JP2005063215A - Assortment proposal system and assortment proposal program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To realize a technique capable of presenting an optimum assortment plan in a specified store. <P>SOLUTION: This assortment proposal system 10 comprises an index single item/classification DB 22 for defining an index single item/classification for determining the family structure of customers; a customer attribute DB 26 for storing static attribute of each customer; a purchase history DB 28 for registering purchased commodities and purchase dates of each customer for every store; a classification definition DB 24 for setting a condition for classifying customers to a plurality of segments; a means for classifying customers based on the static attribute and the purchase history; a means for tabulating segment-based visitor numbers for every store in a set period; a means for tabulating the sales number in each store for an object commodity by segments; a means for calculating an PI value between a plurality of stores by dividing the sales number by segments by the segment-based visitor number in the plurality of stores; a means for calculating the assortment number by segments by multiplying the PI value by the segment-based visitor number of a specified store; and a means for guiding the total assortment number by integrating the assortment numbers of all the segments. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an assortment proposal system and an assortment proposal program, and in particular, categorizes customers into a plurality of segments based on purchase history information and static attribute information, and based on the number of customers visiting the store and sales results by segment. In particular, the present invention relates to a technique for calculating a product assortment quantity and a composition ratio.

In order to improve sales and profits in large-scale retail stores such as supermarkets and department stores, it is extremely important to predict customer demand and purchasing behavior in advance, but so far it has relied on the experience and intuition of field staff. As a result, there were variations in accuracy among people.
For this reason, several systems for supporting objective and quantitative prediction have been proposed.
For example, in the “customer classification method and customer classification device” described in Patent Document 1, each customer is segmented according to the three axes of the average visit interval of customers, the average usage unit price, and the number of days elapsed since the last use date. Therefore, it is intended to predict the number of prospective customers and the expected usage amount in the future.
JP 2003-44653 A

Assuming that future sales can be predicted using this customer classification method and customer classification device, it is possible to determine the upper limit value of the promotional expenses that can be surely spent as the effect of the invention. .
However, in the case of this conventional technology, the future is only forgotten based on the past results, and it cannot be expected to propose an improvement measure for the current situation and actively increase sales.
In order to increase sales at retail stores beyond the current level, assortment policies regarding products other than the so-called “standard” products that can be sold silently are the most important. Instead, it is necessary to derive the number of purchases and composition ratio of each product based on the customer's product preference pattern.
However, a support system that can present an assortment plan for a specific store in consideration of a customer's product preference pattern, including the above-mentioned Patent Document 1, has not existed so far.
In the first place, in the case of the above-mentioned conventional technology, the average visit interval, the average usage price, and the last usage date are used as the criteria for segmenting customers. This will greatly influence the product preferences, such as the family structure and hobbies of customers. The dynamic attributes of the product cannot be fetched, and therefore an appropriate assortment plan for the product cannot be presented.

  This invention was made to solve the above-mentioned problems of the prior art, and based on dynamic attributes such as family composition and hobbies that have a great influence on customer product preference patterns, The purpose is to realize a technology that can present an optimal product assortment plan.

In order to achieve the above object, the assortment suggestion system according to claim 1 is configured such that a product and a product classification, which serve as an index for determining a dynamic attribute such as a customer's family structure and hobby preferences, Individual index / classification storage means defined as classification, customer information storage means for registering static attributes such as age and gender of each customer, and each customer's purchase products, purchase date and time, etc., registered for each store Setting conditions for classifying each customer into multiple segments based on the combination of the purchase history storage means to be stored, the single item purchased by the customer, the product corresponding to the index classification, and the static attribute of the customer Classification definition storage means, and means for classifying customers into a plurality of segments based on the respective static attributes and purchase history with reference to each of the storage means described above,
Referring to the purchase history storage means, the means for counting the number of customers visiting each segment for each store within the set period, and the number of sales within the set period at each store for each preset target product Means for calculating the PI value between a plurality of stores by segment by dividing the number of sales by segment for the target product by the number of customers visiting by segment within a set period at a plurality of stores; By multiplying the PI value by segment between stores by the number of customers visiting by segment at a specific store, a means for calculating the number of assortments by segment within the set period of the store, and for all segments related to the target product And a means for deriving the total number of items in the specific store by accumulating the number of items. doing.
The assortment suggestion program according to claim 5 defines a product and a classification as an index for determining dynamic attributes such as a family structure of a customer and a hobby preference as an index single item and an index classification. Individual index / classification storage means, customer information storage means for registering static attributes such as age and gender of each customer, purchase history storage for registering each customer's purchased products, purchase date and time, etc. for each store Means storage means for setting a condition for classifying each customer into a plurality of segments based on a combination of a single item purchased by the customer and a product corresponding to the index classification and a static attribute of the customer Means for classifying customers into a plurality of segments based on the respective static attributes and purchase history, referring to the respective storage means, and referring to the purchase history storage means for segmenting each store within a set period. The number of customers visiting each store, the number of sales by segment within the set period for each target store, and the number of customers visiting by segment within the set period at multiple stores. A means to calculate the PI value between multiple stores by dividing the number of sales by segment for the target product, multiplying the PI value by segment between multiple stores and the number of customers visiting by segment at a specific store Means for calculating the number of assortments for each segment within the set period of the store, and means for deriving the total number of assortments of the product at the specific store by integrating the number of assortments for all segments related to the target product It is characterized by making it function as.

