WO2011043329A1

WO2011043329A1 - Product purchase assistance system

Info

Publication number: WO2011043329A1
Application number: PCT/JP2010/067437
Authority: WO
Inventors: 重樹上田平
Original assignee: インターマン株式会社
Priority date: 2009-10-08
Filing date: 2010-10-05
Publication date: 2011-04-14

Abstract

A plurality of filter items are displayed on a product recommendation screen of a client computer, and when a user selects one filter item, the past purchase history database is referenced, and according to the filter items selected by the user, based on the history, products with a high probability of being purchased from the history are preferentially displayed. Then, when another filter item is selected, the purchase history database is referenced, and additionally the other filter item is taken into consideration, and products with a high probability of being purchased are preferentially displayed. In addition, by associating and simultaneously displaying products immediately preceding and following, in probability of being purchased, the product that has been preferentially displayed, the probability of purchase increases. As a result, the user can complete purchase of a necessary product in a short amount of time, and the product retailer side can decrease the probability of losing an opportunity of selling the product.

Description

Product purchase assistance system

The present invention relates to a product purchase assistance system, and more particularly, to a product purchase assistance system capable of promptly purchasing a necessary product through an intuitive operation.

In recent years, sales of mail order sales have increased significantly. In particular, the Internet shopping market has reached several trillion yen and continues to expand rapidly. In fact, it is very convenient to deliver to your work or home after a few days with just a few clicks on your computer screen. From the point of view of sales, there is no need for display and there are few restrictions on location conditions, making it easier to handle more products than store sales.

On the other hand, in general, the number of products that can be displayed on a monitor provided in a computer is extremely small compared to the types of products handled by the shopping site, and each shopping site such as keyword search is not good at business opportunities. Various ideas have been devised so as not to do anything.

Special Table 2002-506256

However, in the case of keyword search, since an appropriate keyword is not conceived, it often takes a lot of work to arrive at a desired product. In addition, keyword search has synonyms and fluctuations, and there is a problem that a desired product often does not appear even though the keyword is appropriate. In addition, although there are clear uses, there are cases where people are looking vaguely, such as “things that can be used like this”.

On the other hand, there is also a method of tracing to a desired product in order from the category. However, you may not know which category the product belongs to. In addition, the classification itself is often inappropriate. For this reason, it is possible to wander between categories and eventually move to another site without reaching the desired product.

Also, when there is a certain purpose such as moving, many related products are needed at once. It takes a lot of time to select a product by keyword search. Even when selecting from categories, since many categories are related, it is still laborious. Eventually, you will forget to buy what you need, or it will be a hassle twice.

After all, it is in principle difficult to select the necessary products from the limited information displayed on the narrow monitor screen of the computer. Various ideas have been made for this problem, but no significant improvement has been made (see Patent Document 1).

Therefore, an object of the present invention is to provide a product purchase assistance system that allows a necessary product to be quickly purchased by an intuitive operation without inputting a product name of a desired product.

In order to solve the above problems, the product purchase assistance system according to the present invention transmits data necessary for displaying a screen for introducing products to a user terminal via a network, and by the user via the screen display. A product purchase assistance system that enables product selection and purchase, and includes a database that manages a history of product selection and purchase by a user in association with each user. When a plurality of filter items that can be selected are included and the user selects one filter item, by referring to the database, corresponding to the filter item selected by the user, the user-specific information is obtained from the above history. The purchase tendency is estimated, products with a high probability of purchase are displayed preferentially, and From the history, the probability of purchasing before and after the preferentially displayed product is calculated for the user and each product from the history, and the product having a high purchase probability is displayed at the same time. .

This configuration allows the user to purchase necessary products in a short time, and the product sales side is less likely to miss the opportunity to sell the product.

In a preferred embodiment, the preferentially displayed product and the product with a high purchase probability are connected by an arm emphasized according to the purchase probability. By adopting such a configuration, the user does not forget to buy, and the merchandise sales side can increase sales.

Furthermore, in a preferred embodiment, the arm calculates the probability that the product connected by the arm will be purchased one after the other for the user, and the higher the purchase probability, the thicker the item is displayed. And By adopting such a configuration, the user is promptly guided to the necessary product, so that the purchase can be performed smoothly.

Furthermore, in a preferred embodiment, the arm calculates an average gap in purchase time of the products connected by the arm for the user, and the longer the average gap is, the longer it is displayed. And With such a configuration, the user can grasp the relationship between the products sensuously, so that the purchase can be performed smoothly.

Furthermore, in a preferred embodiment, the product is displayed in a larger size as the purchase probability of the product is higher. By adopting such a configuration, highly important products are emphasized, and an effect can be expected to prevent users from forgetting to buy.

Furthermore, in a preferred embodiment, the plurality of filter items include an item representing a task related to a specific purpose or application. By adopting such a configuration, the user can efficiently carry out purchase procedures when performing any event such as moving or ceremonial occasion, and the product sales side can also increase sales. .

Furthermore, in a preferred embodiment, the plurality of filter items include an item indicating a product category. With such a configuration, the user can efficiently search for products.

Furthermore, in a preferred embodiment, the plurality of filter items include an item indicating a time. When a user specifies a specific time, history data associated with the user before that time is included in the history data. The product to be displayed is determined based on the above. With such a configuration, when the user wants to purchase the same product as the product purchased in the past, the user can quickly open the purchase page for the same product.

