JP4442374B2 - Product share prediction device and program for share prediction - Google Patents

Description

  The present invention relates to a technique for predicting the share after the start of new product sales (the ratio of the number of sales for each product to the total number of sales of a competitive product group).

The main terms used in this specification will be explained.
The “attribute” of a product refers to an item representing product characteristics such as the size, function, performance, and price of the product. For example, if the product is an automobile, it refers to items such as size, displacement, fuel consumption, manufacturer, and price. If the product is a beverage, it refers to items such as type, raw material, and content.
“Level” refers to a specific value related to an attribute. When divided into a plurality of stages, it represents a value representing each stage. For example, if the product is an automobile and the attribute is the displacement, values such as 1.3 liters, 1.5 liters, and 2.0 liters are standard. If the product is a beverage and the attribute is the content, values such as 50 ml, 150 ml, 250 ml, and 350 ml are standard.

In order to plan a new product, it is necessary to accurately predict how much market share can be expected before producing and selling the new product. If the share of a new product after the start of sales of the new product can be accurately predicted, a production plan and a sales plan suitable for the new product can be formulated. In addition, since it is possible to predict how the share of existing products that compete with new products will change, it becomes possible to review production plans for existing products. As a result, not only new products but also existing products can be produced and sold in proportion to the actual demand, and it is possible to avoid the occurrence of large quantities of inventory, returns, or shortages. . Accurately predicting the share of new products is very important for companies.
Patent Documents 1 and 2 disclose techniques for predicting the share of new products. These technologies consider which existing products are similar to the planned new products. Predict the share of new products based on the share of existing products with similar new products.
JP 2002-99706 A Japanese Patent Laid-Open No. 11-134390

In the technique described in Patent Document 1, a product group is plotted on a product map in order to examine which existing product the planned new product is similar to. That is, the position coordinate value on the product map is calculated for each product based on the product specifications and the like. It is assumed that the existing product group plotted at a position close to the new product on the product map is similar to the new product.
In the technique described in Patent Document 2, sales points are digitized from specifications or the like that are important when a consumer purchases a product, and a product map in which products are coordinated based on the numerical value is created. It is assumed that the existing product group existing at a position close to the new product on the created product map is similar to the new product.
However, it is unclear whether or not the item used to create the product map is appropriate for a product such as an automobile that has both hobby and practicality. For example, if items such as “bonnet length” and “rear mirror shape” were selected to create a product map for automobiles, how much these items affect consumer product selection? It is difficult to know. Also, know whether the number of levels in the items (attributes) used to create the product map is close to the details that consumers will consider when selecting products, and whether it is necessary or sufficient It is also difficult. In this way, for products with a mixture of hobbies and practicality, in order to create a product map, what items (attributes) should be selected and how many levels should be classified That itself is very difficult.
Therefore, even if a product map is created based on items (attributes) that are not properly selected and the number of levels, the product map accurately represents the aspect of product similarity seen by consumers. Whether it is or not is unknown. As a result, there is a doubt as to whether the share predicted using the product map is accurate.

For this reason, Non-Patent Document 1 describes a technique for calculating a share for each product using a product map created by a multidimensional scaling method. In the multidimensional scaling method, an object whose map number of dimensions and dimension contents are unclear is coordinated on a multidimensional map in which the number of dimensions and the spatial structure are specified based on similarity data between objects. Arrangement is performed so that the similarity between objects is higher as the distance between the arranged positions is closer, and the similarity between objects is lower as the distance between the positions is farther. In the multidimensional scaling method, it is possible to create a product map that accurately represents the degree of similarity of products viewed from the consumer without specifying the attribute and the number of levels. In the technique described in Non-Patent Document 1, the share for each product is calculated by superimposing the analysis result of the consumer's product preference on the product map created by the multidimensional scaling method.
"New Consumer Analysis-Theory and Application of LOGMAP" by Hidetaka Katahira, University of Tokyo Press

In the technology described in Non-Patent Document 1, a product map that accurately represents the aspect of the similarity of products viewed from the consumer is created by multidimensional scaling, and the share is calculated based on the product map. The prediction accuracy is high. However, since the content of the product map dimension (what the dimension or axis on the map means) is not clear, the relationship between the product position on the product map and the level of each product attribute is not clear. The product position on the product map cannot be calculated from the level for each product attribute, and therefore the product position on the product map cannot be known for a new product. As a result, the share of a new product cannot be calculated, and the share of an existing product after the start of new product sales cannot be predicted.
Non-Patent Document 1 is a technique for analyzing the performance that consumers think is similar, and is not a technique for analyzing new products that have not yet been provided to consumers.

In the present invention, the technology described in Non-Patent Document 1 is used to create a product map using data that represents the results that consumers think are similar, that is, the degree of similarity that consumers feel about the product. Then, the relationship between “the product position on the product map and the level for each attribute of the product” is clarified. Then, using this relationship, the position on the product map of a new product whose level for each attribute is known but whose position on the product map is not known is clarified.
As described above, the position of the new product and the product group competing with it on the product map is clarified, and the share for each product can be calculated from the position.
In the present invention, data representing the degree of similarity felt by consumers is used, so that the labor and cost of data collection can be reduced. Moreover, since the share is calculated based on the aspect of the degree of similarity of the product as viewed from the consumer, the share can be calculated with high accuracy.

(One way to solve the problem)
The present invention can be embodied in a device that predicts the ratio of the number of sales for each product to the total number of sales of the competitive product group, that is, the share. This apparatus includes the following database and analysis means.
“Consumer / Purchase Existing Product / Comparison Existing Product / Priority Order” Database: Stores the purchased existing product, the existing product compared with the purchased existing product, and the priority order in association with each consumer.
“Product map creation means”: “Comparison existing product and its priority order” stored in the “ Consumer / Purchase existing product / Comparison existing product / Priority order” database is changed to “Existing product similar to purchase existing product and its similarity” A product map of existing products is created by multi-dimensional scaling as “order”.
"Existing product / attribute / level" database: Stores the level for each attribute for each existing product.
Calculating means for calculating a - "coordinate value level" function: the coordinate values explanatory variables on the product map of each existing products, "existing product / attribute / level" for each attribute of an existing product every database is stored calculating a - "standard coordinate value" function by regression analysis aimed variable levels, for each attribute, a regression equation for the purpose variable levels as explanatory variables coordinate values on the product map.
"Product profile creation means": for each existing products, together with associates to the axis of the product map and the attributes of the product, to create a product profile that is associated with the coordinate value on the axis and the level of its attributes.
“Contribution rate calculation means”: “ Consumer / existing product / comparison existing product / priority order” stored in the database, the pre-existing product, the comparison existing product, and the priority order, the product profile from the product profile group by conjoint analysis was selected order in which the selected for each consumer, to calculate the contribution of each "axis of the product map attached as an attribute of the product."
“New product / attribute / level” database: Stores the level of each new product attribute.
“New product position coordinate value calculation means”: As an explanatory variable when the level for each attribute of the new product stored in the “ new product / attribute / level” database is the objective variable of the “coordinate value-level” function , to calculate the coordinate value on the trade map of new products.
"Merchandise share calculation unit": the consumer each contribution rate for each "axis of the product map attached as an attribute of the product", a value obtained by multiplying the coordinate value of the product map of each product, for all items Select the value of the exponential function based on the total number of navy values in the “Product Map Axis Corresponding as Product Attributes” from among competing products after the start of new product sales for each consumer. It was calculated for each product as a probability, and calculates the total value for all consumers selection probability calculated for each product as a share of each commodity.

(The action and effect)
This apparatus stores data collected by conducting a questionnaire survey on consumers who have purchased products in a "consumer / purchased existing product / compared existing product / priority order" database. When selecting a product, the product selected in the end should be purchased by comparing with other products and repeating the comparison between the products. Therefore, the product to be compared in the product selection process is a product that the consumer feels similar to the actually purchased product. In the “consumer / purchased existing product / compared existing product / priority order” database, a higher priority order is stored for comparative products that remain in the examination target until the end of the selection process.
The product map creation means performs the multidimensional original scaling method with the priority order of the comparison existing products stored in the database as the order similar to the purchase existing products, and creates a product map as illustrated in FIG. The LOGMAP-M model described in Non-Patent Document 1 can be suitably used for creating a product map by the multidimensional scaling method.
When a product map is created using a multidimensional scaling method, it is not necessary to specify the attributes and the number of levels for creating a product map, and a product map that accurately represents the degree of similarity of products as seen by consumers can be created. . The product groups arranged close to each other on the product map are product groups that the consumer feels similar to, and the products that the consumer thinks are not similar are plotted at distant positions on the product map. Since the product map reproduces only the relative positional relationship representing the similar relationship between products, the directionality of the axis is meaningless and can be rotated in any direction. Therefore, as illustrated in FIG. 3, if the axis is rotated so that the products are distributed in all the areas divided by the axis, it becomes easy to execute the share calculation described below. When the product map is created by the multidimensional scaling method, the situation that the selection of the dimension or axis used for the product map is inappropriate is prevented. However, in the product map based on multidimensional scaling, the meaning of the axis is not clear.
When selecting a product, the consumer should make a comparison between products while evaluating (or not evaluating) the level for each attribute. Accordingly, coordinate values on the product map created by multidimensional scaling (hereinafter sometimes referred to as "position coordinates"), between the level of each attribute can be assumed constant relationship exists . FIG. 13 exemplifies this relationship. Between the position coordinate values of the products (here, automobiles) A to H and the attribute level “4100 mm, 4200 mm, 4300 mm,. There is a regression relationship represented by an ellipse. In the present invention, a method for calculating a function indicating a relationship between a position coordinate value on a product map and a level for each attribute has been created.
Based on the position coordinate value on the product map for each existing product and the level for each attribute for each existing product, the “position coordinate value-level” function calculation means of this device is used to calculate the position coordinate value and the level for each attribute. A function indicating the relationship is calculated. The function illustrated in FIG. 12 is a function that represents the regression relationship in FIG. 13. By using this function, the coordinate value Z1 of the first axis and the coordinate value Z2 of the second axis in the product map illustrated in FIG. It is possible to calculate the level Y of the attribute “total length” of the product “car” plotted in the position of Or conversely, the position coordinate values (Z1, Z2) on the product map of the product whose level of the attribute “full length” is Y can be calculated.
By using this “position coordinate value-level” function, the position coordinate value of the new product on the product map is calculated based on the level for each attribute of the new product stored in the “new product / attribute / level” database. Can be calculated. Since the position coordinate value of the new product can be obtained, it is possible to calculate the share for each product after the new product is sold.