Conventional customer analysis methods estimate the average life stage (married / unmarried, family structure such as the presence of children) based only on static attributes such as customer age and gender. In present-day Japan, where diversification has progressed, it has not been possible to accurately understand the customer's product preference behavior.
On the other hand, this assortment proposal system and program estimate the current family structure and hobbies preference based not only on the customer's static attributes but also on the product actually purchased by the customer. Since it is a mechanism for classifying customers who are likely to show preference trends into a common segment, it is possible to predict future purchase behavior more accurately by analyzing past purchase results for each segment.
Specifically, the number of sales of target products at multiple stores and the number of customers visiting the store are aggregated by segment, and based on these, PI values between multiple stores are calculated, and based on this, the product for each segment at a specific store is calculated. The ideal number of items in the store can be obtained by deriving the number of items and integrating the number of items in all segments.

The assortment proposal system according to claim 2 is based on the system of claim 1 and further divides the total number of the target products by the total value of the total number of the multiple products including the target products. Thus, a means for calculating the assortment composition ratio of the product is provided.
The assortment proposal program according to claim 6 is based on the program according to claim 5, and further, the computer calculates the total assortment of the target product by the total value of the assortment of a plurality of products including the target product. By dividing, it functions as a means for calculating the assortment composition ratio of the product.
As a result, an ideal assortment composition ratio of the target products in the specific store can be obtained, and an assortment policy can be easily planned.

The assortment suggestion system according to claim 3 is an index single item that defines a product and a product category as an index for determining dynamic attributes such as a customer's family structure and hobbies and preferences as an index single item and an index category. Classification storage means, customer storage means for registering static attributes such as age and gender of each customer, purchase history storage means for registering each customer's purchased products, purchase date and time, etc. for each store, Classification definition storage means for setting conditions for classifying each customer into a plurality of segments based on a combination of a single item purchased by the customer and a product corresponding to the index classification and a static attribute of the customer, Refer to each of the above storage means, classify customers into a plurality of segments based on their respective static attributes and purchase history, and refer to the purchase history storage means for each segment for each store within the set period ·time A means of counting the number of customers coming to another store, a means of counting the number of sales within a set period at each store for each preset period, and a segment for each segment within a set period at multiple stores.・ A means to calculate the PI value between multiple stores by segment / time by dividing the number of sales by segment / time by the number of customers visiting by time, and between the stores By multiplying the PI value for each segment / time zone by the number of customers visiting the specific store by segment / time zone, the number of assortments by segment / time zone within the set period of the store is calculated. Means, and means for deriving the total number of assortments by time of the product at a specific store by accumulating the number of assortments of all segments related to the target product. It is characterized by a door.
Further, the assortment proposal program according to claim 7 defines a computer as a single item and an index classification as an index and a product category as an index for determining a dynamic attribute such as a customer's family structure and hobbies and preferences. Individual index / classification storage means to be registered, customer storage means to register static attributes such as age and sex of each customer, purchase history storage to register each customer's purchased products, purchase date and time, etc. for each store Means storage means for setting a condition for classifying each customer into a plurality of segments based on a combination of a single item purchased by the customer and a product corresponding to the index classification and a static attribute of the customer Means for classifying customers into a plurality of segments based on the respective static attributes and purchase history, referring to the respective storage means, and referring to the purchase history storage means for segmenting each store within a set period. A means of aggregating the number of in-store customers by event and time, a means of aggregating the number of sales within a set period at each store for each preset product, and within a set period at multiple stores By calculating the PI value between multiple stores by segment and by time, by dividing the number of sales by segment and by time for the target product by the number of customers visiting by segment and by time By multiplying the PI value by segment and time zone between stores by the number of customers visiting the specific store by segment and time zone, the number of items sorted by segment and time zone within the set period of the store is calculated. A means for calculating, as a means for deriving the total number of assortments by time of the product at a specific store by accumulating the number of assortments for all segments related to the target product. It is characterized in that to function.
As a result, the total assortment of specific products in specific stores by time zone (for example, every 3 hours, by morning / afternoon / night time) is obtained. It is effective for planning.

