JP4008711B2 - Customer evaluation information generation method, customer evaluation information generation device, and program used therefor - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention provides, for example, a method and apparatus for generating evaluation information for each customer based on customer information owned by a retail store such as a department store, a card company, and such information. It relates to the program used.
[0002]
[Prior art]
Conventionally, as a method of analyzing customer information performed by retail stores such as department stores and card companies, etc., various questionnaires have been conducted when becoming a customer or a new customer, and from the response results, Deriving hobbies, preferences, and fields of interest for each customer is conducted. Based on the analysis results, the most appropriate information (product information, event information, etc.) is distributed to each customer by e-mail, DM (direct mail), etc. To improve customer satisfaction.
[0003]
Also, basic attributes such as gender, age, occupation, marriage information, etc. are extracted from the customer database, and optimal information is distributed according to the basic attributes, or data mining is performed by referring to such a database. (For example, predicting a purchase pattern from a purchase record or segmenting customers according to the purchase pattern, etc.) to predict future purchases.
[0004]
[Problems to be solved by the invention]
In the former method, if a similar service is to be provided to all customers, it is necessary to conduct a similar questionnaire for all customers, and considering cost effectiveness, it is difficult to always conduct the questionnaire. For customers who did not conduct the survey, the appropriate information cannot be delivered because appropriate analysis is not possible. Also, over time, life patterns and environments change (marriage, childbirth, etc.), and if you continue to deliver information based on the results of past questionnaires, the information will no longer be of interest to that person. There may be a gap between the information and the information that is actually required.
[0005]
Furthermore, since the consumption environment changes or product changes occur with the passage of time, and the purchase pattern and purchase behavior change accordingly, the latter analysis method is based on the assumption of changes. Is difficult, and there is a gap between the prediction based on the analysis and the actual purchase pattern or behavior.
[0006]
The present invention has been made paying attention to the above-mentioned problems, and in the case where a large number of customer information is possessed by a retail store, a card company, etc., evaluation of each customer is performed in a permanent, efficient and accurate manner. It is an object of the present invention to provide a method for obtaining information, an apparatus for generating such evaluation information, and a program used therefor.
[0007]
[Means for Solving the Problems]
In order to solve the above-mentioned problem, the customer evaluation information generating device according to the present invention is provided for each evaluation item specified in advance. plural Based on the results of questionnaires conducted for some customers among many customers, Each of a plurality of related items is associated with each combination of the questionnaire results. Item value is assigned, and this is sampled customer evaluation data recorded as evaluation information of the part of customers. ; Customer behavior history data recorded for each of the customers; ; For customers who conducted the questionnaire, plural For each related item Assigned to Corresponding to the item value , Relevant from the customer behavior history data Action history target items , Select and narrow down one or more action history items included in the action history data, calculate the appearance ratio of the selected action history target items for some customers having the item value, and calculate the calculated appearance ratio The presence / absence of the action history for the action history target items narrowed down by associating the highest item with the item value of the related item is stored, and by performing these operations for all the related items, the some customers For each item value of related items Analogical pattern generating means for generating an analogical pattern, analogical pattern recording means for recording the generated analogical pattern, and action history target items corresponding to item values of the analogical pattern recorded in the analogical pattern recording means Presence or absence of action history , Behavior history information of each customer who did not conduct the questionnaire Action history and If each customer who collated and did not conduct the questionnaire had a similar behavior history, it was obtained by the analogy pattern Assigned to multiple related items About each customer who did not carry out questionnaire by giving item value , Similar to the evaluation information of some of the customers An analogy pattern application unit for obtaining evaluation information.
[0010]
Above According to the invention, by using customer evaluation information and behavior history information obtained by a questionnaire or the like, the analogy pattern of the questionnaire implementation customer is generated in advance, while for a large number of customers having behavior history information, Since customer evaluation is performed by applying this analogy pattern, it is possible to obtain evaluation information of each customer efficiently and accurately when there is a large amount of customer information, and provide appropriate information desired by the customer. can do.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
[0012]
FIG. 1 is a schematic block diagram showing an apparatus for generating customer evaluation information according to the present invention (hereinafter referred to as “customer evaluation information generating apparatus”).
[0013]
The customer evaluation information generating apparatus 100 is an information processing apparatus such as a computer or a workstation, and includes an arithmetic processing unit (CPU), a main memory (RAM), a read only memory (ROM), and an input / output device (I / O). It is composed of a device equipped with a storage device such as a hard disk device. The ROM, the storage device, or the like stores a program for causing the information processing apparatus to function as the customer evaluation information generation apparatus 100 or a program for causing a computer to execute a customer evaluation data generation method. Is placed on the main memory, and the CPU executes this to implement a customer evaluation information generation device or to execute a customer evaluation information generation method. In addition, the program need not be stored in a storage device in the apparatus, and can be read by a computer such as a magnetic disk, an optical disk, a magneto-optical disk, a CD (Compact Disc), and a DVD (Digital Video Disc). A possible program recording medium or an external device (for example, an ASP (application service provider's server, etc.)) may be provided and loaded in the main memory.