Furthermore, in a preferred embodiment, for the user, a product that is regularly purchased is extracted from the history, its period is estimated, and the product is displayed in accordance with the next purchase date estimated therefrom. With such a configuration, the user can quickly purchase products that are regularly purchased without stress.

Furthermore, in a preferred embodiment, the purchase history stored in the database includes data relating to external factors at the time of the purchase. For example, the external factors include at least one of weather, season, and access source IP address. With such a configuration, the user can quickly purchase a product according to the situation at that time.

Furthermore, in a preferred embodiment, when the preferentially displayed products and the products that have a high probability of being purchased before and after include products that have only a short time elapsed since the previous purchase. Display differently from other products so that the user notices that. With this configuration, the user can avoid unnecessary orders.

The product purchase assistance system according to the present invention has an effect that the user can search for products efficiently and promptly, and the product sales side can steadily increase the sales.

Hereinafter, a flush air cleaner according to an embodiment of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a diagram illustrating a product purchase assistance system according to an embodiment of the present invention. FIG. 2 is a diagram illustrating a configuration of a database used in the product purchase assistance system according to the embodiment of the present invention. FIG. 3 is a diagram showing a database task table used in the product purchase assistance system according to the embodiment of the present invention. FIG. 4 is an example of a display screen provided by the product purchase assistance system according to the embodiment of the present invention, and shows a screen displayed at the beginning of login. FIG. 5 is an explanatory diagram illustrating a related map generation process in the product purchase assistance system according to the embodiment of the present invention. FIG. 6 is a flowchart illustrating an initial screen generation process after login in the product purchase assistance system according to the embodiment of the present invention. FIG. 7 is a diagram illustrating a screen that is displayed when moving / moving is selected from the items for filtering in the commodity purchase assistance system according to the embodiment of the present invention. FIG. 8 is a flowchart showing a related map generation process when one task is selected from items for filtering in the commodity purchase assistance system according to the embodiment of the present invention. FIG. 9 is displayed when moving / moving is selected from the items for filtering in the product purchase assistance system according to the embodiment of the present invention, and then “window / customer service” is selected as the item of the type of business / type of job. FIG. FIG. 10 is a diagram illustrating a hierarchical structure of filtering items in the product purchase assistance system according to the embodiment of the present invention. FIG. 11 is a diagram illustrating a product list screen displayed after a specific product is selected in the product purchase assistance system according to the embodiment of the present invention. FIG. 12 is a diagram illustrating an example of a screen that is updated when a peripheral product displayed on the related map is dragged to a central position in the product purchase assistance system according to the embodiment of the present invention. FIG. 13 is a diagram for explaining the addition of filtering items in the commodity purchase assistance system according to the embodiment of the present invention. FIG. 14 is a diagram showing the configuration of a shopping mall to which the product purchase assistance system according to the embodiment of the present invention can be applied. FIG. 15 is a diagram showing a product list screen displayed after a specific product is selected in the product purchase assistance system according to the embodiment of the present invention. FIG. 16 is a diagram showing a list screen of dealers where a specific product can be purchased in the product purchase assistance system according to the embodiment of the present invention. FIG. 17 is a diagram illustrating a configuration of a market place to which the product purchase assistance system according to the embodiment of the present invention can be applied. FIG. 18 is a diagram showing another configuration of the database used in the product purchase assistance system according to the embodiment of the present invention. FIG. 19 is a diagram illustrating an example of a screen suitable for a mobile phone or a smartphone, which is a user interface in the product purchase assistance system according to the embodiment of the present invention. FIG. 20 is a diagram for explaining a specific example of the operation of the user interface shown in FIG. FIG. 21 is a diagram for explaining a specific example of the operation of the user interface shown in FIG. FIG. 22 is a diagram for explaining a specific example of the operation of the user interface shown in FIG. FIG. 23 is a diagram for explaining a specific example of the operation of the user interface shown in FIG. FIG. 24 is a diagram for explaining a specific example of the operation of the user interface shown in FIG. FIG. 25 is a diagram for explaining a specific example of the operation of the user interface shown in FIG.

Hereinafter, a product purchase assistance system according to an embodiment of the present invention will be described with reference to the accompanying drawings.
<System configuration>

FIG. 1 is a diagram showing a product purchase assistance system according to an embodiment of the present invention. This merchandise purchase assisting system is composed of a server computer 1 and a program installed and executed on the server computer 1, and merchandise is sent to client computers 3-1... 3-N connected to the Internet. Providing services to assist in the purchase of The elements of this program include a WEB server program corresponding to SSL and a database management system used from this server program.

In this embodiment, when the client computers 3-1... 3-N connect to the server computer 1, user authentication is performed. In other words, the user information accessed via the client computers 3-1... 3-N is stored in the server computer 1 in advance, and the user information, access history, product purchase history, etc., as will be described later. Is managed in the database.
<Database configuration>

The past access history and purchase history are managed in a database as shown in FIG. The database includes five tables. That is, a sales table, a customer table, a task table, a product table, and a general name table. These constitute an RDB (relational database).

Each record of the sales table includes fields of sales date / time, customer ID, general product name, product ID, quantity, and task ID. For example, when a certain quantity of products specified by the product ID is purchased by a customer ID user at 22:49:30 on August 9, 2009, it is recorded as one record.