The assortment suggestion system described in claim 4 is based on the system of claim 3, and further, the total assortment of the target products by time zone is set as the total assortment of the plurality of products including the target product by time zone. It is characterized by comprising means for calculating a product assortment ratio by time of the product by dividing by the total value.
The assortment proposal program according to claim 8 is based on the program according to claim 7, and further includes a computer to calculate the total number of assortment of target products by time zone for a plurality of products by time zone including the target products. By dividing by the total value of the total number of assortments, the product functions as a means for calculating the product assortment ratio by time of the product.
As a result, an ideal assortment composition ratio of the target products in the specific store can be obtained for each time zone.

  In the assortment proposal system and the assortment proposal program according to the present invention, as described above, the family structure estimated based on the customer's static attributes and the purchase history of the commodity corresponding to the single index item or the index classification Considering dynamic attributes such as hobbies and tastes, classify customers who may have similar product preference trends into the same segment, and based on the number of customers visiting the store by segment and the number of sales of the target product at multiple stores Calculate the PI value between multiple stores, multiply this by the number of customers visiting each segment at a specific store, derive the number of items by segment, and add up the number of all segments to total the number of items at that store. As a result, it is possible to present an assortment plan that meets the needs of customers, which contributes to improved sales and reduced inventory loss at the store.

  FIG. 1 is a block diagram showing a functional configuration of an assortment proposal system 10 according to the present invention. A customer classification unit 12, a customer count calculation unit 14, a store PI value calculation unit 16, and a store cluster PI Value calculation unit 18, assortment proposal unit 20, index single item / classification database 22, classification definition database 24, customer attribute database 26, purchase history database 28, customer classification database 30, and period designation database 32 A store customer number database 34, a store-specific PI value database 36, a store classification database 38, a store cluster-specific PI value database 40, a proposal target single product database 42, and an assortment proposal database 44 are provided.

Among these, the customer classification unit 12, the store visitor number calculation unit 14, the PI value calculation unit 16 for each store, the PI value calculation unit 18 for each store cluster, and the assortment proposal unit 20 include a dedicated application program and a CPU of the computer. This is realized by executing necessary processing according to the OS.
In addition, index single item / classification database 22, classification definition database 24, customer attribute database 26, purchase history database 28, customer classification database 30, period specification database 32, store visitor number database 34, store-specific PI value database 36, store classification database 38, the store cluster-specific PI value database 40, the proposal target single item database 42, and the assortment proposal database 44 are stored in the hard disk of the computer.

Hereinafter, according to the flowchart of FIG. 2, a process for obtaining an assortment plan at a specific store using the system 10 will be described in order.
(1) Customer classification In this system 10, the process of first classifying registered customers into multiple segments according to their life stages based on the customer's purchase history and static attributes is the customer classification. Executed by part 12.

First, in the index single item / classification database 22, products and product categories effective for estimating the family structure of customers are registered in advance.
The following is an example.
(Assumed family members) (Indicator single item / classification)
Infants: Powdered milk, baby products, baby food,
Infant toothbrush A ...
◆ Children: Girls 'clothing, boys' clothing, character stationery,
Study desk B for elementary school students ...
◆ Men ... Men's clothing, men's cosmetics, men's watch C ...
◆ Women ... Women's clothing, cosmetics, women's thermometer D, ...
As a result, the fact that a certain customer purchased a specific baby powder belonging to “baby goods” in the past was found from the records in the purchase history database 28, or the fact that a baby toothbrush A was purchased in the past was also found. In this case, the customer classification unit 12 can determine that “the infant is included in the family members” for the customer.