[0014]
This apparatus includes an analogy pattern generation unit 101, an analogy pattern recording unit 102, and an analogy pattern application unit 103. The analogy pattern generation unit 101, the analogy pattern recording unit 102, and the analogy pattern application unit 103 are respectively cooperated by a CPU, a main memory, an external storage device, and a program module that is installed in the main memory and causes the CPU to perform a predetermined operation. The means to be realized does not necessarily exist physically, and does not necessarily need to be realized by the physically same information processing apparatus. For example, it is implemented in a plurality of computers connected by a network. It may be a distributed object.
[0015]
The analogy pattern generation means 101 has customer evaluation data (hereinafter referred to as “sampling customer evaluation data”) for each customer (hereinafter referred to as “sampling customer”) extracted from a plurality of customers as a population by a method such as random sampling. 105 and a customer behavior history data relating to each of the sampling customers, and has a function of generating analogy pattern data. In this embodiment, payment history data, specifically, data relating to the purchase history of the customer (hereinafter referred to as “sampled customer purchase history data”) 106 is applied as the customer behavior history data.
[0016]
The analogy pattern recording unit 102 has a function of receiving and recording analogy pattern data from the analogy pattern generation unit 101.
[0017]
The analogy pattern application unit 103 includes analogy pattern data recorded in the analogy pattern recording unit 102 and data related to purchase history regarding a customer other than a sampling customer (hereinafter referred to as “all customers”) among a plurality of customers as a population. (“All customer purchase history data”) 107 is used to generate customer evaluation data (referred to as “all customer evaluation data”) 108 relating to all customers.
[0018]
The analogy pattern application unit 103 preferably performs the above operation in real time (for example, every month). As a result, even if the lifestyle of a person changes, the purchase history also inevitably changes accordingly, so that the customer evaluation data can be changed in the analogy pattern application unit 103 accordingly, which will be described later. The provided information can always be matched to that which matches the person.
[0019]
The questionnaire response data 104, the sampling customer evaluation data 105, the sampling customer purchase history data 106, the all-customer purchase history data 107, and the all-customer evaluation data 108 are managed by a server.
[0020]
Next, the operation procedure of this embodiment in the above-described apparatus configuration will be described. Here, a procedure for generating customer evaluation data will be described using data obtained with respect to a credit card issued by a department store having various stores as an example. In this case, the customer (for example, suppose that there are 10,000 people) purchase history data (what time, at what department, what price bought, etc.) and basic attribute data (gender, age, occupation) , Address etc.).
[0021]
In the present embodiment, since the purchase history data is used as the customer behavior history data on the sampling side as described above, the all customer purchase history data 107 is also used for all other customers accordingly. .
[0022]
[Stage 1; Sampling customer evaluation data acquisition]
A questionnaire for measuring customer evaluation data is performed on some customers (for example, 500 people) extracted by random sampling from a total of 10,000 customers. The items of the questionnaire are as illustrated in FIG. 2, and are composed of question items that can analyze the future consumption trend and consumption behavior of each customer.
[0023]
The questionnaire as shown in FIG. 2 is preferably performed at a predetermined interval (for example, every year). As a result, even if the consumption environment changes over time or product changes occur, customer evaluation data corresponding to the change can be created, and changes in purchase patterns and purchase behavior can be accommodated.
[0024]
Next, a configuration example of the sampling customer evaluation data will be described with reference to FIGS.
[0025]
As shown in FIG. 3, the sampling customer evaluation data 200 has one record R201, R202,... For each customer, and one record stores evaluation information of the corresponding customer. Specifically, as will be described later, the customer evaluation data 200 is expressed by a table in which code values corresponding to combinations of examination item results are associated with each sampled customer for each sampled item. Is done.
[0026]
The evaluation information has at least one evaluation item. Here, the “evaluation item” means an item that can be predicted and analyzed in a multifaceted manner with respect to the customer's future consumption trend, consumption behavior, and the like. In this embodiment, as shown in FIG. In addition, the evaluation information includes three evaluation items 201α, 201β, and 201γ. Specifically, as shown in the figure, as evaluation items, for example, (1) life values (201α), (2) information attitude (201β), (3) consumption tendency (201γ), etc. Are listed.