Even if a user logs in, selects a specific product, and then cancels the purchase of the product, the user is most likely interested in the product. Therefore, a new record is created in the sales table with the detailed product information displayed, and the quantity is set to zero. If the user has purchased a certain quantity, the quantity is entered, but even if the purchase is canceled, a record of quantity 0 is recorded as a valid record.

The task ID of the sales record indicates that the purchase was made after the task specified by the task ID was selected. If the task is purchased without being selected, NULL is input. The task items displayed here are purchased when the user performs an event such as moving or ceremonial occasion, and are related to a specific purpose or application. However, since the relationship between the product and the task is determined by what the user purchases while the task is selected, it is automatically adjusted to the optimal one from time to time.

Each record of the task table includes fields for a task name, a hierarchy level, an upper task ID, and a pointer to a lower task ID list. For example, as shown in FIG. 3, the task name “department transfer”, the hierarchy level is 2, and the task ID “moving” is stored in the upper task. The hierarchy level indicates how many levels below the highest level is 1. If a lower task is included, the lower task can be accessed from a pointer to the ID list.

Each record of the product table includes fields of product ID, product name, general name, category, usage ID, and unit price. Further, a field for a pointer to a list of IDs and frequencies of tasks selected at the time of purchasing the product is provided. For example, if the product is purchased 100 times in the past half year after the task “department transfer” is selected, an item consisting of the task ID of “department transfer” and the frequency 100 is the task ID of the product. Will be included in the list. The usage ID is appropriately used in relation to the usage of the product, and a specific usage example will be described later. The pointer provided in association with the common name is a pointer to the average price of products having the same common name and a list of aliases of the common name.

Each record of the generic name table includes the generic name of the product, the average price of the product having the generic name, and aliases 1 to 4 of the generic name. The average price is updated dynamically as appropriate. If there is no alias, NULL is appropriately substituted.

検索 Search the above database and display on the screen with priority from the ones that the user is likely to purchase. The display name here is usually a general name, and an individual product name is appropriately used. For example, the general name “鋏” is displayed at the beginning of the search, but a specific product name such as “craft scissors long blade 210 mm CBS-1500” is selected and displayed immediately before purchase.

In the present invention, for example, a user who wants to purchase a curing mat can quickly reach the list of curing mats and select a desired product from the list. If it is a curing mat, it may be possible to directly search for a keyword using “curing mat” as it is, but there are actually various problems. First, there is a case where the keyword “curing mat” is not hit and the hit is not “curing sheet”. In addition, the term “curing mat” may not appear. Furthermore, instead of searching for a specific product, there is a case where there is a certain purpose (task) and it is desired to purchase a product necessary for the purpose.

Therefore, unless otherwise specified, in this specification, when an expression such as product i is used, it indicates the entire product having a certain general name. Further, the meaning of displaying the product i is the meaning of displaying the general name. Click on the generic name to display a list of corresponding products. However, in a very special product, the product name may be a general name as it is.

The priority given to a specific display name (general name) is determined as follows. As shown in FIG. 4, a list of items for filtering is displayed on the left side of the screen. This filtering is the basic interface for user product selection.
<Overview of related map generation process>

FIG. 5 is an explanatory diagram showing a related map generation process. In the present embodiment, the priority order list update process and the related map generation process are the main processes. The priority order list is updated by merging the intermediate priority order list obtained by the filtering process of all sales records with the current display priority order list.

Filter processing is processing that extracts records that satisfy given conditions. This condition is given mainly in relation to operations performed by the user on the screen. For example, when a task such as moving is selected, a record related to the task is extracted. A priority is given to the extracted records to make an intermediate priority order list. The intermediate priority order list and the current display priority order list are merged into a new display priority order list.

Based on this display priority order list, products with high priority are selected, products with high relevance to this product with high priority are extracted, and their mutual relationship is visually expressed. It is a related map. When this related map is displayed, if the user performs another operation on the screen, the display priority order list is updated, and the related order is displayed based on the updated display priority order list. The map is reconstructed.
<Initial screen display>

Since the filtering is not performed at the beginning of login, the displayed product is determined only from the purchase history of the logged-in user. FIG. 4 shows an example of a display screen at the beginning of login. Hereinafter, a specific example of the display product determination process will be described.

FIG. 6 is a flowchart showing a process for generating an initial screen after login. First, in step S601, a sales record having the customer ID of this user is extracted from all sales records. The extracted sales records are rearranged in the general name field. In step S602, for each of the m general names included in the n sales records, the priority P (Ti) is set from the sales record having the general name Ti (1 ≦ i ≦ m ≦ n). Calculate using the following formula.
P (Ti) = Σ (W (δj (Ti)) × nj)

Here, W (δj) indicates the weight of the j-th sales record, and Σ means that the sum of the sales records having the general name Ti is obtained. For example, if the 29th to 33rd sales records have the common name Ti, the sum for j = 29 to 33 is obtained.

Where nj is a weight based on the quantity of the jth sales record. In order to consider the case where the purchase is not completed, 1 is set if the quantity is 0, and 2 if the quantity is 1 or more.

Δj indicates the difference between the creation date of the jth sales record and the current date. As the simplest implementation of W (δj), for example, if δj is less than 31 days, W (δj) = 1, and if it is 31 days or more, W (δj) = 0. In this case, the priority considering only the data for the past month is calculated.