In order for the product share calculation means to calculate the share of each product after selling a new product, a contribution rate for each “deemed attribute” for each consumer is required.
The “deemed attribute” is obtained by regarding each axis of the product map as an attribute of the product. In a product map based on multidimensional scaling, it is difficult to interpret what each axis means. In the present invention, each axis is regarded as representing some attribute of the product. At the same time, the position coordinate value for each axis is considered to represent the level of that attribute. The product profile creating means creates a product profile having “deemed level” in “deemed attribute” based on the position coordinate value of each existing product on the product map.
The contribution rate calculation means performs a conjoint analysis using the existing products purchased by the consumer, the existing products compared with them, and their priority order as the selection order when the product profile is selected from the product profile group, The contribution rate for each “deemed attribute” is calculated for each consumer. The contribution rate is the degree to which each “deemed attribute” contributes when a consumer selects a product, in other words, the degree to which a consumer places importance on each “deemed attribute” when selecting a product. Say.
In the conjoint analysis, the LOGMAP-D model described in Non-Patent Document 1 can be suitably used. For example, if there is a profile of products A to H, if the purchase product of the consumer α is A, the first comparison product is B, the second comparison product is D, the third comparison product is E, etc., the selection order Becomes {A, B, D, E,. In the LOGMAP-D model, the probability of obtaining this selection order is calculated, and a vector having the highest probability on the product map is obtained. The vector v4 in FIG. 9 is an example of this, and the direction of the vector v4 is calculated so as to be close to the direction of the vector (vector v3, etc.) of the position coordinate values of the products A, B, D, E,. Thus, it can be said that the vector v4 represents the preference of the consumer α. When each axis is regarded as a “deemed attribute”, the coordinate values of the first axis and the second axis of the vector v4 are the contribution rates for each “deemed attribute” calculated for the consumer α.
The above is merely an example of a method for calculating the contribution rate of “deemed attributes”. It is also possible to calculate the contribution rate of “deemed attribute” after analytically calculating the utility value for each attribute and level from the profile order.

Based on the calculated contribution rate for each “deemed attribute” for each consumer and the position coordinate value on the product map for each product, the product share calculation means calculates the share for each product after the start of new product sales. Can be calculated.
To calculate the share for each product, first, for each consumer, the probability of selecting one product from the product group after the start of new product sales is calculated for each product. The probability is calculated as an exponential value obtained by multiplying the product position coordinate value for each “deemed attribute” by the contribution rate of each “deemed attribute” and totaling all “deemed attributes”. Find it for each product. It is known that the share for each product is proportional to the exponential for each product.
It can be said that the position coordinate value for each product indicates the degree to which the product has some meaning of the “deemed attribute”. Further, it can be said that the contribution rate for each “deemed attribute” indicates the degree to which the “deemed attribute” is emphasized by the consumer when selecting a product. Therefore, if the product position coordinate value is multiplied by the contribution rate, it is possible to estimate how much the consumer is evaluating the “deemed attribute”.

As described above, in the present apparatus, a product map is created using the purchased product, the comparative product, and the priority order that represent the degree of similarity that the consumer feels for the product. Then, a function indicating the relationship between the position coordinate value on the product map and the level for each attribute of the product is calculated. By using this function, it is possible to accurately predict how much the share of each product will be after starting sales of a new product whose level for each attribute is known in advance.
Since this apparatus uses data representing the degree of similarity that consumers perceive for a product, the labor and cost of data collection can be reduced. In addition, it is possible to accurately calculate a share that closely matches the aspect of the degree of similarity of the product as viewed from the consumer.

(Preferred means for solving the problem)
The contribution rate calculation means, for consumers whose comparison existing products are not stored in the “consumer / purchased existing product / comparison existing product order” database, set a predetermined coordinate value on the product map of the purchased existing product. Multiply the coefficient to calculate the contribution with the vector in the same direction as the position coordinate value vector,
The product share calculation means, for a consumer whose comparison existing product is not stored in the “consumer / purchased existing product / comparison existing product order” database, a contribution rate to the “deemed attribute” of the consumer, and a product map Based on the position coordinate value for each product above and the value that is inversely proportional to the distance between the position coordinate value for each product and the position coordinate value of the consumer's purchased existing product, It is preferable to calculate the probability of selecting one product from each product.

(The action and effect)
Some consumers purchase a product without comparing it to other products. For such consumers, there is no priority order for comparison products, and the conjoint analysis as described above cannot be performed. In order to improve the prediction accuracy, it is necessary to obtain a contribution rate for such consumers.
Therefore, in this apparatus, the contribution rate calculation means calculates the contribution rate by using the position coordinate value on the product map of the consumer's existing purchased product. If another consumer has purchased the same existing product and there is a compared existing product, the position coordinate value should have been obtained. As described above, it is considered that the direction of the position coordinate value vector of the purchased existing product is close to the direction of the consumer contribution vector. For this reason, if the position coordinate value of the existing purchased product is used, a contribution rate that is relatively close to the contribution rate calculated by the conjoint analysis can be obtained. The contribution rate vector calculated by the conjoint analysis is scaled so as to have the same magnitude, and the contribution rate is multiplied by a predetermined coefficient for scaling. The product position coordinate value is multiplied by this predetermined coefficient to calculate a contribution rate having a vector in the same direction as the position coordinate value vector.
In addition, the product share calculating means calculates the position coordinate value for each product and the consumer's Use a value that is inversely proportional to the distance between the position coordinate value of the existing product purchased. This value is used to correct an error caused by the contribution rate of consumers who do not have a comparison product. The contribution rate of consumers without comparative products is not calculated by conjoint analysis. For this reason, if the share is calculated using this contribution rate, an error is likely to occur. In order to correct the error, a value that is inversely proportional to the distance between the position coordinate value of each product and the position coordinate value of the existing purchased product is used. In this value, the closer the position of the product is to the position of the existing purchased product, the greater the probability of selecting that product, and the farther the position of the purchased existing product is, the smaller the probability of selecting that product is. Can be corrected appropriately.
As described above, this apparatus calculates a contribution rate even for a consumer who does not have a comparison product, and uses a value for correcting an error that may occur from the calculated contribution rate. Thereby, even when there is a consumer who does not have a comparison product, the prediction accuracy of the share for each product can be maintained.

(Preferred means for solving the problem)
The product share calculating means includes a contribution rate for each “deemed attribute” for each consumer, a position coordinate value for each product on the product map, a position coordinate value for each product, and a position coordinate of the consumer's existing purchased product. It is preferable to calculate, for each product, the probability of selecting one product from the product group after starting a new product based on a value that is inversely proportional to the distance between the values.

(The action and effect)
For example, as shown in FIG. 19, when a consumer's purchase product is C and the first comparison product is G, the consumer's contribution rate vector may be v5. v5 has a direction close to the position coordinate value vector of the product H. Therefore, if the product share is calculated based only on the contribution rate for each “deemed attribute” of the consumer and the position coordinate value for each product on the product map, the probability that the consumer will select the product H is the product C It becomes larger than the probability of selecting (or product G). In order to prevent such a phenomenon, considering the distance between the position coordinates of the purchased product and the position coordinate for each product, the probability of selecting increases as the distance decreases, and the value of decreasing the probability of selecting as the distance increases. use.
As described above, this device appropriately corrects errors that may occur when product shares are calculated based solely on the contribution rate of each consumer's “deemed attribute” and the position coordinate value for each product on the product map. By doing so, the prediction accuracy is further improved.

(Preferred means for solving the problem)
New product position coordinate value changing means for changing the position coordinate value on the product map of the new product so that the share calculated by the product share calculating means becomes a desired value;
It is preferable to further include attribute level calculation means for calculating a level for each attribute of the new product that realizes the changed position coordinate value using the “position coordinate value-level” function.

(The action and effect)
There is a desire to know the level belonging to the attribute of a new product that maximizes the share of the new product. Alternatively, there is a desire to know the level belonging to the attribute of a new product that does not drop the share of the company's existing product. A level that belongs to the attribute of a new product that realizes a desired share can also be calculated using a share prediction device. The desired share here is not limited to making the share of a new product desirable, but includes a desired share when considering the entire share for each competitive product after the introduction of a new product.
The product share calculation means of the share prediction device calculates the share of the product based on the contribution rate for each “deemed attribute” for each consumer and the position coordinate value for each product on the product map. Therefore, conversely, if a desired share is determined in advance after the introduction of a new product, the new product position coordinate value changing means will make the product map based on the contribution rate for each “deemed attribute” so that it becomes the share. The upper position coordinate value can be changed.
The new product position coordinate value calculation means calculates the position coordinate value of the new product on the product map based on the level belonging to the attribute of the new product, using the “product position coordinate value−level” relation function. Conversely, the attribute level calculation means can calculate the level belonging to the attribute of the new product using the “product position coordinate value−level” relation function based on the changed position coordinate value of the new product.
As described above, with this device, it is desirable to create a new product that achieves a desirable share after the introduction of a new product, for example, maximizes the share of a new product, or does not reduce the share of the company's existing product. The level belonging to the attribute can be calculated.

(Preferred means for solving the problem)
The “existing product / attribute / level” database further stores, in correspondence with each existing product, whether it is its own product or another company's product,
The new product position coordinate value changing means has the smallest total share of the existing product group of other companies calculated by the product share calculating means based on whether the new product position coordinate value is stored in the “existing product / attribute / level” database. It is preferable to change the position coordinate value on the product map of the new product.

(The action and effect)
There is a desire to know the level to which the attributes of new products that minimize the share of the existing product group of other companies and the share of the entire product group of the company belong to among the competing product groups. In this case as well, the level belonging to the attribute of the new product that minimizes the share of the existing product group of the other company can be calculated using the share prediction device.
If the “existing product / attribute / level” database stores whether the company's product or the other company's product is stored, the product share calculating means can calculate the total share of the existing product group of the other company. And the new product position coordinate value changing means can change the position coordinate value on the product map based on the contribution rate for each “deemed attribute” so that the total share of the existing product group of the other company is minimized. . Furthermore, the attribute level calculation means can calculate the level belonging to the attribute of the new product that minimizes the total share of the existing product group of the other company from the changed position coordinate value.

(Preferred means for solving the problem)
The “existing product / attribute / level” database further stores, in correspondence with each existing product, whether it is its own product or another company's product,
The new product position coordinate value changing means has the maximum total share of all the product groups calculated by the product share calculating means based on whether it is the own company or another company stored in the “existing product / attribute / level” database. It is preferable to change the coordinate value on the product map of the new product so that

(The action and effect)
There is a demand for knowing the level to which the attribute of a new product that maximizes the total share of all of the company's product groups belongs among competing product groups. In this case as well, the level belonging to the attribute of the new product that maximizes the total share of all the product groups of the company can be calculated using the share prediction device.
If the “existing product / attribute / level” database stores whether the product is a company product or another company's product, the product share calculation means can calculate the total share of all the product groups of the company. And the new product position coordinate value changing means may change the position coordinate value on the product map based on the contribution rate for each “deemed attribute” so that the total share of all the product groups of the company is maximized. it can. Further, the attribute level calculation means can calculate the level belonging to the attribute of the new product that maximizes the total share of all the product groups of the company from the changed position coordinate value.

(Preferred means for solving the problem)
For example, a new product may be planned by changing the model of an existing product, and a new product that is completely different from the existing product may be planned. Alternatively, there may be a case where a plurality of new products for which a model change of an existing product is planned. Furthermore, a plurality of new products that are completely different from existing products may be planned. Thus, when planning a plurality of new products, it is necessary to consider a plurality of new products at the same time.
In this case, it is preferable that the new product position coordinate value changing unit changes each of the coordinate values on the product map of a plurality of new products.