[0027]
The evaluation items 201α, 201β, and 201γ are classified into a plurality of related items α1 to α8, β1 to β4, and γ1 to γ4, respectively. Here, the “related items” are classified by items that make it easier to predict and analyze the evaluation items. In the example shown in FIG. 3, the evaluation items 201α are eight related items, the evaluation items 201β. 201 γ are classified into four related items. Further, in each related item, a plurality of “review items” that are related to each other and more specific are defined.
[0028]
The “living values” in the evaluation item 201α described above is a measure for judging whether or not it is necessary (provided) for events related to life, and is classified here. Related items α1 to α8 include, for example, α1. Items of interest (examination items are, for example, a. Are you interested in romance? B. Are you interested in sports? C. Are you interested in gambling? D. , Etc.?), Α2. Priority items (examination items are, for example, a. Sensitivity, b. Status, c. Good product, etc.), α3. Purpose-of-use items (examined items are, for example, a. A topic store, b. A place suitable for a date, c. A place where a family can go ...) ...
[0029]
In addition, the “information attitude” in the evaluation item 201β described above refers to the side receiving or giving various information, or the frequency of contact with various media (reading leaflets, watching DMs, being easily influenced by family opinions, etc.) , Which is a measure for judging sensitivity (details about fashion, new information is always available, etc.), etc., and related items β1 to β4 classified here are, for example, β1. Information acquisition items (examination items are, for example, a. Magazine, b. DM, c. Television, radio, d. Internet, etc.), β2. ...
[0030]
Further, the “consumption tendency” in the evaluation item 201γ described above is a measure for judging characteristics, trends, and the like related to consumption, and the related items γ1 to γ4 classified here are, for example, γ1. Purchase items (consideration items are, for example, a. Influenced by fashion, b. Solid ...), γ2. ...
[0031]
Of course, these evaluation items 201 are only classified horizontally and hierarchically so that more multifaceted and accurate analysis can be performed, and are classified according to the contents, the number of items, and each evaluation item. A plurality of related items and examination items in each related item are arbitrary and can be appropriately modified.
[0032]
Then, each of the related items α1 to α8, β1 to β4, and γ1 to γ4 is assigned a corresponding item value (code value) according to the result of the examination item in the related item. In the figure, such code values are shown in capital letters.
[0033]
Here, the procedure from the result of the questionnaire shown in FIG. 2 to the provision of the code value shown in FIG. 3 will be described with reference to the related item α1 of the life values 201α.
[0034]
As shown in FIG. 4, the examination items in the related item α1 of the life values 201α include a. Are you interested in love or not? B. Are you interested in sports? C. Are you interested in gambling? D. Are you interested in fashion? It is. Therefore, among the questions of the questionnaire shown in FIG. ~ D. Are extracted in advance, and each item to be examined a. ~ D. Make an association with.
[0035]
And each examination item is examined from the questionnaire result of the customer. For example, Question 1, Question 4,... Question n of the questionnaire shown in FIG. Are related to each of the combinations of the results of these questions 1, 4,..., N. Decide whether is ○ or ×. Thereby, the examination item a. It can be determined whether or not. In this way, other examination items b. ~ D. Also decide whether it is ○ or ×.
[0036]
All considerations a. ~ D. When it is determined whether or not it is o or x, as shown in FIG. 4, a code value corresponding to the related item α1 is given by the combination. Similarly, code values are determined for all related items in the life values 201α. Such an operation is also performed for the information sensitivity 201β and the consumption tendency 201γ..., Thereby obtaining the sampling customer evaluation data shown in FIG.
[0037]
In addition, since the question content of the questionnaire is related to the content of the examination item, it is desirable to change it appropriately according to the examination item when acquiring the customer evaluation data. Moreover, there is no restriction | limiting in particular about which question content of which questionnaire is linked | related with which examination item, It can determine suitably so that a customer analysis as detailed as possible can be performed. Of course, questionnaire questions may be determined so that the response result can be directly obtained for the examination item.
[0038]
[Second stage: Acquisition of sampling customer purchase history data]
FIG. 5 shows a data configuration example of the purchase history data 106 of the sampling customer. The sampled customer purchase history data 106 has one record R301, R302,... For each customer, and each record is a purchase target item 3011, 3012,..., 3013, 3014 indicating an object purchased by the customer in a certain period T. , ..., 3015, 3016, ..., 3017.
[0039]
In the example shown in FIG. 5, the record R301 is a customer having a customer ID “00001” and uses a sports shop (3011) and casual clothing (3013) in a period T, but a ticket center (3012), This shows that formal clothing (3014), French cuisine (3015), Italian cuisine (3016), and alcoholic beverages (3017) were not used. Thereby, it is possible to recognize the purchasing tendency of the sampled customers.