Or, if δj is 31 days or more, W (δj) = 0, and if it is less than 31 days, N is the difference days and W (δj) = (31−N). In this case, a newer record can be given a greater weight.

In another implementation, in addition to the user-specific purchase trend, the overall purchase trend can also be reflected in the display. In other words, the priority P (Ti) is calculated by the following equation in which the coefficient αj is introduced for the jth sales record.
P (Ti) = Σ (W (δj (Ti)) × nj × αj)

Here, αj is set to 1 if the sales record does not belong to the user, and αj is set to a constant value larger than 1 if the sales record belongs to the user. By adjusting this constant value, the weight of the purchase tendency of the user with respect to the whole can be changed. The specific numerical value of αj should be set larger as the number of all users increases. For example, αj of the sales record of the user can be the number u of all users. In this case, the data of the current user is evaluated to the same extent as the data of the other users as a whole.

By the above processing, a list consisting of general names and their priorities is obtained. By rearranging this list in descending order of priority, a general name priority order list ranked by priority is obtained (step S603). As will be described in detail later, the display of the product is determined based on the rank in this list.

Also, when displaying, the priority value is reflected in the size of the bubble displaying the general name. That is, the displayed general name bubble is displayed larger as the priority is higher. Also, the general name with the highest priority is displayed in the center. Here, the general name is displayed in the bubble as a closed frame, and the priority is expressed by the size of the bubble. However, the priority is expressed by the size of the character string of the general name without using the closed frame. Also good. In that case, a sense of stability can be given to the display by appropriately underlining.

In the example of FIG. 4, since the product having the highest priority is “copy paper”, “copy paper” is displayed as the largest bubble in the center of the screen. As will be described later, the selection of products displayed in the vicinity is performed based on the degree of association with the product displayed in the center. In the “copy paper” bubble, the product name of the product most likely to be purchased among the “copy paper” is also displayed as auxiliary data. Click on the product name to open the purchase page for that product directly.

Next, in step S604, the relationship between the product with the general name having the highest priority and the other products is quantified from the difference in purchase date and time. In other words, if a certain product is frequently purchased at the same time or with a small time difference with another specific product, the convenience of the user is enhanced if they are displayed side by side. In the example shown in the figure, the relevance between the central product i (here, “copy paper”) having the highest possibility of purchase and the other display product j is evaluated.

For example, the following values are calculated for all sales records within two days before and after the purchase date of each sales record of the product i with the highest priority. Here, A (i, j) indicates the degree of association between the central product i and the product j, the purchase date of each sales record of the product i is dik, and the product purchased within two days before and after this purchase date. The purchase date of j sales record is set to djl, and a number sij is introduced such that sij = 2 for simultaneous purchase and sij = 0 for non-simultaneous purchase, and added for all combinations of dik and djl.
A (i, j) = Σ ((3−2 × | dik−djl | + sij) × nj)

In this formula, if | dik−djl | is 1, that is, if it is a day off, the value in parentheses is 1. In addition, | dik−djl | is 0, that is, the same day, and if it is not simultaneous purchase, the value in parentheses is 3. Furthermore, | dik−djl | is 0, that is, the same day, and if it is purchased at the same time, the parenthesis is 5. Whether or not they are simultaneous purchases is determined by whether or not the purchase dates and times are exactly the same.

Then, in step S605, a certain number of products are displayed from those having a high relevance frequency A (i, j) with the product i. Specifically, the product having the largest relevance frequency A (i, j) with the central product i is disposed in the vicinity of the product i, and is disposed farther as the relevance frequency A (i, j) becomes smaller. Further, as shown in FIG. 4, the central product i (copy paper) and other products are connected by an arm. The arm has a thickness proportional to the related frequency A (i, j). For example, a product with a relevance frequency of 10 is connected to the product i at the center with an arm twice as thick as the arm of the product with 5 items. In this specification, such a diagram is referred to as a relation map.

For example, in the example of FIG. 4, it can be seen that there is a high possibility that the toner is purchased at the same time as the copy paper with a small time difference. As described above, the “toner” is located near the “copy sheet”, so that it is more conspicuous, and the possibility of purchase is further increased. In addition, it is understood that “postcard” is highly likely to purchase “copy paper” after a while.

Also, “toner” and “copy paper” are common in terms of consumables for copy machines. Commodities having such a relationship are given the same application ID in advance. Then, compare the usage ID of the product in the center of the related map with the usage IDs of the neighboring products, and if there is the same, change the color of the arm connecting them and emphasize that they are in a special relationship. May be. For example, a normal arm is blue, and if the same application ID is given, for example, orange, the arm can be made conspicuous.

Further, the related map also has a degree of freedom in the vertical and horizontal directions. In order to make use of this, here, the arrangement of the products constituting the related map is determined from the sales data as follows. In other words, the degree of association U (Ti) with the user is calculated according to the following formula, and the larger degree of association U (Ti) is displayed on the upper side.
U (Ti) = Σ (W (δj (Ti)) × βj)

Here, W (δj (Ti)) is as described above. If the sales record does not belong to the user, βj is set to 0, and if the sales record belongs to the user, βj is set to 1. In this case, basic products that are generally purchased, that is, products that have been purchased for all users are displayed on the lower side.

Also, the distribution of the related map to the left and right positions is performed according to the average price. That is, a product with a high average price is displayed on the left and a cheap product is displayed on the right. This rule is clearly specified in advance, but the user unconsciously sees the related map with such a rule in mind during the repetition. Thus, by utilizing the degree of freedom of display, the user can more easily select a product.