Here, “indicator single item” refers to a specific product on the assumption of the existence of a specific family member, and “indicator category” refers to a product category including such a product.
For example, men's underwear is originally associated with a product category that is directly linked to the personal attribute (male) of “gentleman's clothing”. Belongs to the general category “toothbrush”, and is registered in the index single item / category database 22 in units of specific products.

In the classification definition database 24, conditions for segmenting customers by life stages are set.
FIG. 3 is a chart showing the setting contents, and it is assumed that the life stages of each customer are classified into 1 to 14 segments.
The “gentleman's goods” listed in the “Purchase history” column in this chart is an index single item indicating the existence of men in the family, and the indicator classification. , It means that the family has a purchase history that can be estimated that only men exist.
Similarly, “Women's goods” is a single indicator and index classification indicating that women are in the family, and “Women's only purchasing history” is a purchase history that can be estimated that only women in the family exist. It means having.
“Infant supplies” is a single indicator and index classification indicating that there are infants in the family. “Infant supplies purchase history” is a purchase history that can be estimated that at least an infant exists in the family. It means having.
Similarly, “children's goods” refers to a single indicator and an indicator classification indicating that there are children in the family, and “with children's goods purchase history” is a purchase history that can be estimated to have at least children in the family. It means that it has.

In the customer attribute database 26, static attribute information such as sex and age collected when a customer joins a card member is registered in association with the customer code.
The purchase history database 28 stores detailed information when a customer makes a purchase by presenting a card at an affiliated store. Examples of the detailed information registered in the purchase history database 28 include store code, customer code, JAN code, sales date and time, sales quantity, sales amount, and the like.

The customer classification unit 12 refers to the customer attribute database 26, the purchase history database 28, the index single item / classification database 22, and the classification definition database 24 regularly or as needed, and corresponds each customer to the life stage. The data is classified into segments, and the classification results are registered in the customer classification database 30 (S10 to S18).
Hereinafter, a specific example of customer segmentation by the customer classification unit 12 will be described.

(Case 1)
○ Static attribute of customer A:
28-year-old male ○ Customer A purchase history:
Purchase history of infant items: None Purchase history of children's items: None Purchase history of men's items: Yes Purchase history of women's items: None In this case, the customer classification unit 12 assigns the customer A segment ID according to the classification definition of FIG. The classification result (segment ID: 1) is registered in the customer classification database 30 in association with the customer code.

(Case 2)
○ Static attribute of customer B:
32-year-old female ○ Customer B purchase history:
Purchase history of infant products: None Purchase history of children's products: None Purchase history of men's products: Yes Purchase history of women's products: None In this case, the customer classification unit 12 assigns the customer B segment ID according to the classification definition of FIG. 4 and classifying the result of the classification (segment ID: 4) in the customer classification database 30 in association with the customer code.
By the way, in this segment ID: 4, ID: 10, and ID: 14, those who are buying women's goods even though they are women, or women's goods even though they are men Because the person is presumed to have a spouse, the name “Family's Family” is given. Of course, if there is a purchase history of both men's goods and women's goods, it is classified as a "couple's family" regardless of gender.

(Case 3)
○ Static attribute of customer C:
42-year-old male ○ Purchase history of customer C:
Purchase history of infant supplies: Yes Purchase history of children's goods: No Purchase history of gentlemen's goods: Yes Purchase history of ladies' goods: Yes In this case, the customer classification unit 12 assigns the customer C to the segment ID according to the classification definition of FIG. No. 5 is classified into “the family of the smallest child is an infant”, and the classification result (segment ID: 5) is registered in the customer classification database 30 in association with the customer code.
In the case of this customer C, according to the static attribute and purchase history, it corresponds to “middle-aged couple's family” of segment ID: 10, but the priority of segment ID: 5 is “1”. In order to win the priority “3” of the segment ID: 10, the customer C is classified as “the smallest child is an infant family”.

(Case 4)
○ Static attribute of customer D:
26-year-old male ○ Customer D purchase history:
Purchase history of infant items: None Purchase history of children's items: Yes Purchase history of men's items: Yes Purchase history of women's items: No In this case, the customer classification unit 12 assigns the customer D to the segment ID according to the classification definition of FIG. No. 6 is classified into “the smallest child is a child's family”, and the classification result (segment ID: 6) is registered in the customer classification database 30 in association with the customer code.
In the case of this customer D, according to the static attribute and purchase history, it corresponds to “young single man” of segment ID: 1, but the priority of segment ID: 6 is “2”, and segment ID , Customer D is classified as “the smallest child is the child's family”.