(The action and effect)
If the new product position coordinate value change means can change each position coordinate value on the product map of multiple new products, it can calculate the share after introducing multiple new products, It becomes possible to consider the product group at the same time.

(One way to solve the problem)
The present invention can be further embodied in a program for predicting the ratio of the number of sales for each product to the total number of sales of the competitive product group, that is, the share. This program executes the following processing on the computer:
・ For each consumer, the existing product purchased, the existing product compared with the purchased existing product, and the priority order are associated with each other and stored in a computer-readable memory “consumer / purchased existing product / compared existing product / priority” Process to access the "order" database,
・ "Comparison existing product and its priority order" stored in the " Consumer / Purchased existing product / Comparison existing product / Priority order" database is defined as "Existing product similar to purchased existing product and its similarity order" as a multi-dimensional scale Process to create a product map of existing products by law,
For each existing product, a process of accessing an “existing product / attribute / level” database in which the level for each attribute is stored in a computer-readable manner;
・Regression formula with the coordinate value (position coordinate value) on the product map for each existing product as an explanatory variable, and the level for each attribute of each existing product stored in the “existing product / attribute / level” database as the objective variable and the process for calculating the - "level coordinate value" function, is
· The existing products each, together with associated with the axis of the product map and attributes of the product, and the process of creating a product profile associated with the coordinate values for that axis and the level of its attributes,
・ Purchase existing product, comparison existing product, and comparison existing product stored in the “ Consumer / Purchased Existing Product / Comparison Existing Product / Priority Order” database, and their priority order, selection order when a product profile is selected from the product profile group and the conjoint analysis that, for each consumer, a process of calculating the contribution of each "axis of the product map attached as an attribute of the items (deemed attribute)",
A process of accessing a “new product / attribute / level” database in which the level for each attribute of the new product is stored in a computer-readable manner;
- the level of each attribute of the new products that have been stored in the "new product / attribute / level" database - as an explanatory variable of time for the purpose variable of "coordinate value level" function, the coordinates on the commodity map of new products Processing to calculate a value (position coordinate value) ;
And consumer every "axis of the product map attached as an attribute of the items (deemed attribute)" and each of the contribution rate, the value obtained by multiplying the coordinate value of the product map of each product, for all items " As the probability of selecting the value of the exponential function based on the total number of navy values in the product map axes associated as product attributes from among the competitor product groups after the start of new product sales for each consumer A process of calculating for each product, and calculating a value obtained by adding the selection probabilities calculated for each product for all consumers as a share for each product;
Is executed.
(The action and effect)
This program uses a purchased product that represents the degree of similarity a consumer feels for a product, a comparative product, and its priority order, and then starts selling new products whose levels belong to attributes in advance, and then shares Can be predicted with high accuracy.

The best mode for carrying out the invention is listed below.
(Mode 1) The contribution rate for each “deemed attribute” for each consumer is calculated by a different method for a consumer having a comparison product and a consumer having no comparison product. For consumers having comparative products, the contribution rate is calculated by the LOGMAP-D model having the same principle as the conjoint analysis using the purchased product comparison priority order. For a consumer who does not have a comparison product, the position coordinate value of the same existing product as the existing product for purchase is multiplied by a predetermined coefficient to calculate a contribution rate having a vector in the same direction as the position coordinate value vector.

  In the present embodiment, the share of the newly planned car is predicted using the share prediction apparatus illustrated in FIG. The application target of the present invention is not limited to automobiles, and can be widely applied to products that are provided in a variety of types because there are various demands.

(First embodiment)
The share prediction apparatus in FIG. 1 is configured by a computer apparatus, and the hardware configuration is the same as a normal one. This apparatus includes a processing device (CPU), a storage device (an optical storage medium, a magnetic storage medium, or a semiconductor memory such as a RAM or a ROM), an input device, an output device, and the like. Various databases store information that can be used by computers. Various processing devices are realized by a computer operating according to an installed program.

The purchased product comparison priority order database (hereinafter abbreviated as DB) 2 in FIG. 1 stores the product purchased by the consumer and the product compared with the product at the time of purchase in association with the comparison priority order. . This information is collected from the results of questionnaire surveys with consumers.
FIG. 2 shows a data configuration example of the purchased product comparison priority order DB2. For example, consumer α has purchased the product A, the product compared to the first place is B, the product compared to the second place is D, the product compared to the third place is E, etc. ing. Some consumers do not have the compared products. For example, it is shown that the consumer γ purchased the product F but did not compare it with other products.

  The product map creation means 4 of FIG. 1 extracts and uses data of consumers having products compared at the time of product purchase from the purchased product comparison priority order DB 2. Here, data of consumers who do not have a product compared at the time of product purchase is not used. For example, in FIG. 2, consumers α, β, δ, etc. have comparative products, whereas consumers γ, etc. have no comparative products, so only using data on consumers α, β, δ, etc. Create a product map.

The product map creation means 4 creates a product map by a multidimensional scaling method. In this embodiment, a LOGMAP-M model of multidimensional scaling is used. This model can create a highly accurate product map using less data than traditional multidimensional scaling.
For example, FIG. 2 shows that the products compared when the consumer α purchased the product A are the product B, the second product D, the third product E, etc. . Considering the rank of the compared products as the order representing the similarity to the product A, the order of the similarity of the pair consisting of all the products included in the data of the consumer α is
{AB, AC, AE, ...}
It becomes. In this data, among all the pairs, AB is the pair with the highest similarity. In the LOGMAP-M model, this is interpreted as “AB is selected from a set of all pairs according to the criterion of similarity”. Then, the probability P (pi / pn) that a certain pair pi is selected from the set of all pairs pn is considered to be obtained by subtracting the lottery whose pair pi is the weight δi.

  Here, for example, the value of δi of the pair AB depends on the position coordinates of the product A and the position coordinates of the product B on the product map, and increases as the distance between the coordinates is closer. That is, as the distance between the coordinates is closer, the probability P (pi / pn) that the pair is selected increases. In the LOGMAP-M model, coordinate values that maximize the probability P (pi / pn) are estimated values of the position coordinates of the product A and the position coordinates of the product B. Similarly, the probability P (pi / pn) is obtained for all pairs using the above formula, and the coordinate value that maximizes the probability P (pi / pn) is used as the estimated value of the position coordinate of each product. In this way, a product map in which products are coordinate-arranged at each product position can be created.

FIG. 3 illustrates the created product map. Here, for convenience of understanding, a two-dimensional product map in which the number of dimensions is omitted is shown, but in actuality, it is an N-dimensional product map.
As described above, the distance between the coordinates of the position of each product represents the probability that the product pair is selected. The probability that a pair is selected represents the similarity of the pair. The similarity in this embodiment is a similarity from the viewpoint of “consumer's product purchase”. Therefore, if the distance is small, the similarity is high and the degree of competition between the products is high. If the distance is large, the similarity is similar. The degree of competition is low and the degree of competition between products is low. For example, since the distance between the position coordinate of the product A and the position coordinate of the product B is smaller than the distance between the position coordinate of the product A and the position coordinate of the product F, the product A is more similar to the product B than the product F. )
Note that it is difficult to interpret the meaning of the axes in a product map created by the multidimensional scale method. Therefore, what each axis means cannot be specified and is unknown here.

The coordinate values of the position of each product on the product map illustrated in FIG. 3 are as shown in FIG. For example, the position of the product A has a first axis coordinate value of −0.52, a second axis coordinate value of 0.46, a third axis coordinate value of 0.81,.
The coordinate value for each axis of the position for each product is stored in the product position coordinate value DB 6 of FIG.

  Returning to FIG. The contribution rate calculation means 8 in FIG. 1 calculates the contribution rate for each “deemed attribute” for each consumer. Here, a contribution rate is calculated by different processing for a consumer who has a product compared at the time of product purchase and a consumer who does not have a product compared at the time of product purchase. The detailed procedure of this process will be described with reference to the flowchart of FIG.

First, in step S102, a consumer having a product compared at the time of product purchase is distinguished from a consumer having no product compared at the time of product purchase. In the following, it is assumed that YES is determined in step S102, and processing for a consumer having a comparison product will be described.
In step S104, the product position coordinate values of the purchased product and the comparison product are read from the product position coordinate value DB 6. Since the product position coordinate value DB 6 stores the position coordinate values of the consumer's purchased product and the comparative product having the comparison product, the product position coordinate value shown in FIG. 4 is read.
In the subsequent step S106, the axis of the product map is determined as the “deemed attribute” of the product. Specifically, the axis of the product map in FIG. 3 is regarded as an attribute of the product. As described above, since the meaning of the axes is unknown, it is impossible to specify which attribute each axis corresponds to. Therefore, it is assumed here that each axis corresponds to some attribute. Hereinafter, the first axis is referred to as a first attribute, the second axis as a second attribute, the third axis as a third attribute, and so on.
In step S108, the position coordinate values of the purchased product and the comparative product are determined as “deemed level” values. It can be said that the position coordinate value indicates the degree of expression of how much the product represents some meaning of the axis. Such a position coordinate value is regarded as a level value set in the “deemed attribute” here. For example, for the product A shown in FIG. 4, the coordinate value −0.52 of the first axis is regarded as the value of the level set as the first attribute, and the coordinate value 0.46 of the second axis is set as the second attribute. The coordinate value 0.81 of the third axis is regarded as the level value set in the third attribute. In the following conjoint analysis, these “deemed levels” are handled in the same manner as the level values “4100 mm, 4200 mm, 4300 mm,.
In step S110, the product group for which the “deemed attribute” and the “deemed level” are determined as described above is handled as a profile (virtual product) group created by the conjoint analysis. In the normal conjoint analysis, for example, a profile of “car A having an attribute level of 4200 mm, an engine attribute level of FF, and a color attribute level of silver or the like” is created. Here, for example, the product A is handled as a profile of “the first attribute level is −0.52, the second attribute level is 0.46, the third attribute level is 0.81, and the like”.

In step S112, the priority order of the purchased product and the comparative product is read from the purchased product comparison priority order DB2.
Next, in step S114, the priority order of the purchased product and the comparative product is regarded as the order selected from the profile group, and the contribution rate for each “deemed attribute” is calculated for each consumer by conjoint analysis.
In normal conjoint analysis, a value survey is conducted on consumers, and the result is used to calculate the contribution rate for each attribute. Specifically, the group of profiles is shown to the consumer, and the consumer ranks the answers in the order in which he / she wants to purchase the group of profiles shown. The contribution rate for each attribute is calculated using the collected response results.
In this embodiment, the consumer has already purchased the product, and the priority order of the product compared at that time is obtained. Accordingly, in step S114, a conjoint analysis is performed using the purchased product comparison priority order, and a contribution rate for each “deemed attribute” is calculated. Purchased products in the order of preference for purchased products are treated as 1st selected products, 1st compared products as 2nd selected products, 2nd compared products as 3rd selected products, etc. To do.
For example, in FIG. 2, the consumer α is A for the purchased product, B for the first comparative product, D for the second comparative product, E for the third comparative product, etc., and the preference order is {A, B, D , E,... Here, the probability that this preference order is obtained is calculated using the LOGMAP-D model having the same principle as the conjoint analysis. For example, when the preference rankings of three products are (1>2> 3), the probability P of this preference ranking is
P ₁₂₃ = exp (V ₁ ) / {exp (V ₁ ) + exp (V ₂ ) + exp (V ₃ )} ·
exp (V ₂ ) / {exp (V ₂ ) + exp (V ₃ )}
Is calculated.
A vector having the highest probability P calculated based on the preference order {A, B, D, E,. FIG. 9 illustrates the obtained vector v4. It can be seen that the direction of the vector v4 is close to the direction of the position coordinate value vector of the products A, B, D, E,..., And it can be said that the vector v4 represents the preference of the consumer α. That is, the coordinate value of the vector v4 such as the first attribute can be said to be a contribution ratio of the first attribute to the consumer α. Here, scaling is performed so as to align the size of the contribution rate vector of each consumer, and the contribution rate is multiplied by a predetermined coefficient.
The contribution ratio calculated for each consumer in this way is illustrated in FIG. For example, the contribution rate for each “deemed attribute” of the consumer α is calculated as 0.27 for the first attribute, 1.19 for the second attribute, 3.31 for the third attribute, and the like. It can be said that these contribution ratios indicate the degree to which the “deemed attribute” is emphasized when the consumer purchases the product.