In addition, the sales record of the product i extracted from all the sales records is used for the calculation of the example related frequency A (i, j). However, if there are more than a certain number of sales records, if only the sales record of the product i having the current user ID is targeted, the relevance frequency A (i, j that more strongly reflects the current user's sales trend. ) Can be calculated.

In order to obtain a more effective relevance frequency A (i, j), it is better to calculate the relevance frequency A (i, j) by using the following formula, using the data used for the calculation as sales records for all users' products i .
A (i, j) = Σ ((3−2 × | dik−djl | + sij) × nj × u)

Here, the coefficient u is 1 for sales records of users other than the current user. In the case of sales records of the current user, appropriate weighting is performed. The specific numerical value should be set larger as the number of all users is larger. For example, αj of the sales record of the user can be the number u of all users. In this case, the data of the current user is evaluated to the same extent as the data of the other users as a whole.

Therefore, the generic name of the product with the highest possibility of purchase is displayed in the center with the largest bubble, and the product that is more relevant to the product is displayed closer to the center and connected to the product with a thicker arm Is done. The size of the bubble indicates the high possibility of purchase. Such a display method is common to all the following display examples, and thus a duplicate description is omitted.

In the display of only the related map, the product with the largest number of sales records and the product with little relation are not displayed even if the number of sales records is large. Therefore, in this embodiment, a certain number (two in the illustrated example) of the products with the largest number of sales records among the products not included in the related map is displayed as candidates. When one of the candidates is clicked, the related map is reconstructed and displayed again by the same process with the candidate as the center.

Furthermore, the same user may purchase the same product periodically. Many of these products are consumables, but a record of the same product is extracted in advance from each user's sales record, and its purchase cycle is calculated. More specifically, the product quantity is normalized by using the most frequently used product quantity as a unit in the product quantity of one purchase. Each purchase interval is also normalized by the reciprocal of the normalized product quantity. Then, the average of the normalized purchase intervals is set as the purchase cycle.

Specifically, considering m samples, the unit product quantity is n, the i-th purchase quantity is Ni, the i-th and i + 1-th purchase intervals are Di, and the purchase cycle C is given by Calculated. Note that Σ is the sum of i from 1 to m−1.
C = Σ (Di × n / Ni) / (m-1)

And if it displays as a recommended product according to the period and promotes purchase, the probability of purchase will increase. Also, from the user's perspective, there is no such thing as no stock when needed. For example, when a box tissue is purchased regularly at an interval of two months on average, the box tissue may be displayed as a recommended product after a month and a half. In this way, when recommended products are displayed at the timing when purchase is expected, it is convenient that the user does not forget to purchase from the viewpoint of the user.
<Filtering>

In the initial state of FIG. 4, the possibility of purchase by the user is judged and displayed only from the database information. Now suppose that the user is looking for what he needs to move or move. In that case, the user clicks and selects “moving / moving” from the list of items for filtering displayed on the left side of the screen.

When selected, the item “moving / moving” is displayed so that the selected state can be seen, such as a font change. This is because the selected item is known when a plurality of items are selected at the same time. From this selection information, the system updates the priority order list, and the related map is reconstructed as shown in FIG.

FIG. 8 is a flowchart showing a related map generation process when moving / moving is selected from the items for filtering. First, in step S801, sales records that are moving / moving with a task ID selected are extracted from all sales records. From the extracted sales record data set, a record associated with the common name Ti is extracted and rearranged in the general name field.

In step S802, the sales record priority P (Ti) is calculated. That is, the difference between the creation date and time of the sales record having the common name Ti and the current date and time is δj (Ti), and for each of the m common names included in the n sales records, the common name Ti (1 ≦ 1 The priority P (Ti) is calculated from the sales record having i ≦ m ≦ n) by the following formula. The meaning of this formula is the same as that of the above formula, except that the target data set is different.
P (Ti) = Σ (W (δj (Ti)) × nj × αj)

Then, in step S803, an intermediate priority order list of general names ranked by this priority is created. Again, this list includes the general name and its priority.

In this embodiment, thereafter, in step S804, the intermediate priority order list created in relation to the selected task is merged with the priority order list currently used for display. However, the already displayed general name is deleted from the previous priority order list. This is because, if there is a necessary general name already displayed, it is considered that it has been clicked.

Specifically, first, the intermediate priority is set so that the priority of the top of the intermediate priority order list is, for example, twice the priority of the top of the priority order list immediately before the general name already displayed is deleted. Weight all priorities in the order list. For the same common name, the priorities are added together to form one entry, and all entries in both lists are arranged in order of priority to form one priority order list for display.

Next, a related map centering on the general name at the top of the updated priority order list is created in the same manner as in the above example. That is, in step S805, the related frequency A (i, j) is calculated, and in step S806, a related map is created and displayed. However, here, the relevance frequency A (i, j) is calculated using the following formula described above for the sales record data set after filtering, that is, the sales record having the moving / moving task ID. .
A (i, j) = Σ ((3−2 × | dik−djl | + sij) × nj × u)

Then, the created related map is displayed on the screen. Again, a certain number (two in the illustrated example) of the products with the largest number of sales records among the products not included in the related map is displayed as candidates. When one of the candidates is clicked, the related map is reconstructed and displayed again by the same process with the candidate as the center.