(Case 5)
Customer E static attributes:
72 years old, female Customer E purchasing history:
Purchase history of infant items: None Purchase history of children's items: None Purchase history of men's items: None Purchase history of women's items: None In this case, the customer classification unit 12 assigns the customer E a segment ID according to the classification definition of FIG. Classification is made into 13 “senior women (other)”, and the classification result (segment ID: 13) is registered in the customer classification database 30 in association with the customer code.
In the case of this customer E, there is no purchase history of the product corresponding to the index single item / index classification, and other food items are purchased, so it is not possible to discriminate between a single person and a married person, and only gender and age are known. Classified as “senior women (others)”.

This system 10 is based on the assumption that customers belonging to the same life stage have similar product preference trends.
The classification definition shown in FIG. 2 is merely an example, and it is naturally possible to classify customer life stages from other viewpoints, in which case the corresponding classification conditions are set in the classification definition database 24. You just have to.
It is also possible to classify customers into multiple segments according to scales other than the life stage, such as lifestyles based on hobbies, preferences, age, and gender. The classification and hobbies / preferences should be registered in association with each other, and the classification condition based on the combination of the customer's static attribute and the purchase history of the single index / index classification should be set in the classification definition database 24. .

In the above case, if the customer has purchased a product that falls under the index single item / index classification even once within a certain period in the past, it will be classified into the corresponding segment. The frequency may be set in the classification definition database 24 as a classification condition.
For example, even if a single man asks his sister who has a child to purchase a disposable diaper once, the customer is classified as “the smallest child is an infant's family” in the above classification process. However, with regard to segment ID: 5, by setting the condition that “buy baby products at least twice in 3 months”, the customer classification unit 12 will make the customer segment according to the actual situation. Can be realized.

(2) Calculation of the number of customers visiting by store When the segmentation of all registered customers by the customer classification unit 12 is completed as described above, the customer number calculation unit 14 is activated, and the customer classification database 30 and the purchase history database 28 are activated. And the period designation database 32 is referred to (S20).
Then, the store visitor number calculation unit 14 aggregates the number of customers who visited each store in the series within the period set in the period designation database 32 by segment (S22) and stores it in the store visitor number database 34 (S24). ).
The registration example is shown below.
Store code: 001
Target start date: 2003/01/01
Target end date: 2003/03/31
Target segment: Middle-aged couple's family Customers: 1,000

The number of customers visiting the store means the total number of customers visiting the store during the period. When the customer presents a membership card at the cash register and the customer code is read by the reader, the number is counted as one person.
When the same customer visits the store several times on different days within the specified period, each person is counted as one person.
On the other hand, when the same customer purchases a product multiple times at the store on the same day, even if a plurality of records are left in the purchase history database 28, one person is counted.

(3) Calculation of store-specific PI value At the stage where the number of customers visiting each store is calculated, the store-specific PI value calculation unit 16 is activated, and the store visitor number database 34, the proposed single product database 42, the customer classification database 30, and The purchase history database 28 is referred to (S26).
Then, the PI value for each segment in each store is calculated for each target product registered in the proposal target single item database 42 (S28).
Here, “PI (Purchase Index) value” is defined as follows.
PI value = number of sales of the target product within the period / total number of customers visiting the store during the period

For example, if the number of products α sold by “middle-aged couples” at store X during the period is 8, and the total number of customers visiting “middle-aged couples” at store X is 1,000. , PI value is 8 / 1,000 people = 0.008.
The store-specific PI value calculation unit 16 calculates a segment-specific PI value for each product set in the proposal target single product database 24 for each store, and stores it in the store-specific PI value database 36 together with the sales point information for each segment. (S30).

By analyzing the PI value by segment, it becomes possible to extract the support tendency by customer segment among similar products.
For example, even if it is the same sports drink, P drink has a relatively strong support from customers who have children in the family, while A drink has a support from other young people Can be fed back to product planning and advertising strategies.

(4) Calculation of store cluster PI value When the PI value at each store is calculated, the store cluster PI value calculation unit 18 is started, and the store PI value database 36, store classification database 38, and store The customer number database 34 is referred to (S32).
Then, the PI value calculation unit 18 for each store cluster calculates the PI value for each store cluster to which each store belongs, for each customer segment and each target product (S34), and registers the calculation result in the PI value database 40 for each store cluster ( S36).