Next, the process for calculating the contribution rate of consumers who do not have a comparison product will be described. For example, the consumer γ shown in FIG. First, if NO is determined in step S102 of FIG. 5, the process proceeds to step S118.
In step S118, the product position coordinate value of the purchased product purchased by the consumer who does not have the comparison product is read from the product position coordinate value DB6. For example, for the product F purchased by the consumer γ shown in FIG. 2, if another consumer has purchased and there is a product compared at the time of purchase, the product position should be obtained. Therefore, the position coordinate value of the product F can be read.
In step S120, the predetermined coefficient used for scaling the contribution vector for each “deemed attribute” in the conjoint analysis is called. The predetermined coefficient used in step S114 is called.
In step S122, for each consumer, a contribution rate for each “deemed attribute” is calculated based on the position coordinate value of the same product as the purchased product and the predetermined coefficient used for scaling in the conjoint analysis. For consumers who do not have a comparison product, there is no data of priority order for purchase product comparison. Therefore, since the conjoint analysis cannot be used like the consumer having the comparative product, the contribution rate is calculated by a different method.
Specifically, as illustrated in FIG. 6, if the purchased product is F, a vector v2 in the same direction as the position coordinate value vector v1 of the product F is set as the consumer contribution rate vector. As described above, it can be said that the coordinate value of the product position indicates the degree of expression of how much the product represents some meaning of the “deemed attribute”. Further, it can be said that the contribution ratio of the consumer indicates the degree to which the “deemed attribute” is emphasized by the consumer at the time of purchase. It is considered that “deemed attributes” tend to be emphasized by the consumer if the degree of expression of the product is large, and not tend to be emphasized by the consumer if the degree of expression of the product is small. Therefore, when the consumer i purchases the product j, it can be said that the position coordinate value vector of the product j and the contribution rate vector of the consumer i tend to be similar. Here, it is assumed that the position coordinate value vector v1 and the contribution rate vector v2 are very similar, and the directions are the same.
From the above, when the position coordinate value of the product j is (L1j, L2j), the contribution rate (U1j, U2j) of the consumer i is obtained by the following equation.
U1 = a × L1,
U2 = a × L2
Here, a is the predetermined coefficient called in step S118.

In step S124 of FIG. 5, the contribution rate calculated as described above is stored in the contribution rate DB 10 of FIG. A data configuration example of the contribution rate DB 10 is shown in FIG.
In the normal conjoint analysis, the contribution rate is a value indicating the degree of importance of the consumer to the attribute. For example, it can be seen that, among the three “deemed attributes” of the first attribute, the second attribute, and the third attribute, the consumer α tends to place the highest importance on the third attribute when purchasing a product.

By the processing so far, the coordinate value of the position on the product map of each product was obtained, and the contribution rate was obtained for each consumer. The product share calculation means 12 of FIG. 1 calculates the share for each product using these coordinate values and the contribution rate.
In the first embodiment, the formula shown in the A column in the merchandise share calculating means 12 is used. The procedure of the share calculation process of the first embodiment executed by the product share calculation means 12 will be described with reference to the flowchart of FIG.
First, in step S202, the product position coordinate value for each product is read from the product position coordinate value DB 6, and in step S204, the contribution rate for each "deemed attribute" for each consumer is read from the contribution rate DB 10.

In step S206, one consumer is selected as consumer i, and in step S208, one product is selected as product j.
In step S210, the probability that the consumer i selects the product j from all the product groups is calculated. Here, different calculation methods are used for a consumer who has a product compared at the time of product purchase and a consumer who does not have a product compared at the time of product purchase. That is, as shown in the column A in the product share calculating means 12 of FIG. 1, the share is calculated by the equation (1) for the consumer having the comparison product, and the equation (2) for the consumer not having the comparison product. Calculate the share by
First, a case where the consumer i is a consumer who has a comparison product will be described. In this case, the share can be calculated by the following equation (1).

here,
L1j: the coordinate value of the first axis of the position of the product j,
L2j: the coordinate value of the second axis of the position of the product j,
...
L1n: the coordinate value of the first axis of the position of the product n,
L2n: coordinate value of the second axis of the position of the product n,
...
U1i: Contribution rate of the first attribute of consumer i
U2i: Contribution rate of the second attribute of consumer i
...
It is.

As described above, it can be said that the coordinate value of the product position indicates the degree of expression of how much the product represents some meaning of the “deemed attribute”. In addition, it can be said that the contribution rate indicates the degree to which the “deemed attribute” is emphasized by the consumer at the time of purchase. Therefore, by multiplying the position coordinate value of the first axis of the product j by the contribution rate of the first attribute of the consumer i, it is estimated how much the consumer i evaluates with respect to the first attribute of the product j. be able to. That is, in the above formula, for example, (L1j × U1i + L2j × U2i +...) Can be said to be a value representing a comprehensive evaluation of how much the consumer i evaluates the product j.
From this, it is possible to know the probability that the consumer i selects a certain product j from all the product groups if the comprehensive evaluation value for the product j of the consumer i is obtained and the ratio of the exponential is obtained. .

In the above formula, for example, (L1j × U1i + L2j × U2i +...) Can be transformed into (L1j, L2j,...) · (U1i, U2i,...). As shown in FIG. 9, this is equivalent to calculating the inner product of the coordinate value vector v3 of the product position and the contribution vector v4 of the consumer's purchased product. By calculating the inner product, the degree of similarity between the coordinate value vector v3 and the contribution vector v4 can be determined.
As described above, it is considered that “deemed attributes” tend to be emphasized by consumers if the degree of representation of the product is large, and do not tend to be emphasized by consumers if the degree of representation of the product is small. Therefore, when the consumer i purchases the product j, it is considered that the degree expressed by the product j and the degree to which the consumer i emphasizes are approximate, and the contribution of the coordinate value vector of the product j and the purchase product j of the consumer i It can be said that the rate vectors tend to be similar.
From this, the inner product is calculated, the similarity between the coordinate value vector of the product j and the contribution rate vector of the purchase product of the consumer i is obtained, and the ratio of the exponential is obtained. The probability of selecting a certain product j from the product group can be known.

Next, a case where the consumer i is a consumer who does not have a comparison product will be described. In this case, the share can be calculated by the following equation (2).

here,
d: A value that is inversely proportional to the distance between the position coordinates of the product j and the position coordinates of the consumer i's purchased product.

d is a value that decreases the probability of selecting the product j as the distance between the position coordinates of the product j and the position coordinates of the purchased product of the consumer i increases. For example, the following values can be used as d.
Example 1: Reciprocal of power of t (distance between position coordinates of product j and consumer i purchased product). t can be selected as appropriate, such as 1 or 2, for example.
Example 2: The power of t (the distance between the position coordinates of a product j and the position coordinates of a consumer i purchased product). t can be appropriately selected, for example, exp, 2 or the like.
The reason for multiplying by d is to correct an error caused by the contribution rate of a consumer who does not have a comparison product. The contribution rate of consumers without comparative products was not calculated by conjoint analysis. Since the contribution rate is a value obtained by multiplying the position coordinate value of the same product as the purchased product by a predetermined coefficient, an error is likely to occur if the share is calculated using this contribution rate. Therefore, the error can be corrected by multiplying by d.

When the calculation is performed as described above, in step S212 in FIG. 8, it is determined whether or not all the products (the products A to H in this embodiment) are selected as the products j. If NO is determined, the next product is selected as the product j in step S214, and the calculation in step S210 is performed for the product j. When the calculation in step S210 is completed for all the products, YES is determined in step S212, and the process proceeds to step S216.
In step S216, it is determined whether or not all consumers have been selected as consumer i. If it is determined NO, the next consumer is selected as the consumer i in step S218, and the processing from step S210 to step S214 is executed. When the process is executed for all consumers, YES is determined in step S216.
FIG. 10 shows the result of the processing so far. For example, the probability that the consumer α selects the product A is calculated as 0.5, the probability that the product B is selected is 0.14, and the probability that the product C is selected is 0.05. For each product, the total probability of all the consumers is calculated, which is 15.4 for product A, 7.6 for product B, 11.7 for product C, and the like.

  Finally, in step S220 of FIG. 8, a share is calculated for each product based on the probability of selecting each product for each consumer. The share for each product can be calculated by the following formula.

From the above formula, for example, the share of the product A is calculated as 15.4 ÷ (15.4 + 7.6 + 11.7 + 8.6 + 9.2 + 23.5 + 13.3 + 10.7) = 15.4%.
The share for each product calculated as described above is stored in the product share DB 14 in FIG.

Returning to FIG. The existing product specification / price DB 16 in FIG. 1 stores details of specifications and prices of existing products.
FIG. 11 shows a data configuration example of the existing product specification / price DB 16. For example, the specification a (displacement) of the product A is 1.8L, the specification b (driving method) is FF, the specification c (full length) is 4190 mm, it is an in-house product, and the price is 1.8 million yen. This is shown. For the following calculation, it is preferable to substitute numerical values in advance such as 1 for FF, 2 for FR, and 3 for 4WD for specifications that are not expressed numerically as in the specification b.