[0040]
[Third stage; generation of analogy pattern]
Next, using the sampling customer evaluation data and the sampling customer purchase history data described above, the purchase pattern of the sampling customer is inferred. Here, the operation of the analogy pattern generation unit 101 will be described with reference to FIGS. FIG. 6 is a flowchart showing an analogy pattern generation method executed by the analogy pattern generation unit 101.
[0041]
An analogy pattern is generated for each related item. For example, as shown in FIG. 3, the evaluation item “living values 201α” has eight related items α1 to α8, and each related item is given an item value (code value) by the above-described procedure. Yes. In this case, eight analogy patterns are generated in the life values 201α, and four analogy patterns are generated for the other evaluation items “life values 201β” and “201γ”.
[0042]
First, the analogy pattern generation unit 101 identifies related items that are the target of the analogy pattern (step S601). For example, the related item α1 in the first evaluation item group “life values” 201α is specified.
[0043]
Next, the analogy pattern generation unit 101 extracts a record having the same code value for the identified related item from the sampling customer evaluation data 105, and stores a corresponding customer identifier (step S602). Thereby, it is possible to list customers who have the same customer evaluation for the related item α1.
[0044]
Next, the analogy pattern generation unit 101 acquires a purchase history record having the customer identifier stored in step S602 from the sampling customer purchase history data 106 (step S603). Thereby, the purchase history of the customer who is the same customer evaluation about the related item α1 can be obtained.
[0045]
Next, the analogy pattern generation unit 101 narrows down purchase target items (step S604). The narrowing down of the purchase target items is performed by selecting one or a plurality of purchase target items from all purchase history items included in the purchase history data. The selected purchase target item can be determined independently for each evaluation item, each related item, or each code value.
[0046]
For example, in the evaluation item 201α, when the purchase target items are narrowed down by the code value I of the related item α1, only “d. I am interested in fashion” is ○ as shown in FIG. As described above, the items to be purchased are narrowed down to only the casual clothing (3013) and the formal clothing (3014).
[0047]
In addition, in the evaluation item 201α, when the purchase target items are narrowed down by the code value J of the related item α1, “a. I am interested in love” and “d. I am interested in fashion” are ○. As items to be purchased, items to be purchased are narrowed down to four items: a ticket center (3012), casual clothing (3013), French cuisine (3015), and Italian cuisine (3016).
[0048]
Such narrowing is performed by predicting stores and events that will be related to or will be used according to the examination items, and such narrowing is performed in addition to, for example, statistical probability. It can be determined appropriately such as inferring based on this. In some cases, it is not necessary to narrow down by referring to all purchase histories.
[0049]
As means for narrowing down as described above, the analogy pattern generation means 101 has a narrowing table in which purchase target items and related items (review items to which codes are assigned) are stored in association with each other. Although a method of narrowing down is conceivable, the purchase target items may be narrowed down by other methods. Further, for each related item of each evaluation item shown in FIG. 4, it is necessary to determine in advance which purchase target item to narrow down, but the association between each related item and the purchase target item can be changed as appropriate. .
[0050]
Next, the analogy pattern generation unit 101 calculates an appearance ratio for each purchase history content from the record acquired in step S603 (step S605). FIG. 7 is a diagram for explaining an example of an appearance ratio calculation method for each purchase history content. Here, the “purchase history content” refers to the result (whether or not the customer uses) the purchase history item selected by the narrowing-down in step S604.
[0051]
As described above, since the code value J extracted in S602 has a high interest in fashion and love as shown in FIG. 4, the purchase history has been narrowed down to clothing. Only two items of “casual clothing” and “formal clothing” are selected as purchase history items.
[0052]
In calculating the appearance ratio for each purchase history content, all purchase histories may be referred to without narrowing down purchase history items as described above. In this way, the appearance ratio for each purchase history content can be calculated more accurately.
[0053]
Of the records acquired in step S603, P1 indicates the number of records indicating that both casual clothing and formal clothing are purchased, P2 indicates the number of records indicating that only casual clothing is purchased, and purchases only formal clothing. Let P3 be the number of records indicating that the user is present and P4 be the number of records indicating that neither casual clothing nor formal clothing has been purchased.
[0054]
The analogy pattern generation unit 101 calculates the percentages of P1 to P4 (these calculation results are referred to as appearance ratios). The example shown in FIG. 7 shows a state where P1 = 20% (region R701), P2 = 70% (region R702), P3 = 7% (region R703), and P4 = 3% (region R704).
[0055]
Next, the analogy pattern generation unit 101 determines the item having the highest appearance ratio from the calculation result of the appearance ratio (step S606), and stores the purchase target item corresponding to the determined appearance ratio and the content thereof (step S606). S607). In the example shown in FIG. 7, the highest appearance ratio is “70%” in the region R702, and the item to be purchased and the content corresponding thereto are “casual clothing ○, no formal clothing”.
