer. 10. One parameter is a list of a variety of similar products, and the other parameter is a list of motivation candidates that can be a motive for determining purchase of a specific type of product from a group of products of the same type. An edge map storage unit that stores a group of motivations for deciding the purchase of a product type, a purchase product list storage unit that stores a purchase product list for each customer, and a product type included in the purchase product list for each customer. Means for calculating a customer-edge correspondence map by accumulating the corresponding purchase motives, and means for searching for purchase motives of a group of unpurchased customers of a specific product type using the purchase motives of the specific product type as a key, A device that extracts customers who meet the purchase motivation for that particular product type. 11. One parameter is a list of a variety of similar products, and the other parameter is a list of motivation candidates that can be a motive for determining purchase of a specific type of product from a group of products of the same type. A step of constructing an edge map storing a group of motivations for deciding the purchase of a product type; a step of collecting a list of purchased products for each customer; and a purchasing motive corresponding to a product type included in the purchased product list for each customer. A process of calculating a customer-edge correspondence map by accumulating the data, and a process of searching for a purchase motivation group of an unpurchased customer group of a specific product type by using the purchase motivation group of the specific product type as a key. How to identify customers who fit the product type motivation. 12. One parameter is a list of a variety of similar products, and the other parameter is a list of motivation candidates that can be a motive for determining purchase of a specific type of product from a group of products of the same type. An edge map storage unit that stores a group of motivations for deciding the purchase of a product type, a purchase product list storage unit that stores a purchase product list for each customer, and a product type included in the purchase product list for each customer. Means for constructing a customer-edge correspondence map by accumulating the corresponding purchase motives; and emotional personality map storage means for storing a specific purchase motive in the purchase motive candidate group corresponding to the personality at the time of the purchase decision. The purchase motivation group of the unpurchased customer group of the specific product type is stored in the emotional personality map storage means in association with the personality corresponding to the specific product type. Having means for retrieving the purchase incentive group as a key, an apparatus for extracting customers with compatible personality purchase motivation of that particular product type. 13. One parameter is a list of similar and diverse products, and the other parameter is a list of motivation candidates that can be a motive for determining purchase of a specific type of product from a group of products of that type. A step of constructing an edge map storing a group of motivations for deciding the purchase of a product type; a step of collecting a list of purchased products for each customer; and a purchasing motive corresponding to a product type included in the purchased product list for each customer. Accumulating the data and constructing a customer-edge correspondence map; and constructing an emotional personality map storing a specific purchasing motivation in the purchasing motivation candidate group in accordance with the personality at the time of purchasing decision. The purchase motivation group stored in the emotional personality map is associated with the purchase motivation group of the unpurchased customer group of the product type and the personality corresponding to the specific product type. Run the step of searching in a method of extracting customers with personality fits the particular product type of purchase motivation.