The product position-specification / price relation calculating means 18 in FIG. 1 uses the product position coordinate value DB 6 and the data of the existing product specification / price DB 16 to express the relationship between the product position on the product map and the specification / price. Is calculated. That is, a function is calculated that represents the relationship between what position and position on the product map the specification and price will be.
Specifically, a regression analysis is performed with the coordinate value of the position on the product map as an explanatory variable and the specification / price as an objective variable, and a function is calculated from the result of the analysis. The distribution of product positions on the product map can be expressed with higher accuracy not only by analysis by linear regression but also by analysis by curve regression as shown in FIGS.
The explanatory variable for curve regression can be generally expressed as Z1 ⁿ Z2 ^m . A constant term C that does not depend on Z1 and Z2 is also used. Here, Z1 is a coordinate value for the first axis of the product position, and Z2 is a coordinate value for the second axis of the product position. If it is determined how many times (n + m) is to be raised, it is possible to determine an explanatory variable on condition that 0 <n + m ≦ nm. For example, when it is determined to be up to the square, 0 or a positive integer n and m satisfying 0 <n + m ≦ 2 are (2, 0), (0, 2), (1, 1), (1, 0), (0, 1), and the explanatory variables are Z1 ² , Z2 ² , Z1Z2, Z1, and Z2. A specific example of the explanatory variable determined in this way is shown in FIG. In FIG. 12, the objective variable is “Y total length”, and the regression equation for calculating “Y total length” is
Y total length = α1 × Z1 ² + α2 × Z2 ² + α3 × Z1Z2 + α4 × Z1 + α5 × Z2 + C Total length is calculated. α1 to α5 are coefficients for the explanatory variable when the objective variable is “Y full length”, and the C full length is a constant term when the objective variable is “Y full length”.
FIG. 13 shows a regression line drawn from this regression equation on the product map. For example, it can be seen that the products A, B, and D belong to a category with a total length of 4100 mm, and the products C and G belong to a category with a total length of 4300 mm.
FIG. 14 shows a regression line drawn from the following regression equation on the product map, with the objective variable being “Y minimum ground clearance”.
Y minimum ground clearance = β1 × Z1 ² + β2 × Z2 ² + β3 × Z1Z2 + β4 × Z1 + β5 × Z2 + C minimum ground clearance β1 to β5 are coefficients for explanatory variables when the objective variable is “Y minimum ground clearance”. High is a constant term when the objective variable is “Y lowest ground height”.
In this way, regression equations (functions) representing relationships with positions on the product map are calculated for all specifications / prices stored in the existing product specification / price DB 16.

  In addition, regarding the regression equation described above, the determination of how many times (n + m) can be made can be performed as follows. First, regression analysis is performed for each value while increasing the values of N and M where 0 <n + m ≦ NM (N and M are integers). When the determination coefficient or correlation coefficient at each value does not increase more than a certain value with respect to the value before increasing N and M, the values of N and M at that time are determined as the values of n and m. To do. Since the determination coefficient and the correlation coefficient indicate the degree to which the regression is compatible with the distribution, if the determination coefficient and the correlation coefficient are determined by this method, a regression equation that fits well with the distribution can be obtained.

  The obtained regression equation is stored in the product position-specification / price relationship function DB 20 of FIG. 1 as a product position-specification / price relationship function.

As described above, when the method for calculating the product share and the function representing the relationship between the product position-specification / price are obtained, it is possible to execute the process of calculating the share of the planned new product candidate.
This process is executed by the new product candidate position determining unit 24 and the product share calculating unit 12 shown in FIG. The new product candidate position determining means 24 is based on the specification / price of the new product candidate stored in the new product candidate specification / price DB 22 and is stored in the product position-specification / price relation function DB 20. The position on the product map of the new product candidate is determined using the price relation function. The determined position of the new product candidate is temporarily stored in the product position coordinate value DB 6 and then input to the product share calculating means 12. The product share calculation means 12 calculates the share of the new product candidate using the position coordinates of the new product candidate.
By the time this process is executed, the specifications / prices of the planned new product candidates are provisionally determined and stored in the new product candidate specifications / price DB 22.

First, the process executed by the new product candidate position determining unit 24 will be described. The detailed procedure of this processing is shown in the flowchart of FIG.
First, in step S302, the product position-specification / price relation function for all specifications / prices is read from the product position-specification / price relational function DB 20.
Next, in step S304, position coordinates of a new product candidate as an initial value are set on the product map. The initial value may be determined in advance or may be set by a random number generation method. When the new product candidate is a model change, the position coordinates of the product before the model change may be set. For example, in FIGS. 16 and 17, coordinates indicated by asterisks are set as initial values.

In step S306 in FIG. 15, the specification / price is calculated based on the set coordinates using the product position-specification / price relational function. For example, in FIG. 16, the set coordinates are calculated as “total length 4370 mm”, and in FIG. 17, the set coordinates are calculated as “minimum ground clearance 155 mm”.
When the specification / price of the new product candidate is read from the new product candidate specification / price DB 22 in step S308 of FIG. 15, the specification / price “Yi” calculated from the set coordinates in step S310 and the new product candidate The difference ΔYi between the specification / price “Y new i” is calculated for each specification / price. ΔYi can be calculated by the following equation.
ΔYi = γ ² i × | Y new i−Yi | × αi
Here, γ ² i is a determination coefficient or a correlation coefficient for the specification i. By multiplying by γ ² i, the degree of contribution for calculating the position of a new product candidate can be increased as the specification or price has a higher correlation with the position coordinates.
Αi is a coefficient for standardizing the difference in magnitude of ΔYi caused by the difference in units of Yi. For example, in | Y new i-Yi |, the difference between the total length of 100 mm and the displacement of 100 cc is the same numerical value, but the total length and the unit of displacement are different. Y new i-Yi | and the displacement | Y new i-Yi | are not necessarily different from each other by the same amount. Therefore, the difference in the magnitude of ΔYi is normalized by multiplying by αi. αi can be determined in advance. Alternatively, αi can be determined such that a minimum level and a maximum level are set for each specification / price and the difference between the minimum level and the maximum level is multiplied by αi. For example, the minimum level of the total length is set to 1500 mm and the maximum level is set to 1800 mm. The total length of α is set so that the difference of 300 mm is 1 when multiplied by the total length of α. It can be set to 2000 cc, and the α displacement can be determined so that the difference of 100 mm is 1 when multiplied by the α displacement.
For example, if the “full length 4370 mm” is calculated from the coordinates set in FIG. 16 and the new product candidate is “full length 4430 mm”, | Y new i−Yi | By multiplying the number γ ² i by αi, ΔYi can be obtained.
Subsequently, in step S312 of FIG. 15, ΔYi calculated for each specification / price is totaled for all specifications / price to calculate ΔY.
In step S314, it is determined whether ΔY is equal to or less than a specified value. Here, it may be determined whether or not ΔY has become the minimum value. If YES is determined, the specification / price approximate to the specification / price of the new product candidate can be derived from the position of the set coordinates. On the other hand, if the determination is NO, the specification / price approximate to the specification / price of the new product candidate has not been derived, and the process proceeds to step S316 to reset the coordinates. This is indicated by an arrow m1 in FIGS.
Returning to step S306 in FIG. 15, the series of processing up to step S314 is repeated, and in step S316, coordinates are variously reset as indicated by a plurality of arrows in FIGS. 16 and 17. Finally, YES is determined in step S314. Then, the position coordinates of the new product candidate are determined.
In this case, the process proceeds to step S318, and the coordinates set at this time are added to the product position coordinate DB shown in FIG.

When the position coordinates of the new product candidate are determined as described above, the product share calculation unit 12 in FIG. 1 calculates the product share after the new product candidate is input. The merchandise share calculation process is executed according to the procedure shown in the flowchart of FIG.
The product share after introducing the new product candidate is stored in the product share DB 14 of FIG. In this case, the product share before introducing a new product candidate may be overwritten or coexisted.

As described above, in this embodiment, a product map is created by the multidimensional scaling method using the purchased product comparison priority order. Regardless of product specifications, etc., a product map is created by directly using the data that represents the degree of similarity that consumers perceive to the product. A product map can be created.
In this example, the product map created in this way is used to calculate the product position-specification / price relation function, thereby determining the position of the new product candidate and the product after the new product candidate has started to be sold. Predict the share of each.
Since the share is predicted based only on the order of priority for comparing the purchased products, which represents the degree of similarity of products that consumers feel, the labor and cost of data collection can be reduced, and the degree of similarity of products as seen by consumers A well-matched share can be calculated with high accuracy.

(Second embodiment)
In the second embodiment, the merchandise share calculating means 12 in FIG. 1 uses the formula described in the B column. That is, the share is calculated not only for the consumer who does not have the comparison product but also for the consumer who has the comparison product by the formula (2).
This is due to the following reason. For example, as illustrated in FIG. 19, when the purchased product of the consumer i is C and the first comparative product is G, the contribution rate vector of the consumer i may be v5. v5 has a direction close to the position coordinate value vector of the product H. For this reason, if it calculates with Formula (1), the probability that the consumer i will select the goods H will become larger than the probability which selects the goods C (or goods G). In order to prevent such a phenomenon, the distance between the position coordinates of the purchased product (or the position coordinates of the comparison product) and the position coordinate of the target product j is taken into consideration, and the equation (2) reduces the probability of selection as the distance increases. Is used.

here,
d is a value that is inversely proportional to the distance between the position coordinates of the product j and the position coordinates of the purchased product of the consumer i (or the position coordinates of the comparison product).
When there are a plurality of comparison products, the distance between each comparison product and the position coordinates is obtained, and the maximum distance is used.

  As described above, according to the present embodiment, it is possible to correct an error that may occur when calculating a product share using Equation (1), and it is possible to further improve the prediction accuracy.

(Third Example) In the first example and the second example, the share of a new product candidate could be calculated.
In the third embodiment, the product share calculating means 12 for calculating the share of new product candidates and the product position-specification / price relation function are utilized for other processing. For example, (1) it is used for the process of determining the specifications / price of a new product on the condition that the share of the new product is maximized. Also, (2) it is used for processing to determine the specifications and prices of new products on the condition that the shares of specific existing products are not reduced and the new products are desired. Alternatively, (3) it is used for the process of determining the specifications and prices of new products on the condition that the share of the existing product group of other companies is minimized. In addition, it is used for the process of determining the specifications and prices of new products on the condition that (4) the total share of the entire company products is maximized. In addition, (5) When planning multiple new products, including new products whose existing products are to be remodeled, maximize the total share of the entire company's products and reduce the share of new products to be remodeled from existing products. In addition, it is used for processing to determine the specifications and prices of each of a plurality of new products on condition that the share of other new products is maximized.
For these processes, the new product position determination unit 26 in FIG. 1 utilizes the product share calculation unit 12 and (1) the position on the product map of the new product on the condition that the share of the new product is maximized. To decide. In addition, (2) the position on the product map of the new product is determined on the condition that the share of the specific existing product is not reduced and the share of the new product is desired. Alternatively, (3) the position on the product map of the new product is determined under the condition that the total share of the existing product group of the other company is minimized. In addition, (4) the position on the product map of the new product is determined under the condition that the total share of the entire company product is maximized. In addition, (5) When planning multiple new products, including new products whose existing products are to be remodeled, maximize the total share of the entire company's products and reduce the share of new products to be remodeled from existing products. In addition, the position of each new product on the product map is determined on condition that the share of other new products is maximized.
The new product specification / price determination means 28 determines the specification / price of the new product using the product position-specification / price relation function based on the position of the new product on the determined product map.