That is, the affiliated stores belonging to the supermarket chain can be classified into a plurality of clusters from the viewpoint of the scale and location conditions. For example, a large-scale suburban store with a large-scale parking lot, a small-scale store targeting commuters attached to a terminal station, or a medium-sized store with a relatively small parking lot located in an urban area The customer base and top-selling products are different.
In this system 10, the conditions of the affiliated stores are input in advance into a cluster analysis program and classified into a plurality of clusters, and the results are registered in the store classification database 38.
For this reason, the PI value calculation unit 18 for each store cluster refers to the store classification database 38, and totals the number of stores by customer segment among a plurality of stores belonging to the same cluster and the number of sales of the target product. The value is calculated.
For example, if the number of customers of “middle-aged couples” during the entire cluster to which the X store belongs is 40,000 and the sales volume of the product α is 400, the PI value by store cluster is as follows: Become.
400 / 40,000 = 0.01

(5) Calculation of the total number of assortments When the PI value for each store cluster is calculated as described above, the assortment proposing unit 20 is activated and calculates the number of assortments of target products in each store (S38). .
For example, in the store cluster to which the X store belongs, the PI value for each store cluster related to the product α by the “middle-aged couple's family” is 0.01, and the number of customers of the “middle-aged couple's family” at the X store during the same period is 1,000. In the case of a person, an ideal assortment quantity is calculated by the following formula.
1,000 people × 0.01 = 10

That is, from the fact that the product α has been sold to 10 “middle-aged couples” in the past 3 months, despite the fact that it has sold only 8 in the past three months, It turns out that there were two lost opportunities.
By executing this calculation for all customer segments and obtaining the aggregated value, the sales forecast quantity for the product in X store for three months = the total number of assortments is derived (S40).
This calculation result (total number of assortments) is registered in the assortment proposal database 44 by the assortment proposal unit 20 (S42).

(6) Calculation of Assortment Composition Ratio Next, the assortment proposal unit 20 calculates an ideal assortment composition ratio of the target products at each store.
Here, the “product assortment ratio” means a ratio that the product should occupy in the same type of product category in a specific store.
In other words, the total number of products sold in the three months can be predicted by the calculation process of the total number of items in (5) above, but if an ideal occupancy rate in the sales floor of the product can be derived one step further, This makes it possible to strongly support the planning of assortment policies at the store.

For example, in determining the proportion of a particular brand / size chocolate to be displayed at a confectionery corner in a supermarket, the assortment proposing unit 20 refers to the proposal target single item database 42 and displays it in the confectionery corner. The JAN codes of all products to be acquired are acquired, and the assortment proposal database 44 is referenced to obtain the total assortment of each product calculated in (5).
Next, the assortment proposing unit 20 calculates the composition ratio by totaling the total assortment of each product and dividing the total assortment of chocolates by this total value (total sales forecast number during the period) (S44). ).
Here, if the total sales forecast for all products belonging to the confectionery corner is 2,000, and the total number of chocolate items is 82, 82 ÷ 2,000 × 100 = 4.1%. An appropriate composition ratio is required.
This calculation result (assortment composition ratio) is registered in the assortment proposal database 44 by the assortment proposal unit 20 (S46).
On the store side, it is possible to reduce the occurrence of inventory loss and opportunity loss by referring to this component ratio when purchasing products. In addition, when a variety of products are displayed at the storefront, the limited shelf space can be effectively used by referring to the composition ratio.

A registration example of the assortment proposal database 44 is shown below.
Assortment proposal pattern ID: 001
Store code: 001
JAN code: 4900000000001
Target start date: January 01, 2003 Target end date: March 31, 2003 Number of assortment proposals: 82
Assortment proposals (all categories): 2,000
Product composition ratio: 4.1%

Although the example which calculates the ratio which a specific chocolate should occupy in all the goods of a confectionery corner was shown in the above, this invention is not limited to this.
In other words, there are “selling products” that can be sold at a certain number or more in any store, and it can be said that there is little room to devise an assortment for these items.
Similarly, there are “dead-skinned products” that can hardly be sold regardless of how they are devised, and there is little room to devise an assortment of goods for these items.
Therefore, for example, the assortment composition ratio is calculated only for the top 5% to 20% products excluding the best selling products whose sales volume is within the top 4% of the total and the top sales products whose sales volume is 21% or less. Can be operated as follows.
Specifically, it is only necessary to register in the proposal target single item database 42 only the products that belong to the top 5 to 20% of sales.