Here, when the user clicks another filtering item, for example, “window / service”, a sales record in which the task ID is “moving / moving” and the type / job type is “window / service” Extract from all sales records. Thereafter, the processing from step S802 onward is performed in the same manner, the priority order list is updated, the related map is reconstructed, and displayed as shown in FIG.

However, the calculation of the relevance frequency A (i, j) performed in step S805 is related to “moving / moving” and “window / serving” because “moving / moving” and “window / serving” are selected. This is done for sales records.

Therefore, in this case, in the updated priority order list for display and related maps, for the purpose of moving / moving, products that are purchased by users whose type of industry / job type is “Contact / Service” are given priority. Is done. Furthermore, since the first priority order list is based on the current user data and is weighted to the current user data, it also reflects the purchase trend of the current user.
Here, if the related map includes a product that the user has just ordered, there is a possibility that an unnecessary order may be inadvertently placed. Of course, the product may be really necessary for some reason, so it is considered inconvenient not to display the product itself. Therefore, although the display itself is performed, it is desirable to perform a display different from other products so that the user notices that. For example, the user may be alerted not to place duplicate orders by displaying in gray.
If this gray-colored item is clicked here, in addition to the details of the item, the latest order date and quantity are also displayed. If you have forgotten your last order, you can check this display. If you really need it, you can place an order. In this way, the user can check the stock on the spot and avoid wrong orders.

In addition, when the selection of the task ID is canceled, the sales record that is the “selection / customer service” that is the remaining selection item is extracted from all the sales records, and the same processing is performed again. In this way, selection and deselection can be repeated until a desired product is displayed.

Filtering items include categories and manufacturers as shown in Fig. 4, etc., but these correspond to each sales record uniquely in the database, so they are the same as the above tasks and industries / job types Can handle.

Each filtering item has a hierarchical structure. For example, if you click on office supplies, the writing items etc. will be displayed as the filtering items one level below, and if you click on the writing implements, it will be ballpoint pen, magic, pencil, etc. Expanded (see FIG. 10). In addition, there is an item “others” for designating a filtering condition not belonging to any of them. Furthermore, there is an item “add” at the end of the filtering item. A processing example when “add” is selected will be described later.

フィルタリング Filtering by the date of sales record in the “Timing” filtering item. Here, you can specify the first and last date of the target period. By default, it is from January 1, 2000 to the current date (here, October 10, 2009), and since this is intended for all sales records, virtually no filtering is performed.

For example, in the example shown in FIG. 12, since the last date of the target period is designated as December 31, 2008, only the sales records before this date are extracted and then the same as above. Perform the process. That is, processing is performed so that it is displayed when the same operation is performed on December 31, 2008. This is effective when a certain event is held at the end of 2008 and a similar event is performed, and it is desired to make a similar purchase.

Also, if the first date of the target period is specified, the data before that can be invalidated. For example, this is effective when there is a significant change in business at a certain time and it is desired to eliminate the influence of previous data.

Also, if a product is put in the cart on the way, the product is displayed in the center of the related map. For example, when a steel cart is selected in FIG. 7, a list of product names of the steel cart is displayed as shown in FIG. The user selects a product from the list and places it in the cart. Then, the related map displayed next is centered on the steel cart. This is because if a product purchased at the same time as the steel cart is displayed, the possibility of purchase is high.

Note that the actual purchase procedure performed via the cart from the product list screen and the subsequent payment method are the same as the online sales system that is currently generally used, so the details are omitted here.

Similarly, when a product other than the product in the center of the related map is moved to the center on the way, for example, when “postcard” is dragged to the center with the mouse in FIG. 4, as shown in FIG. Moves to the center position. Then, this time, the related map is reconstructed around “postcard”. In this case, there is no change in the priority order list, but the products displayed on this related map are excluded when updating.

Note that the statistical criterion at the upper right of the screen displayed in FIG. 4 or the like can be switched from the setting of the current user (here, “Mr. S”) to “general” that does not distinguish the user. By using these options as appropriate according to the purpose and purpose, it is possible to reliably reach the necessary products.
<Update and customization>

∙ Some of the above filtering items are used very frequently and rarely. This is influenced by time and changes in the user composition, but it is inefficient to display things that are rarely used in prominent positions. As described below, this display is dynamically adjusted according to utilization.

First, the number of filtering items that belong directly under the highest level classification such as task, category, industry / job type is fixed according to the utilization rate. For example, it is assumed that the number of filtering items belonging directly under the task is 6, while the number of filtering items of a manufacturer with a lower utilization rate is 4. In this case, except the last “others”, the items are displayed in order from the highest utilization rate. When the usage rate decreases, the display position also decreases. When the usage rate ranks 6th, it is included in “Others”. Similarly, the display position of the lower filtering items is adjusted according to the ranking of the utilization rate. The usage rate R (Ti) is calculated by the following formula for each filtering item Ci.
R (Ti) = Σ (W (δj (Ci)) × αj)

Here, W (δj) indicates the weight of the j-th sales record, and Σ means that the sum of the sales records related to the filtering item Ci is obtained in the same manner as the above equations. Again, for example, if the 29th to 33rd sales records are related to the filtering item Ci, the sum for j = 29 to 33 is obtained. Again, αj is set to 1 if the sales record does not belong to the user, and αj is set to a constant value larger than 1 if the sales record belongs to the user. By doing in this way, even if the number of sales records of the user is small, it is possible to display efficiently from the overall tendency.