[0056]
Next, the analogy pattern generation unit 101 writes and stores the purchase content item stored in step S607 in the analogy pattern recording unit 102 as an analogy pattern of the code value. In the example illustrated in FIG. 7, the purchase content item “casual clothing ○, no formal clothing” is stored as an analogy pattern corresponding to the code value “J” of the related item α1 of the evaluation item group “life values” 201α.
[0057]
Then, it is determined whether or not the processing from step S601 to step S607 has been performed for all code values (step S608), and if processing has been performed for all code values of the related item, the process proceeds to the next related item. (Steps S610, S611) (This is sequentially performed as α2, α3,...) On the other hand, if not completed, the process proceeds to the next code value (Step S609), and all code values are processed. To do.
[0058]
In this way, analogy patterns are generated and stored for all related items and code values included in each evaluation item.
[0059]
In addition, in the generation of analogy patterns, for sampling customers, clustering from the customer evaluation data side and clustering from the purchase history data side, create a matrix between these two clusters, and calculate the appearance ratio of each By doing so, an analogy pattern may be generated.
[0060]
FIG. 8 shows an example of analogy pattern data recorded in the analogy pattern recording unit 102 when the analogy pattern generation unit 101 generates analogy patterns for all evaluation items that can be included in the customer evaluation data. . Here, “◯” is given to items with a purchase history, blanks are given to items with no purchase history, and “−” is assigned to items that have not been narrowed down.
[0061]
[Fourth stage: Application of analogy pattern]
Next, the customer evaluation information generation device 100 extracts the above-mentioned analogy pattern data for each of the customers who did not answer the questionnaire (customers other than the sampling customer), and generated the purchase history data in the third stage. Apply analogy pattern data.
[0062]
Here, the purchase history narrowed down as described above and the corresponding purchase history are extracted, and the same code value is assigned when they match.
[0063]
FIG. 9 is a flowchart illustrating an example of a method for applying analogy pattern data executed by the analogy pattern application unit 103.
[0064]
First, the analogy pattern application unit 103 sets a pointer to the record of the code value of the first related item of the analogy pattern data recorded in the analogy pattern recording unit 102 (step S901).
[0065]
The analogy pattern application unit 103 stores the item to be purchased and the contents of the record set with the pointer (step S902). For example, in the analogy pattern data shown in FIG. 8, when the pointer is placed in a record indicating the code value “J” of the related item α1 of the “life values 201α” of the evaluation item group, the analogy pattern application unit 103 The purchase target item “casual clothing” is “○” and “formal clothing” and “none” are stored.
[0066]
Next, the analogy pattern application unit 103 sets a pointer to the first record of all customer purchase history data 107 (step S903). Further, the analogy pattern application unit 103 determines whether or not the item to be purchased and the content stored in step S902 match the content of the record of all customer purchase history data 107 to which the pointer is attached in step S903 (step S903). S904). For example, it is determined whether the purchase target item “casual clothing” in the record of the all customer purchase history data 107 is “◯” and “formal clothing” is “none”.
[0067]
If the contents of both records match, the code value is written in the related item of the customer evaluation data of the customer corresponding to the record (step S905). For example, the code value “J” is written in the related item α1 of the evaluation item group “life values 201α” of the customer. If they do not match, nothing is written. This is performed for all records of customer purchase history data, that is, all customers (steps S906 and S907). When the analogy pattern application process is completed for all customer records, the process proceeds to the next code value process (step S908). For example, when the related item α1 and the code value “J” of the evaluation item group “life values 201α” of the customer are finished, the process shifts to the related item α1 and the code value “K” to reach the code value “P”. Similarly, the analogy pattern application process continues.
[0068]
Next, it is determined whether or not processing has been completed for all code values of all related items (step S909). If completed, generation of customer evaluation data is terminated. If there is a related item or code value that has not yet been processed, steps S902 to S908 are further repeated.
[0069]
The customer evaluation data completed as described above is output as all customer evaluation data 108 by the analogy pattern application unit 103. In this case, the total customer evaluation data is configured similarly to the data shown in FIG.
[0070]
Thus, customer evaluation data is created for all customers based on the questionnaire results for some customers. Note that, as described above, in the method of narrowing down and calculating the appearance ratio and evaluating all customers according to the one with the highest appearance ratio, a code value is not assigned to a specific item itself. In this case, a process of assigning code values to all items may be performed by further reducing the appearance ratio by one rank. Also, depending on the narrowing down method, a plurality of code values may be assigned. In this case, a plurality of code values may be assigned as they are.