By the way, if a product during the specified period is subject to sale, it is appropriate to calculate the PI value by store (3) and the PI value by store cluster (4) above based on the sales volume. Since a product assortment proposal result cannot be obtained, it is desirable to provide a mechanism for excluding the sales volume at the time of special sale.
Specifically, a database that stores the JAN code of the product to be sold and the sale period is set up, and the store-specific PI value calculation unit 16 refers to the database when calculating the store-specific PI value. The PI value may be calculated by excluding the number of products sold during the special sale period.

In the above, an example is shown in which the PI value is calculated for each store cluster, and the total number of products and the product composition ratio of each store are calculated based on the PI value. However, the present invention is not limited to this. .
For example, when the types of affiliated stores are common, the PI value in all affiliated stores can be calculated, and based on this, the total number of items in the store and the product composition ratio can be derived.

By introducing a time element called “time zone” into the system 10, it becomes possible to make a more practical assortment proposal.
For example, in the case of prepared food products, small packages for a single person sell well for office workers and OLs at lunch, whereas housewives often buy large packs for families at dinner, so there are many selling points depending on the time of day. There is a tendency to change.
For this reason, depending on the target product, it is effective to calculate a PI value for each time zone and to finely derive the number of assortment proposals and the assortment proposal ratio based on this.

Specifically, it is realized by installing a database defining time zones in the system 10 and registering the definition pattern ID set therein in the proposal target single product database 42.
An example of the definition pattern is shown below.
Pattern 1: Calculate in units of 3 hours from opening to closing Pattern 2: Calculate by AM / PM / Night For example, if pattern 1 is set for a product of potato salad (large), the PI value calculation unit for each store 16 Adds up the data in the purchase history database 28 in units of 3 hours and calculates the PI value for each time zone. Similarly, the store cluster specific PI value calculation unit 18 also calculates the store cluster specific PI value by time period.
Further, the assortment proposing unit 20 calculates the total assortment for each product for each time period and also calculates the assortment composition ratio for each time period.

It is a block diagram which shows the basic composition of an assortment proposal system. It is a flowchart which shows the process sequence of an assortment proposal system. It is a chart which shows the example of a setting in a classification definition database.

Explanation of symbols

10 Assortment proposal system
12 Customer Classification Department
14 Number of customers visiting store
16 Store-specific PI value calculator
18 PI value calculator by store cluster
20 Assortment Proposal Department
22 Single indicator / classification database
24 Classification definition database
26 Customer attribute database
28 Purchasing history database
30 customer classification database
32 timed database
34 Customer Number Database
36 PI value database by store
38 Store classification database
40 PI cluster database by store cluster
42 Proposed single item database
44 Assortment proposal database

Claims (8)