(1) First, with reference to the flowchart of FIG. 18, the procedure of the processing executed by the product position determination unit 26 and the new product specification / price determination unit 28 under the condition that the share of the new product is maximized will be described. To do.
First, in step S402, the product position determining means 26 sets the coordinates of a new product as an initial value on the product map. The initial value may be determined in advance or set by a random number generation method. In addition, if the new product is a model change product, the position coordinates of the product before the model change may be set, or if the new product is a completely new product, the position coordinates of the competing existing product may be set. . The set coordinates are temporarily stored in the product position coordinate value DB 6 once. If the set coordinates are the coordinates before the model change or the coordinates of the competing existing product, the information is also stored together.
Next, in step S404, the share of the new product is calculated using the product share calculation means 12. Specifically, the position coordinates set above are used as the product position coordinates in step S202 of FIG. 8, and a series of processing from step S204 to step S220 is executed to calculate the share of the new product (here, Step S212 and step S214 are not executed).
In step S406 of FIG. 18, the product position determination unit 26 determines whether the share calculated by the product share calculation unit 12 is the maximum. When determined as YES, the maximum share can be derived from the set coordinates. On the other hand, if it is determined NO, the maximum share cannot be derived, and the process proceeds to step S408 to reset the coordinates. This is executed in the same manner as the arrow m1 in FIGS.
Returning to step S404 in FIG. 18, after passing through step S406, in step S408, various coordinates are reset as shown by the plurality of arrows in FIG. 16 and FIG. The position coordinates of the new product that maximizes is determined.
In this case, the process proceeds to step S410, and the coordinates set at this time are stored in the product position coordinate DB 6 as the coordinates of the finally determined new product.

Next, in S412 of FIG. 18, the new product specification / price determination means 28 reads the product position-specification / price relationship function for all specifications / prices from the product position-specification / price relationship function DB 20.
In step S414, the new product specification / price determination means 28 calculates the specification / price of the new product based on the coordinates set above using the product position-specification / price relation function.
Finally, in step S416, the calculated new product specification / price is stored in the new product specification / price DB 30.
As described above, it is possible to determine the specifications / price of a new product with the largest share.

(2) In addition, when determining the specifications / price of a new product under the condition that the share of a new product is desired without dropping the share of a specific existing product, the processing in step S406 in FIG. 18 is different.
Under this condition, in step S406, it is determined whether or not the difference between the share calculated in step S404 and the desired share is minimum. If it is determined YES, a share that approximates the desired share can be derived from the set coordinates. On the other hand, if it is determined as NO, since a share that approximates the desired share cannot be derived, the process proceeds to step S408 and the process of resetting the coordinates is executed. If the coordinates are variously reset and finally determined to be YES in step S406, the position coordinates of the new product closest to the desired share are determined.
As described above, when it is not desired to reduce the share of a specific existing product, it is possible to determine the specification / price of a new product to be a desired share.

(3) When determining the specifications and prices of new products under the condition that the total share of existing product groups of other companies is minimized, the process is executed according to the procedure shown in the flowchart of FIG. Hereinafter, this process will be described.
First, in step S502 of FIG. 20, the product position determination means 26 sets the coordinates of a new product as an initial value on the product map. The set coordinates are temporarily stored in the product position coordinate value DB 6 once. If the set coordinates are the coordinates before the model change or the coordinates of the competing existing product, the information is also stored.
Next, in step S504, the share for each product is calculated using the product share calculation means 12. In step S506, based on the share for each product, the share of the new product, the total share of the company's existing product group, and the total share of the other company's existing product group can be calculated. For existing products, information on the company / other companies is stored in the existing product specification / price DB 16, and by using this information, each share can be calculated. Each share calculated here is the “first” share.
The subsequent step S508 is not executed in the first process, and the process proceeds to step S510. In step S510, the coordinates of the new product are reset as in the case of the arrow m1 in FIGS.

Returning to step S504 in FIG. 20, the share for each product is calculated. In step S506, the share of the “second” new product, the total share of the company's existing product group, and the total share of the existing product group of other companies are calculated. Is calculated. In step S508, the new position determination unit 26 calculates a difference Δ between the share calculated at the second time and the share calculated at the first time as follows.
Δ (Share of new products)
= Share of new product for the second time Share of new product for the first time Δ (Total share of the company's existing products)
= Total share of existing product group of the company for the second time-Total share of existing product group of the company for the first time Δ (Total share of existing product group of other companies)
= Total share of existing product group of the other competitor for the second time-Total share of existing product group of the other company of the second time Further, the following calculation is performed.
Δ (Total share of all products)
= Δ (share of new products) + Δ (total share of the company's existing products)
In this step, it is determined whether “Δ (total share of existing product group of other company) <0 and total share of existing product group of other company is minimum”. Here, if “Δ (total share of existing products of other companies)> 0”, even if “Δ (share of new products)> 0” (that is, the share of the second new product is Even if the share of the new product is increased for the first time, the share of all of the company's products is decreasing due to the coordinate setting of the second new product. In such a case, increasing the market share of new products will not work for the entire company's products. Therefore, “Δ (total share of existing product group of other company) <0 and the total share of existing product group of the other company is minimum” is set as a criterion.
It should be noted that at the stage of the “second time-first time” processing, since other “total share of existing product groups of other companies” for comparing whether or not “minimum” is obtained is not obtained, the process proceeds to step S510. . In step S510, a series of processing from step S504 to step S510 is repeated a predetermined number of times or more while the coordinates are variously set as indicated by a plurality of arrows in FIGS. In step S508, a difference Δ between the share calculated each time and the share calculated immediately before is calculated.
If “YES” is determined in the step S508, the coordinates of the new product are set so as to minimize the total share of the existing product group of the other company. On the other hand, if the determination is NO, the coordinates that minimize the total share of the existing product group of the other company are not set, so the process proceeds to step S510, as indicated by the arrow m1 in FIGS. The coordinates are reset, and a series of processing from step S504 to step S510 is repeated.
If YES is finally determined in step S508, the coordinates of the new product that minimizes the total share of the existing product group of the other company are determined.
In this case, the process proceeds to step S512, and the coordinates of the new product set at this time are stored in the product position coordinate DB 6 shown in FIG.

Next, in S514 of FIG. 20, the new product specification / price determination means 28 reads the product position-specification / price relationship function for all specifications / prices from the product position-specification / price relationship function DB 20.
In step S516, the new product specification / price determination means 28 calculates the specification / price of the new product based on the coordinates of the new product set as described above, using the product position-specification / price relation function.
Finally, in step S518, the calculated specification / price of the new product is stored in the new product specification / price DB 30.
By the above, it is possible to determine the specifications and prices of new products that minimize the share of existing product groups of other companies.

(4) When determining the specifications / price of a new product under the condition that the total share of the entire company product is maximized, the process is executed according to the procedure shown in the flowchart of FIG. Hereinafter, this process will be described.
First, in step S602 of FIG. 21, the product position determination means 26 sets the coordinates of a new product as an initial value on the product map.
Next, in step S604, the share for each product is calculated using the product share calculation means 12. In step S606, based on the share for each product, the share of the new product, the total share of the existing product group of the company, and the total share of the existing product group of the other company can be calculated. Each share calculated here is the “first” share.
The subsequent step S608 is not executed in the first process, and the process proceeds to step S610. In step S610, the coordinates of the new product are reset as in the case of the arrow m1 in FIGS.

Returning to step S604 of FIG. 21, the share for each product is calculated. In step S606, the share of the “second” new product, the total share of the company's existing product group, and the total share of the existing product group of other companies are calculated. Is calculated. In step S608, the new position determining unit 26 calculates a difference Δ between the share calculated at the second time and the share calculated at the first time as follows.
Δ (Share of new products)
= Share of new product for the second time Share of new product for the first time Δ (Total share of the company's existing products)
= Total share of existing product group of the company for the second time-Total share of existing product group of the company for the first time Δ (Total share of existing product group of other companies)
= Total share of existing product group of the other competitor for the second time-Total share of existing product group of the other company of the second time Further, the following calculation is performed.
Δ (Total share of all products)
= Δ (share of new products) + Δ (total share of the company's existing products)
In this step, it is determined whether “Δ (total share of all products of the company) ≧ 0 and the total share of all products of the company is maximum”.
Note that, at the stage of the “second to first” process, since “the total share of all the products of the company” for comparing whether or not it is “maximum” has not been obtained, the process proceeds to step S610. In step S610, a series of processing from step S604 to step S610 is repeated a predetermined number of times or more while the coordinates are variously reset as indicated by a plurality of arrows in FIGS. In step S608, a difference Δ between the share calculated each time and the share calculated immediately before is calculated.
If YES is determined in step S608, the coordinates of a new product that maximizes the total share of all the products of the company are set. On the other hand, if NO is determined, coordinates that maximize the total share of all the products of the company are not set, so the process proceeds to step S610 and coordinates as indicated by an arrow m1 in FIGS. Is set again, and a series of processing from step S604 to step S610 is repeated.
If it is finally determined YES in step S608, the coordinates of the new product that maximizes the total share of all the products of the company are determined.
In this case, the process proceeds to step S612, and the coordinates of the new product set at this time are stored in the product position coordinate DB 6 shown in FIG.

Next, in S614 of FIG. 21, the new product specification / price determination means 28 reads product position-specification / price relationship functions for all specifications / prices from the product position-specification / price relationship function DB 20.
In step S616, the new product specification / price determination means 28 calculates the specification / price of the new product based on the coordinates of the new product set as described above, using the product position-specification / price relation function.
Finally, in step S618, the calculated specification / price of the new product is stored in the new product specification / price DB 30.
By the above, it is possible to determine the specifications and prices of new products that maximize the share of all the products of the company.

(5) Furthermore, when planning multiple new products, including new products whose model is to be remodeled, maximize the total share of the entire company's products and reduce the share of new products to be remodeled from existing products. Furthermore, when determining the specifications and prices of each new product on condition that the share of other new products is maximized, the processing is executed according to the procedure shown in the flowcharts of FIGS. Here, there are two new products planned, one is a completely new product (hereinafter referred to as “new product”), and the other is an existing product (hereinafter referred to as “ The case of “changed product”) will be described.

First, an outline of this process will be described.
First, in the process of FIG. 22, while changing the coordinate m as the initial value of a brand new product, the total share of all the products of the company is not dropped, and the brand new product maximum share of the brand new product is maximized. Search for the provisionally determined coordinate p.
Subsequently, in the process of FIG. 23, while changing the coordinates of the changed product n, the total share of all the products of the company is maximized, and the maximum share of the brand-new product obtained from p is not decreased, and is obtained from p. The temporary determination coordinate r of the changed product that does not drop the share of the changed product is searched.
Next, in the process of FIG. 24, if the provisionally determined coordinate r of the changed product cannot be searched in the process of FIG. The tentative decision coordinates u of the brand new product that does not drop the maximum share of the brand new product obtained from p and does not drop the share of the changed product obtained from p are searched again. Thereafter, by returning to FIG. 22 or FIG. 23, the temporarily determined coordinate r of the changed product is searched.
Finally, “Maximize the total share of all the company's products, maximize the share of brand-new products, and not drop the share of the changed products before the model change”. Until the search, the processing of FIGS. 22 to 24 is repeatedly executed.
After searching for the coordinates of the brand new product and the coordinates of the changed product by the above processing, the process proceeds to the processing of FIG. Decide each.