[0071]
[Fifth stage: Transmission and distribution of information according to customers]
The customer evaluation data created as described above is used to transmit and distribute information according to the customer. FIG. 10 is a block diagram schematically illustrating a configuration example of a customer-specific information distribution device for transmitting and distributing information according to a customer using customer evaluation data.
[0072]
The apparatus 1000 evaluates the customer evaluation data storage unit 1001 that stores the customer evaluation data generated by the customer data generation apparatus 100 described above and stores / events in advance, and evaluation information similar to the customer evaluation data Store / event evaluation data storage unit 1002 for storing evaluation data having the above, matching means 1003 for comparing the customer evaluation data with the store / event evaluation data and selecting a customer that matches the store / event, and each customer A destination data storage unit 1004 that stores information transmission destinations (e-mail address, postal address, etc.), a transmission / send content data storage unit 1005 that stores information to be sent to customers such as store / event contents, Send matching store / event information to the customer selected by the matching means 1003 And a transmission-sending processing means 1006 subjecting.
[0073]
As for the store / event evaluation data, for example, data may be obtained from each store or event company and the evaluation data may be prepared in advance. As shown in FIG. 11, the customer evaluation data shown in FIG. It is composed of similar items.
[0074]
The customer-specific information distribution apparatus 1000 stores all customer evaluation data generated by the customer evaluation information generation apparatus 100 in the customer evaluation data storage unit 1001. The matching unit 1003 extracts the customer evaluation data from the customer evaluation data storage unit 1001 and also extracts the store / event evaluation data from the store / event evaluation data storage unit 1002. To match. Then, this store / event information is provided to matching customers by various providing means such as e-mail, DM, and telephone.
[0075]
Specifically, a database that stores customer evaluation data generated as described above is constructed for all customers, and a database that stores information such as stores and events is constructed on the other hand. . Then, for each customer, the optimum information is extracted by the matching means 1003 and provided to each customer via transmission and sending means (e-mail, DM, TEL, etc.).
[0076]
As an actual information provision procedure, the store / event information stored in the transmission / delivery content data storage unit 1005 is output to the transmission / delivery processing unit 1006, and the destination of the customer providing the information is designated as the destination data storage unit 1004. To the transmission / delivery processing means 1006, and the transmission / delivery processing means 1006 transmits / sends the store / event information to the corresponding customer. That is, based on the result of matching between customer evaluation information and store / event evaluation information acquired in advance, store / event information suitable for the customer is accurately distributed to the customer.
[0077]
In providing information to each customer, it is preferable to refer to data obtained by the “information attitude” of 201β in the evaluation items described above. That is, in this “information attitude”, since the most preferable method for obtaining information is analyzed for each customer, an optimal information transmission means may be selected and used for each customer. For example, an apparatus (for example, a printer for printing address labels, a telephone operator, etc.) for implementing a plurality of types of information transmission means (for example, guidance by DM (direct mail), guidance by telephone operator, guidance by electronic mail, etc.) A terminal device for displaying a customer list including telephone numbers, a computer for sending e-mail, etc.) to a selectable customer-specific information distribution device, and this information according to the data on the customer's information attitude The information selection means most suitable for the customer is selected from the transmission means, and the guidance information is provided. In this way, customer satisfaction can be further improved.
[0078]
As described above, in the present invention, an analogy pattern is generated using sampling customer evaluation data and purchase history data obtained by a questionnaire conducted for some customers, and the generated analogy pattern is converted to a purchase history. Since all customer evaluations are performed by applying to a large number of customers who have data, evaluation information of each customer can be obtained efficiently and accurately when there is a large amount of customer information. It is possible to provide different appropriate information.
[0079]
As mentioned above, although embodiment of this invention was described, this invention is not limited to embodiment mentioned above, It is possible to implement in various changes. For example, in the above-described embodiment, a case is described in which purchase history data (payment history data) is used when generating an analogy pattern or applying the generated analogy pattern to another customer. In addition to such payment history data, for example, payment history, borrowing history, and the like can be referred to. In addition, although purchase history data is used as the customer behavior history, in the present invention, as such customer behavior history, other than that, WEB history (access history) by the Internet, point product exchange history, by telephone Response history (voice response history), campaign participation history, service usage history (various insurance systems, member discount system, loan borrowing system, etc.), store visit history, etc. Can also be used in combination.
[0080]
Moreover, although the example in a department store is demonstrated in embodiment mentioned above, this invention is used in field | areas, such as a store represented by various retail stores and restaurants, highways, finance, and Internet shopping. It can be applied to various credit cards.
[0081]
In addition, by applying the basic attribute data (gender, age, occupation, address, etc.) and the range of action of each customer (card usage area, date of use, etc.) to the embodiment described above, a certain customer Therefore, it is possible to provide information with higher accuracy, such as reducing information that is not essentially required.