An index single item / classification storage means for defining a product and a product classification as an index for determining dynamic attributes such as a customer's family structure and hobby preferences, as an index single item and an index classification;
Customer information storage means for registering static attributes such as age and gender of each customer;
A purchase history storage means for registering each customer's purchased products, purchase date and time, etc. for each store;
Classification definition storage means for setting conditions for classifying each customer into a plurality of segments based on a combination of a single item purchased by the customer and a product corresponding to the index classification and a static attribute of the customer,
Means for classifying the customers into a plurality of segments based on the respective static attributes and purchase history with reference to each of the storage means;
Referring to the purchase history storage means, a means for counting the number of customers visiting each segment for each store within a set period;
For the target product set in advance, means for totaling the number of sales within each set period in each store by segment,
By calculating the PI value between multiple stores by segment by dividing the number of sales by segment for the target product by the number of customers visiting by segment within the set period at multiple stores,
Means for calculating the number of assortments by segment within the set period of the store by multiplying the PI value by segment between the plurality of stores by the number of customers visiting by segment at a specific store;
An assortment suggestion system comprising: a means for deriving the total number of assortments of the product at a specific store by accumulating the number of assortments of all segments related to the target product.   A means for calculating an assortment composition ratio of the product by dividing the total assortment of the target product by a total value of the total assortment of a plurality of products including the target product. Item 3. The assortment proposal system according to item 1. An index single item / classification storage means for defining products and classifications as indicators for determining dynamic attributes such as customer family structure and hobbies preferences as index single items and index classifications;
Customer storage means to register static attributes such as age and gender of each customer,
A purchase history storage means for registering each customer's purchased products, purchase date and time, etc. for each store;
Classification definition storage means for setting conditions for classifying each customer into a plurality of segments based on a combination of a single item purchased by the customer and a product corresponding to the index classification and a static attribute of the customer,
Means for classifying the customers into a plurality of segments based on the respective static attributes and purchase history with reference to each of the storage means;
Means for counting the number of customers visiting by segment and by time zone for each store within the set period with reference to the purchase history storage means;
With respect to the target products set in advance, means for counting the number of sales in each store within the set period, by segment and by time zone,
By dividing the number of customers by segment and time period for the above target products by the number of customers visiting by segment and time period within the set period of multiple stores, the PI value between multiple stores can be calculated by segment and time period. Means for calculating separately;
Multiplying the above-mentioned PI values by segment and time zone between multiple stores by the number of customers visiting the specific store by segment and by time zone, the product lineup by segment and time zone within the set period of the store Means for calculating the number;
An assortment proposal system comprising: means for deriving the total number of assortments of a particular store by time zone by integrating the number of assortments of all segments related to the target product.   Means for calculating the assortment composition ratio by time of the product by dividing the total assortment by time of the target product by the total value of the total assortment by time of multiple products including the target product The assortment proposal system according to claim 3, further comprising: An index single item / classification storage means for defining a product and a product classification as an index for determining a dynamic attribute such as a family structure of a customer and a hobby preference as an index single item and an index classification;
Customer information storage means to register static attributes such as age and gender of each customer,
Purchase history storage means for registering each customer's purchased products, purchase date and time, etc. for each store,
Classification definition storage means for setting conditions for classifying each customer into a plurality of segments based on a combination of a single item purchased by the customer and a product corresponding to the index classification and a static attribute of the customer,
Means for classifying customers into a plurality of segments based on the respective static attributes and purchase history with reference to each of the storage means above;
Means for referring to the purchase history storage means and totaling the number of customers visiting each segment for each store within the set period;
A means for totaling the number of sales within each set period for each target segment for each target segment,
A means for calculating the PI value between multiple stores by dividing the number of sales by segment for the target product by the number of customers visiting by segment within the set period at multiple stores,
Means for calculating the number of assortments by segment within the set period of the store by multiplying the segment-specific PI value between the plurality of stores by the number of customers visiting by segment at a specific store;
An assortment suggestion program that functions as means for deriving the total number of assortments of a product at a specific store by accumulating the number of assortments of all segments related to the target product.   The computer is caused to function as a means for calculating an assortment composition ratio of the product by dividing the total assortment of the target product by the total value of the total assortment of the plurality of products including the target product. The assortment proposal program according to claim 5. An index single item / classification storage means for defining a product and a product classification as an index for determining a dynamic attribute such as a family structure of a customer and a hobby preference as an index single item and an index classification;
Customer storage means to register static attributes such as age and gender of each customer,
Purchase history storage means for registering each customer's purchased products, purchase date and time, etc. for each store,
Classification definition storage means for setting conditions for classifying each customer into a plurality of segments based on a combination of a single item purchased by the customer and a product corresponding to the index classification and a static attribute of the customer,
Means for classifying customers into a plurality of segments based on the respective static attributes and purchase history with reference to each of the storage means above;
Means for counting the number of customers visiting by segment and by time zone for each store within the set period with reference to the purchase history storage means;
A means for totaling the number of sales within each set period for each target and each time zone for the target product set in advance,
Divide the number of sales by segment / time for the above products by the number of customers visiting by segment / time during the set period at multiple stores to calculate the PI value between multiple stores by segment / time Means for calculating separately,
Multiplying the above-mentioned PI values by segment and time zone between multiple stores by the number of customers visiting the specific store by segment and by time zone, the product lineup by segment and time zone within the set period of the store Means for calculating the number,
An assortment suggestion program that functions as a means for deriving the total assortment of goods for each specific time period at a specific store by accumulating the assortment numbers of all segments related to the target product. The computer calculates the assortment composition ratio by time of the product by dividing the total assortment by time of the target product by the total of the total assortment by time of multiple products including the target product. 8. The assortment suggestion program according to claim 7, wherein the product assortment proposing program is made to function as a means.

Images