Next, details of this processing will be described.
First, in the process of FIG. 22, while changing the coordinate m as the initial value of a brand new product, the total share of all of the company's products is not dropped, and the brand new product's share is maximized. The determined coordinate p is searched.
First, in step S702, the merchandise position determining means 26 sets the coordinates m of the brand new merchandise and the coordinates n of the changed merchandise as initial values on the merchandise map. The coordinate n of the changed product is stored in the product position coordinate value DB 6. The set coordinates m and n are temporarily stored in the product position coordinate value DB 6 once.
Next, in step S704, the share for each product is calculated using the product share calculation means 12. In step S706, based on the share for each product, the share of the brand new product, the total share of the company's existing product group (including the changed product), and the total share of the existing product group of the other company are calculated. Each share calculated here is the “first” share.
Since the subsequent step S708 cannot be executed in the first process, the process proceeds to step S710. In step S710, the coordinates of the brand new product are reset as in the case of the arrow m1 in FIGS.

Returning to step S704 of FIG. 22, the share for each product is calculated, and in step S706, the share of the “second time” brand new product, the total share of the company's existing product group, and the total of the other company's existing product group. Calculate share. In step S708, the new position determination unit 26 calculates a difference Δ between the share calculated at the second time and the share calculated at the first time as follows.
Δ (Share of brand new products)
= Share of brand-new products for the second time -Share of brand-new products for the first time Δ (Total share of the company's existing products)
= Total share of existing product group of the company for the second time-Total share of existing product group of the company for the first time Δ (Total share of existing product group of other companies)
= Total share of existing product group of the other competitor for the second time-Total share of existing product group of the other company of the second time Further, the following calculation is performed.
Δ (Share of all products)
= Δ (Share of brand new products) + Δ (Total share of the company's existing products)
In this step, it is determined whether or not “Δ (total share of all products of the company) ≧ 0 and Δ (share of brand new products)> 0 and the share of brand new products is the maximum”. In other words, the share of the company's products for the second time is greater than the share of all the products of the company for the first time, and the share of the brand-new products for the second time is larger than the share of the brand-new products for the first time. Is determined to be the maximum.
Note that at the stage of the “second time-first time” processing, since another “share of brand-new products” for comparing whether or not the share is maximum is not obtained, the process proceeds to step S710. In step S710, a series of processing from step S704 to step S710 is repeated a predetermined number of times or more while the coordinates are variously reset as indicated by a plurality of arrows in FIGS. In step S708, a difference Δ between the share calculated each time and the share calculated immediately before is calculated.
If it is finally determined as YES in step S708, the share of all the products of the first company is larger than the share of the first new product, and the share that maximizes the share is new. The coordinates of the product are searched (each share calculated in step S706 at this stage is set as the “p-th” share). In step S711, the coordinates of the searched new product are set as the provisionally determined coordinates p.
In the process so far, the provisionally determined coordinate p of the brand new product has been searched, and is thus replaced with the coordinate m stored in the product position coordinate value DB 6.

Subsequently, in the process of FIG. 23, while changing the coordinate n of the changed product, the total share of all the products of the company is maximized, and the share of the brand-new product obtained from p is not decreased, and is obtained from p. The temporary determination coordinate r of the changed product that does not drop the share of the changed product is searched.
First, in step S712, the coordinate n of the changed product is changed to an arbitrary coordinate q. The changed coordinate q is temporarily stored in the product position coordinate value DB 6 once.
In step S714, the share for each product is calculated using the product share calculation means 12. In step S716, based on the share for each product, the share of the brand new product, the total share of the company's existing product group (including the changed product), and the total share of the existing product group of the other company are calculated. Each share calculated here is set as the “q-th” share.

Subsequently, the process proceeds to step S718 in FIG. Here, the new position determination means 26 calculates the difference Δ between the p-th share and the q-th share, and distinguishes between the following three cases.
(a) Δ (total share of all products)> 0 and Δ (share of brand new products) ≧ 0,
And Δ (share of changed product) ≧ 0
(b) Δ (total share of all products of the company)> 0 and Δ (share of brand new products) <0,
And Δ (share of changed product) ≧ 0
(c) Δ (total share of all products of the company)> 0 and Δ (share of changed products) <0,
Or, Δ (total share of our products) ≦ 0
(a) is a criterion for introduction into the process of searching for the provisionally determined coordinate r of the changed product. In (a), “Δ (total share of all products of the company)> 0” means that the total share of all products of the qth company is the pth of all products of the company calculated in step S706 of FIG. This is because it is required to be larger than the total share. In (a), the share of the qth new product is equal to or greater than the share of the pth new product, and the share of the qth changed product is equal to or greater than the share of the pth changed product. Request. When these criteria are cleared, the process proceeds to step S720, and the process of searching for the provisionally determined coordinate r of the changed product starts.
In (b), the criteria of “Δ (total share of all products of the company)> 0” and “Δ (share of changed products) ≧ 0” have been cleared, but the share of brand-new products has dropped. Is the case. In this case, after step S726, the process proceeds to a process (described later) for re-searching the provisionally determined coordinates u of the brand new product in FIG.
In (c), the criteria of “Δ (total share of all products of the company)> 0” has been cleared, but the share of the changed product has dropped or “Δ (total of all products of the company) This is a case where “share” ≦ 0 ”and“ Δ (total share of all products of the company)> 0 ”is not cleared. In this case, the process returns to step S718 through steps S728 and S730, and a series of processing is repeated.
As described above, in FIG. 23, the temporarily determined coordinate r of the changed product is searched for in step S722, and the process of (a), (b), or (c) is finally performed until it is determined YES in step S724. Repeat.

If (a) is determined in step S718, the process proceeds to step S720, and the process of searching for the provisionally determined coordinate r of the changed product is entered. In step S720 and step S722, while maintaining the condition (a), the change of the coordinates of the changed product is repeated until the total share of all the products of the company is maximized. If “YES” is determined in the step S722, the coordinate of the changed product at this time is set as the temporarily determined coordinate r in a step S723.
In the process so far, the provisionally determined coordinate r of the changed product has been searched, and is replaced with the coordinate q stored in the product position coordinate value DB 6.

  Subsequently, in step S724, the “pth” total share of all products of the company calculated from the provisionally determined coordinates p of the brand new product in step S708 of FIG. 22 and the provisionally determined coordinates r of the changed product “ The total share of all the products of the company of “r” is compared, and it is determined whether the total share of all of the products of the “r” is maximum. If the answer is YES, in addition to the result that “the total share of all the products of the company is the largest”, “more than the share of brand-new products obtained at the p-th” and “more than the share of changed products obtained at the p-th” The provisionally determined coordinate p of the brand new product and the provisionally determined coordinate r of the changed product are obtained. That is, “maximize the total share of all the company's products, maximize the share of brand-new products, and do not drop the share of the changed products before the model change”. The determined coordinate r is obtained. In this case, the processing of FIG. On the other hand, if NO is determined, the process returns to step S710 in FIG. 22 to reset the coordinates p of the brand new product until a result that “the total share of all products of the company is maximum” is obtained. The series of processing from step S704 is repeated.

  On the other hand, if (b) is determined in step S718, the process proceeds to step S726, and the coordinate q of the changed product changed in step S712 is returned to the first coordinate n. The returned coordinate n is temporarily stored in the product position coordinate value DB 6 once.

In the process of FIG. 24, the total share of all the products of the company is maximized while resetting the provisionally determined coordinates p of the brand-new product, and the share of the brand-new product obtained from p is not reduced. The temporary determination coordinates u of the brand new product that does not drop the share of the obtained changed product are searched again.
In the first step S732, the provisionally determined coordinate p of the brand new product is reset to t. The reset coordinate t is temporarily stored in the product position coordinate value DB 6 once.
In step S734, the share for each product is calculated using the product share calculation means 12. In step S736, based on the share for each product, the share of the brand new product, the total share of the company's existing product group (including the changed product), and the total share of the existing product group of the other company are calculated. Each share calculated here is set as the “t-th” share.

Subsequently, the process proceeds to step S738. Here, the new position determination means 26 calculates the difference Δ between the p-th share and the t-th share, and distinguishes between the following three cases.
(d) Δ (total share of all products)> 0 and Δ (share of brand new products) ≧ 0,
And Δ (share of changed product) ≧ 0
(e) Δ (total market share of all products)> 0 and Δ (share of brand new products) <0,
And Δ (share of changed product) ≧ 0
(f) Δ (total share of all products of the company)> 0 and Δ (share of changed products) <0,
Or, Δ (total share of our products) ≦ 0
(d) is a criterion for introduction into the process of re-searching the provisionally determined coordinates u of a brand new product. In (d), “Δ (total share of all products of the company)> 0” means that the total share of all products of the t-th company is the p-th all products of the company calculated in step S706 of FIG. This is because it is required to be larger than the total share. In (a), the share of the t-th new product is equal to or greater than the share of the p-th new product, and the share of the t-th new product is greater than the share of the p-th new product. Request. When these criteria are cleared, the process proceeds to step S740 to enter a process for re-searching the provisionally determined coordinates u of the brand new product.
In (e), the criteria of “Δ (total share of all products of the company)> 0” and “Δ (share of changed product) ≧ 0” are cleared, but the share of brand-new products has dropped. Is the case. In this case, the process returns to step S730 in FIG. 23 via step S746, and a series of processes for searching for the provisionally determined coordinate r of the changed product is repeated.
In (f), the criteria of “Δ (total share of all products of the company)> 0” has been cleared, but the share of the changed product has dropped or “Δ (total of all products of the company) This is a case where the standard of “share”> 0 ”is not cleared (“ Δ (total share of all products of the company) ≦ 0 ”). In this case, the process returns to step S738 via steps S748 and S750, and a series of processes is repeated.
As described above, in FIG. 24, the temporarily determined coordinate u of the brand new product is searched for in step S742, and the process of (d), (e), or (f) is finally performed until YES is determined in step S742. Repeat.

If (d) is determined in step S738, the process proceeds to step S740, and the process for re-searching the provisionally determined coordinates u of the brand new product is entered. In step S740 and step S742, while maintaining the condition (d), the setting of the coordinates of the brand new product is repeated until the total share of all the products of the company is maximized. If YES is determined in step S742, in addition to the result that “the total share of all products of the company is the maximum”, “more than the share of brand-new products obtained at the r-th time” and “change obtained at the r-th time” The provisionally determined coordinates of the brand new product and the provisionally determined coordinates r of the changed product, which are “commodity share or more”, are obtained. In this case, in step S743, the coordinates of the brand new product at this time are set as the temporarily determined coordinates u.
Since the provisionally determined coordinate u of the brand new product has been obtained by the processing so far, the coordinate t stored in the product position coordinate value DB 6 is replaced.

Subsequently, in step S744, the “r-th” total share of all the products of the company calculated from the temporarily determined coordinate r of the changed product in step S724 of FIG. The total share of all the products of the company in the “u-th” is compared to determine whether the total share of all the products of the company in the “u-th” is the maximum.
However, at this stage, the total share of all the products of the company “r-th” in step S724 of FIG. 23 is not obtained. The total share of all the products of the “r-th” company can be obtained after step S718 in FIG. 23 is determined as (a), and then steps S720 to S624 are executed, and the process returns to step S710 in FIG. Thereafter, a series of processing is performed from step S704, and it is determined in step S718 that (b) is entered and the processing of FIG. 24 is entered. Therefore, it is determined as NO at this stage, and the process returns to step S712 in FIG. 23 to repeat a series of processes for searching for the provisionally determined coordinate r of the changed product.
After such a series of processing has been performed, if YES can be determined in step S744, in addition to the result that “the total share of all the products of the company is the maximum”, “the new product obtained at the r-th time” The provisionally determined coordinate u of the brand new product and the provisionally determined coordinate r of the changed product, which are “share over” and “over the share of the changed product obtained at the r-th”, are obtained. In other words, “maximize the total share of all of the company's products, maximize the share of brand-new products, and do not drop the share of the changed products before the model change”. The determined coordinate r is obtained. In this case, the processing of FIG.