However, the value of αj will be different depending on the highest classification. Here, when the highest classification is a manufacturer, the user-specific orientation is prioritized for the filtering items that belong directly below the manufacturer. This is because filtering by a manufacturer is often selected when the user wants to reflect his / her preference. For example, if the sales record belongs to the user, αj is doubled the total number of users, and the orientation of the user is given higher priority.

Also, users can add their own filtering items. For example, when an emerging manufacturer gets a lot of hot selling products, they may want to add the name of that manufacturer. In such a case, it is sufficient to select manufacturer → other → addition. When “Add” is clicked, a text box is displayed and a new item can be added (see FIG. 13).

The added filtering item is displayed only to the user who made the addition. However, when many users add the name of the manufacturer and the number of sales records increases, the ranking of the utilization rate increases and it is displayed to all users. In this way, the system can be maintained in a more appropriate state according to the circumstances of the user by user customization or purchase operation.
<Mounted application>

As a typical example of the above-described embodiment, it is possible to mount a server application that implements a net shop on a user interface. That is, the sales record is a sales record at the net shop. In addition, if sales records from many online shops other than your own online shop are used as data, product purchase assistance can be performed more efficiently as the number of samples increases.

Specifically, the above embodiment can be implemented in a shopping mall as shown in FIG. In this case, since the data exchanged between the member store (shops A to Z) and each user passes through the shopping mall management server, the purchase data at each member store can be used as the sales record.

When the user accesses this shopping mall, the screen shown in FIG. 4 is displayed. Then, the purchased product is determined according to the flow described above. Up to this point is based on the operation of the present embodiment, and an interface having the screen configuration as described above is provided as appropriate. However, the page indicating the purchased product candidate is displayed as shown in FIG. 15 instead of FIG. That is, the button for entering the cart is a button for displaying a list of stores. Here, if a specific product is selected by clicking a store list button, a list of stores that can be purchased as shown in FIG. 16 can be displayed. Furthermore, if an arbitrary store is selected, it is possible to move to the product purchase page of the store.

In the above example, the case of a shopping mall has been described. As a similar system, there is a market place shown in FIG. In the case of a shopping mall, the shop needs to become a member store after paying the membership fee after examination and a certain rent is required. In the case of marketplaces, individuals can freely buy and sell without the need for special membership procedures or rent. In this case as well, the above embodiment can be applied.

Furthermore, it can be implemented separately from online sales. That is, in the above embodiment, the present invention is applied to the front end of the sales system. In this case, participation in shopping malls and marketplaces is a premise because it is involved in the center of the sales process of the store. Therefore, the sales data has a certain range. If this is implemented separately from online sales, such a range can be eliminated.

More specifically, the list of purchased product candidates shown in FIG. 11 is almost the same system as in the above embodiment. However, when one of the purchased product candidates is selected, the screen moves to a detailed explanation screen of the purchased product candidate. This detailed explanation screen is displayed by linking a URL such as the homepage of the manufacturer of the purchased product candidate. Therefore, in this case, rather than purchasing a product, the purpose is to present a purchase product candidate and assist the user's selection.

In addition, since it is separated from purchasing and sales here, it is necessary to investigate and update the product information from time to time. In addition, the product information is obtained by keyword search performed by the user, and the update is performed by adding the product information. In addition, regarding the acquisition of sales records, if you leave the product purchase assistance system site and move to the shop site, data cannot be obtained. Therefore, since the sales record is obtained by the information of the operation performed by the user until the selection from the list of purchased product candidates, the sales record is acquired as information indicating the intention of purchase, not the actual purchase information. The operating cost of this system is covered by affiliates and advertising revenue.

In addition, after one of the purchased product candidates is selected, instead of moving to the detailed explanation screen of the purchased product candidate, the user may move to a net auction page. That is, when a purchased product is determined from the list of purchased product candidates shown in FIG. 11 and FIG. 15, the purchased product is searched by a net auction and moved to a page showing the result. When targeting a plurality of net auctions, the search results are displayed side by side.
<Reference of external factors>

In the product purchase assistance system according to the embodiment of the present invention, external factors for purchase such as weather, season, access source IP address can be taken into consideration. Such external factors can be statistically processed by recording them in a database.

An example of such a database is shown in FIG. In this example, fields of weather and IP address are further provided in addition to the sales record in the database shown in FIG. In addition, the usage ID of the product record includes information on outdoor use or indoor use and information on a typical number of users.

That is, when registering a sales record, the weather at that location and the IP address of the access source are recorded at that time. The IP address can be used to determine whether the client terminal is a mobile phone or a smartphone. When accessing a web server from a mobile phone, the IP address of the transmission source notified to the web server side is in a certain band for each mobile phone carrier and can be used for the purpose of distinguishing it from access from other terminals. .

If the client terminal is a mobile phone, the location can be specified using the location information service of the mobile phone carrier using the location of the base station. If the IP address is that of a general terminal other than that, it is possible to specify the area from the range of the IP address. After the location of the client terminal is specified, the weather there is obtained using the Web API of the weather data providing service. Then, weather information is input by setting a specific bit in the weather field.