[0082]
Furthermore, in the present invention, the generation of customer evaluation information and the distribution of the generated customer evaluation information may be configured by software (program).
[0083]
[Other variations]
(1) In the above embodiment, the sampled customer evaluation data has been described as being created based on the answer of a questionnaire conducted with respect to the sampled customer. You may make it produce sampling customer evaluation data using. For example, instead of a questionnaire response, the result of a proper diagnostic test, the result of a personality diagnostic test, etc. may be used, or any other information that reflects attributes such as taste, personality, and propensity of other customers is used. Also good.
[0084]
(2) In the analogy pattern in the above embodiment, the information stored in each purchase history item is the content of whether or not the product corresponding to the purchase history item has been purchased. Is not limited to such information. For example, the information stored in each purchase history item may be information indicating whether the accumulated amount of purchase is within a certain threshold or exceeding it, or may be the number of purchases. Alternatively, information reflecting the degree of interest of other customers may be used.
[0085]
(3) In this embodiment, the matching system for matching a store / event to a customer based on the customer evaluation information has been described. However, the target that can be matched with the customer based on the customer evaluation information is the store / event. It is not limited to events. By using the customer evaluation information according to the present invention, it is possible to match a customer with all objects such as all products / services, people, things, companies, and organizations.
[0086]
(4) In the above-described embodiment, as shown in FIG. 4, for the related items of the evaluation items, a code value is given by a combination of ○ and × of the examination items, and the sampling customer evaluation data shown in FIG. 3 is created. However, the configuration of the customer evaluation data (including both sampling customer evaluation data and all customer evaluation data) can be variously modified.
[0087]
For example, FIG. 12 shows a configuration example of customer evaluation data applicable to the present invention. In this example, it is assumed that the evaluation items included in the customer evaluation data are composed of a plurality of examination items, and each of the examination items can take a plurality of values representing the strength, matching degree, etc. of a certain characteristic as a code value. In FIG. 12, five characteristics of “not interested”, “not so interested”, “normal”, “somewhat interested”, and “interested” are set for the examination item “Is interested in love?” The code value corresponding to the characteristics of the customer is stored in the examination item. For other examination items, code values representing the strength of the corresponding characteristic, the degree of coincidence, and the like are stored.
[0088]
In addition, when conducting a questionnaire survey with sampling customers, each review item is set to a survey item that can be evaluated in multiple stages, and each review item can be analyzed based on strength. You can leave it.
[0089]
FIG. 13 shows a configuration example of customer evaluation data when the code values shown in FIG. 12 are used. As shown in the figure, each customer record is composed of a set of code values representing the strength of the customer's characteristics, the degree of match, and the like.
[0090]
FIG. 14 shows a configuration example of analogy pattern data when the code values shown in FIG. 12 are used. This configuration example is basically the same as the configuration example of the analogy pattern data shown in FIG. 8. For each code value of each examination item, there is a purchase history corresponding to each purchase target item. Items that are not present are blank, and items that are not narrowed are given "-".
[0091]
FIG. 15 shows a configuration example of the store / event evaluation data when the code values shown in FIG. 12 are used. This store / event evaluation data is a set of code values representing the strength of a characteristic (examination item), the degree of matching, etc., corresponding to the customer evaluation data shown in FIG.
[0092]
The configurations of the customer purchase history data 106 and 107, the configuration / operation of the customer evaluation information generation device 100, and the configuration / operation of the customer-specific information distribution device 1000 can be the same as those described in the above embodiment. That is, the customer evaluation information generation device 100 has the configuration shown in FIG. 1, and this customer evaluation information generation device 100 has the customer purchase history data 106 having the data configuration shown in FIG. 5 and the data shown in FIG. Using the sampled customer evaluation data having the configuration, the analogy pattern generation operation shown in FIG. 6 is performed to generate analogy pattern data as shown in FIG. 14, and this analogy pattern data is compared with all customer purchase history data 107. Thus, all customer evaluation data 108 having a configuration as shown in FIG. 13 can be generated.
[0093]
Further, the customer-specific information distribution apparatus 1000 holds all customer evaluation data 108 having the configuration shown in FIG. 13 in the customer evaluation data storage unit 1001 and stores / event evaluation having the data configuration shown in FIG. Data can be stored in the store / event evaluation data storage unit 1002 and these data can be compared by the matching means 1003 to select a store / event suitable for each customer.
[0094]
According to the configuration using the code value representing the strength of the customer's characteristic, the degree of coincidence, etc., it becomes possible to perform analysis and matching based on the strength of each examination item.