On the other hand, if (e) is determined in step S738, the process moves to step S746, and the provisionally determined coordinate t of the brand new product reset in step S732 is once returned to p. The returned temporarily determined coordinate p is temporarily stored in the product position coordinate value DB 6.
Subsequently, the process returns to step S730 in FIG. 23, and after changing the coordinate of the changed product to s, a series of processes for searching the provisionally determined coordinate r of the changed product from step S718 is repeated.

As described above, the series of processing of FIGS. 22 to 24 is repeated until YES is determined in step S724 of FIG. 23 or YES is determined in step S742 of FIG.
If it is determined as YES in step S724 of FIG. 23, in addition to the result that “the total share of all the products of the company is the maximum”, “the maximum share of the brand new product obtained in the p-th time (step S708 of FIG. 22)” The provisionally determined coordinate p of the brand new product and the provisionally determined coordinate r of the changed product, which are “above” and “more than the share of the changed product obtained at the p-th”, are obtained.
Alternatively, if it is determined as YES in step S744 in FIG. 24, in addition to the result that “the total share of all the products of the company is the maximum”, “the new product obtained in the r-th time (step S724 in FIG. 23)” The provisionally determined coordinate u of the brand new product and the provisionally determined coordinate r of the changed product, which are “greater than the maximum share” and “more than the share of the changed product obtained at the r-th”, are obtained.
In other words, in both cases, “maximize the total share of all the company's products, maximize the share of brand-new products, and not drop the share of the changed products before the model change”. The provisionally determined coordinates of the product are obtained.
As described above, when the provisionally determined coordinate p of the brand new product and the provisionally determined coordinate r of the changed product, or the provisionally determined coordinate u of the brand new product and the provisionally determined coordinate r of the changed product are obtained, the process proceeds to FIG.

In the process of FIG. 25, the specification / price of the brand new product and the specification / price after the model change of the changed product are determined.
First, in step S752, the provisionally determined coordinate p of the brand new product and the provisionally determined coordinate r of the changed product are stored in the product position coordinate DB 6 as the determined coordinate of the new product and the determined coordinate of the changed product, respectively. Alternatively, the provisionally determined coordinate u of the new product and the provisionally determined coordinate r of the changed product are stored in the product position coordinate DB 6 as the determined coordinate of the new product and the determined coordinate of the changed product, respectively.
Next, in step S754, the new product specification / price determination means 28 reads product position-specification / price relationship functions for all specifications / prices from the product position-specification / price relationship function DB 20.
In step S756, the new product specification / price determination means 28 uses the product position-specification / price relation function to determine the specification / price of the new product and the model change of the changed product based on the coordinates stored in step S752. Calculate specifications and prices.
Finally, in step S758, the calculated specification / price of the new product and the specification / price after the model change of the changed product are stored in the new product specification / price DB 30.
As described above, it is possible to determine the specifications and prices of new products that maximize the total share of all of the company's products, maximize the share of brand-new products, and do not drop the share of the changed products before the model change. Thereby, it is possible to make a lineup of products that are advantageous as a whole of the company's products.

  As described above, according to the present embodiment, the specification of the new product that becomes the desired share by utilizing the product share calculating means for calculating the share of the new product candidate and the product position-specification / price relation function.・ Price can be determined.

Specific examples of the present invention have been described in detail above, but these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and changes of the specific examples illustrated above.
In addition, the technical elements described in the present specification or the drawings exhibit technical usefulness alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. In addition, the technology illustrated in the present specification or the drawings achieves a plurality of objects at the same time, and has technical utility by achieving one of the objects.

The figure which illustrates a share prediction apparatus. The figure which shows the data structural example of purchase merchandise comparison priority order DB2. The figure which shows the goods map by a multidimensional scaling method. The figure which shows the coordinate value of the position of each goods on a goods map. The flowchart which shows the procedure of the contribution rate calculation process for every "deemed attribute" for every consumer. The figure explaining the calculation method of the contribution rate for every "deemed attribute" of the consumer who does not have a comparison product. The figure which shows the contribution rate for every "deemed attribute" for every consumer. The flowchart which shows the procedure of a share calculation process. The figure explaining the relationship between a goods position coordinate value vector and a contribution rate vector. The figure which shows the probability that a consumer will select goods. The figure which shows the data structural example of the existing goods specification and price DB16. The figure which illustrates the function which shows the relationship between a position coordinate value and a level. The figure which illustrates the relationship between a position coordinate value and a level. The figure which illustrates the relationship between a position coordinate value and a level. The flowchart which shows the procedure of a new goods candidate position determination process. The figure explaining new product candidate position determination processing. The figure explaining new product candidate position determination processing. The flowchart which shows the procedure of the specification / price determination process of the new product which implement | achieves the desired share. The figure explaining the problem of the contribution rate vector for every "deemed attribute" of the consumer who has a comparison product. The flowchart which shows the procedure of the specification / price determination process of the new product of the new product which minimizes the total share of the existing product group of the other company. The flowchart which shows the procedure of the specification / price determination process of the new product of the new product which minimizes the total share of the existing product group of the other company. The procedure for determining the specification / price determination process of the changed product / brand new product that maximizes the total share of all of the company's products and does not drop the share of the modified product from the existing products and maximizes the share of the brand new product. flowchart. The procedure for determining the specification / price determination process of the changed product / brand new product that maximizes the total share of all of the company's products and does not drop the share of the modified product from the existing products and maximizes the share of the brand new product. flowchart. The procedure for determining the specification / price determination process of the changed product / brand new product that maximizes the total share of all of the company's products and does not drop the share of the modified product from the existing products and maximizes the share of the brand new product. flowchart. The procedure for determining the specification / price determination process of the changed product / brand new product that maximizes the total share of all of the company's products and does not drop the share of the modified product from the existing products and maximizes the share of the brand new product. flowchart.

Explanation of symbols

2: Purchased product comparison priority order DB,
4: Product map creation means,
6: Product position coordinate value DB,
8: Contribution rate calculation means,
10: Contribution rate DB,
12: Product share calculation means,
14: Product share DB
16: Existing product specifications / price DB,
18: Product position-specification / price relation function calculating means,
20: Product position-specification / price relational function DB,
22: New product candidate specification / price DB,
24: New product candidate position determination means,
26: New product position determination means,
28: New product specification / price determination means,
30: New product specifications / price DB

Claims (2)

The ratio of the number of sales for each product to the total number of sales of the competitive product group, that is, a device for predicting the share,
For each consumer, a “consumer / purchased existing product / compared existing product / priority order” database that stores the purchased existing product, the existing product compared with the purchased existing product, and the priority order in association with each other;
“Comparison existing product and its priority order” stored in the “consumer / purchased existing product / comparison existing product / priority order” database by the computer as “existing product similar to purchase existing product and its similar order” A product map creation means for creating a product map of an existing product by a dimensional scaling method;
For each existing product, an “existing product / attribute / level” database storing the level for each attribute,
By computer, regression analysis using the coordinates on the product map for each existing product as explanatory variables and the level for each attribute of each existing product stored in the “existing product / attribute / level” database as the objective variable, A calculation means for calculating a “coordinate value-level” function that is a regression equation with the coordinate value on the product map as an explanatory variable and the level as an objective variable,
Product profile creation means for creating a product profile that associates the axis of the product map as an attribute of the product for each existing product and associates the coordinate value related to the axis as the level of the attribute,
When the product profile is selected from the product profile group by the computer, the existing purchase product, the comparison existing product and the priority order stored in the “consumer / existing product / comparison existing product / priority order” database. A contribution rate calculating means for calculating a contribution rate for each "product map axis associated as a product attribute" for each consumer by conjoint analysis as a selection order;
A “new product / attribute / level” database storing levels for each attribute of new products;
As an explanatory variable when the level for each attribute of the new product stored in the “new product / attribute / level” database is set as an objective variable of the “coordinate value-level” function by the computer, it is displayed on the product map of the new product. New product position coordinate value calculating means for calculating coordinate values;
The product multiplied by the contribution rate for each “product map axis associated as a product attribute” for each consumer and the coordinate value on the product map for each product, As the probability of selecting the value of the exponential function based on the number of navy of the total value in the `` Product Map Axis Associated as '' from the competitive product group after the start of new product sales for each consumer, for each product A product share calculating means for calculating and calculating a value obtained by summing up the selection probabilities calculated for each product for all consumers as a share for each product;
Share prediction device having The ratio of the number of sales for each product in the total number of sales of the competitive product group, that is, a program for predicting the share, the computer performs the following processing:
For each consumer, “consumer / purchased existing product / compared existing product / priority order” stored in the computer in association with the existing product purchased, the existing product compared with the purchased existing product, and the priority order. Process to access the database,
Multidimensional scaling method with “Comparison existing product and its priority order” stored in the “Consumer / Purchasing existing product / Comparison existing product / Priority order” database as “Existing product similar to purchase existing product and its similarity order” To create a product map for existing products,
For each existing product, a process of accessing an “existing product / attribute / level” database in which the level for each attribute is stored in a computer-readable manner;
For each attribute, a regression analysis using the coordinate value on the product map for each existing product as the explanatory variable and the level for each attribute for each existing product stored in the “existing product / attribute / level” database as the objective variable. A process of calculating a "coordinate value-level" function that is a regression equation with the coordinate values on the product map as explanatory variables and the level as an objective variable;
For each existing product, associating the axis of the product map as an attribute of the product and creating a product profile that associates the coordinate value related to that axis as the level of the attribute,
The “Purchasing Existing Product / Comparison Existing Product / Priority Order” stored in the “Consumer / Purchasing Existing Product / Comparison Existing Product / Priority Order” database, the priority order, and the selection order when the product profile is selected from the product profile group. With the conjoint analysis, for each consumer, a process of calculating a contribution rate for each “product map axis associated as a product attribute”
A process of accessing a “new product / attribute / level” database in which the level for each attribute of the new product is stored in a computer-readable manner;
As an explanatory variable when the level of each new product attribute stored in the “new product / attribute / level” database is the objective variable of the “coordinate value-level” function, the coordinate value on the product map of the new product is used. Processing to calculate,
For each product, “Association as product attributes” is calculated by multiplying the contribution rate for each “product map axis associated as product attributes” for each consumer and the coordinate value on the product map for each product. The value of the exponential function based on the number of navy of the total value in the `` Product Product Map Axis '' is calculated for each product as the probability of selecting from the competitive product group after the start of new product sales for each consumer, A process of calculating the sum of the selection probabilities calculated for each product for all consumers as a share for each product,
A program for share prediction that makes you run.

Images