[0095]
【The invention's effect】
As described above, according to the present invention, an analogy pattern for a customer is generated in advance using customer evaluation information and purchase history information obtained by a questionnaire or the like, and for a large number of customers having purchase history data, Since customer evaluation is performed by applying this analogy pattern, it is possible to obtain evaluation information of each customer efficiently and accurately when there is a large amount of customer information, and provide appropriate information desired by the customer. can do.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram showing a configuration of a customer evaluation data generation apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram showing items of a questionnaire conducted for some customers in order to generate sampling customer evaluation data.
FIG. 3 is a diagram showing sampling customer evaluation data.
FIG. 4 is a diagram for explaining a configuration example of evaluation items for obtaining sampling customer evaluation data;
FIG. 5 is a view showing purchase history data of a sampling customer.
FIG. 6 is a flowchart showing an analogy pattern generation method executed by the analogy pattern generation unit shown in FIG. 1;
FIG. 7 is a diagram for explaining an example of a method for calculating an appearance ratio for each purchase history content;
FIG. 8 is a view showing an example of analogy pattern data recorded in the analogy pattern recording unit shown in FIG. 1;
FIG. 9 is a flowchart showing an example of a method for applying analogy pattern data executed by an analogy pattern application unit.
FIG. 10 is a block diagram illustrating a configuration example of a customer-specific information distribution device for performing transmission / distribution of information according to a customer using customer evaluation data.
FIG. 11 is a diagram showing an example of store / event evaluation data.
FIG. 12 is a view for explaining a modified example of evaluation items for obtaining sampling customer evaluation data.
FIG. 13 is a diagram showing another configuration example of customer evaluation data.
FIG. 14 is a diagram showing another configuration example of analogy pattern data.
FIG. 15 is a diagram showing another configuration example of store / event evaluation data;
[Explanation of symbols]
100 ... Customer evaluation information generating device
101 ... Analogical pattern generation means
102. Analogical pattern recording means
103 ... analogy pattern application means
104 ... Questionnaire response data
105 ... Sampling customer evaluation data
106 ... Sampling customer purchase history data
107 ... All customer purchase history data
108 ... All customer evaluation data
1000 ... Customer-specific information distribution device
1001 ... Customer evaluation data storage unit
1002 ... Store / event evaluation data storage unit
1003 ... matching means,
1004 ... Destination data storage unit
1005 ... Transmission / send contents data storage unit
1006 ... Transmission / delivery processing means

Claims (3)

A plurality of related items are set for each evaluation item specified in advance, and for each of some of the customers based on the results of a questionnaire conducted on some of the customers. Sampling customer evaluation data in which item values associated with each combination of questionnaire results are assigned to each of the plurality of related items , and this is recorded as evaluation information of the part of the customers ; and customer action history data recorded for each part of the customer; using, for customers who carried out the questionnaire, corresponding to the item value assigned to each of the plurality of related items, the customer behavior history data the action history object items that are related from with to filter one or more selected from all the action history items included in the action history data, this for some customers with the item value Calculate the appearance ratio for the action history target item that has been included, and store the presence / absence of the action history for the action history target item that has been narrowed down by associating the highest appearance ratio with the item value of the related item Then, by performing these operations for all related items, analogy pattern generation means for generating an analogy pattern for each item value of the related items for the part of customers ,
Analogy pattern recording means for recording the generated analogy pattern;
The presence / absence of the action history in the action history target item corresponding to the item value of the analogy pattern recorded in the analogy pattern recording means is collated with the action history of the action history information of each customer who did not conduct the questionnaire, and the questionnaire If each customer who did not conduct the survey has a similar behavior history, each customer who did not conduct the questionnaire by assigning item values assigned to each of a plurality of related items obtained by the analogy pattern For analogy pattern application means for obtaining evaluation information similar to the evaluation information of some of the customers ,
A customer evaluation information generating apparatus comprising: The action history information includes payment history data, Internet web history (access history), point product exchange history, telephone response history (voice response history), campaign participation history, service usage history, and store visit history. The customer evaluation information generating device according to claim 1 , wherein the customer evaluation information generating device is obtained from at least one of them. A customer evaluation data storage unit for recording customer evaluation data obtained from the customer evaluation information generating device according to claim 1 ;
For each of a large number of stores / events, a related item is set for each evaluation item specified in advance, and a predetermined item value is assigned to the related item for each of the large number of stores / events. A store / event evaluation data storage unit to record as event evaluation data;
A matching means for comparing the customer evaluation data with the store / event evaluation data and selecting a customer that matches the store / event,
Distribution means for distributing store / event information suitable for the selected customer to the customer based on the selection result of the matching means;
An information distribution apparatus comprising:

Images