For example, the weather field includes at least three bit pairs, and if the first bit pair is “01”, it is clear, “10” is rain, and “11” is cloudy. “00” indicates that it is unknown or no information is included. Also, if the next bit pair is “01”, the maximum temperature is 3 degrees or more higher than normal, “10” is 3 degrees or more lower than normal, and “11” is otherwise (ie, almost normal). It is shown that. “00” indicates that it is unknown or no information is included. Further, if the next bit pair is “01”, the humidity is 80% or higher, “10” indicates that the humidity is 50% or lower, and “11” indicates the other (ie, intermediate value). “00” indicates that it is unknown or no information is included.

Information regarding outdoor use or indoor use is input by setting a specific bit pair of usage ID when registering a product record. For example, “01” is suitable for indoor use, “10” is suitable for outdoor use, and “11” is suitable for both outdoor use and indoor use. “00” indicates that it is unknown or no information is included.

Furthermore, the information regarding the “number of people” is also input by setting another bit pair of the usage ID when registering the product record. For example, “01” is suitable for one person, “10” is suitable for two persons, and “11” is suitable for three or more persons. “00” indicates indefinite, unknown or no information is included.

From the sales record as described above, the priority P (Ti) is calculated by the following formula.
P (Ti) = Σ (W (δj (Ti)) × nj × αj) + γj

As described above, W (δj) indicates the weight of the jth sales record, nj indicates the weight based on the quantity of the jth sales record, and αj indicates whether the jth sales record belongs to the user. Σ means that the sum of sales records having the general name Ti is obtained. That is, the subscript j represents the divisor of the sales record having the general name Ti.

Regarding the priority P (Ti), a factor indicating an external factor is further added as γj here. γj is calculated by the following equation.
γj = Cd × | Σ (91−Dj) | + Cw × | Σf (Iw, wj) | + Ch × | Σf (Ih, hj) | + Cg × | Σh (Ig, gj) |

Here, Dj indicates the seasonal difference between the date of purchase of the jth sales record and the current date in days. That is, the month and day are extracted from the current time, and the current date is the c-th day of the year (1st if 1/1, 365th or 366th if 12/31), and similarly jth Dj is calculated as 365- | cdd if | cd |> 182 as the d-th day of the sales record of the year, otherwise Dj is | cd | Is calculated. Therefore, Dj is a maximum of 182 days. If the data is unrelated to the season, Dj is considered to have an average value of approximately 91 days, and the contribution to γ in that case is lost.

Iw and wj are the above bit pairs representing sunny / rainy / cloudy, where Iw is the current value and wj is the value of the jth sales record. Further, Ih and hj are the above bit pairs representing humidity, Ih is the current value, and hj is the value of the jth sales record. The function f (x, y) follows the following formula.
f (10,10) = 1, f (10,01) =-1, f (10,11) = 0, f (10,00) = 0
f (01,10) =-1, f (01,01) = 1, f (01,11) = 0, f (01,00) = 0
f (11,10) = 0, f (11,01) = 0, f (11,11) = 1, f (11,00) = 0
f (00,10) = 0, f (00,01) = 0, f (00,11) = 0, f (00,00) = 0

The value of this function f (x, y) is selected to reflect the value of γ if there is any correlation between the sunny / rainy / cloudy or humidity information and the sales record, and not to reflect if there is no correlation. Has been. The accuracy can be further improved by adjusting the value according to the general distribution of sunny / rainy / cloudy and the general bias of humidity.

Ig and gj are true / false values (1, 0) indicating whether the client terminal is a mobile phone or a smartphone, Ig is a current value, and gj is a value of the jth sales record. . The function h (x, y) follows the following formula.
h (1,1) = 1, h (1,0) = 0, h (0,1) = 0, h (0,0) = 1

In the formula, Cd, Cw, Ch, and Cg are constants for balancing with other items. Further, the degree of association A (i, j) between the central product i and the product j in the related map is calculated by the following equation.
A (i, j) = Σ ((3−2 × | dik−djl | + sij) × nj + ηij

As described above, the purchase date of each sales record of product i is set to dik, the purchase date of the sales record of product j purchased within two days before and after this purchase date is set to djl, and sij = 2 for simultaneous purchase If it is not simultaneous purchase, a number sij where sij = 0 is introduced, and all combinations of dik and djl are added.

Ηij is a factor for adjusting the degree of association A (i, j) between the product i and the product j, and is based on information on outdoor use or indoor use, information on typical number of users, and IP address information. It is calculated by the following formula.
ηij = Co x f (oi, oj) + Cp x f (pi, pj)

Here, oi and oj are the above-described bit pairs representing information regarding outdoor use or indoor use, oi indicates the value of the i th product record, and oj indicates the value of the j th product record. Further, pi and pj are the above-described bit pairs representing information on the typical number of users, pi indicates the value of the i-th product record, and pj indicates the value of the j-th product record. The function f () is the same as the function f (). Again, Co and Cp are constants for balancing with other items.
<Optimization for mobile phones>

* The display screen is small on mobile phones and smartphones, and the entire Web page provided by the above product purchase assistance system cannot be displayed. In recent years, it has been desired to utilize a multi-touch display that has been widely used in mobile phones and smartphones.

In view of such circumstances, it is expected that the operability can be improved even on a small display screen by simplifying the design of the Web page provided by the product purchase assistance system. Hereinafter, an example of such an implementation will be described.

One of the important elements in the product purchase assistance system is the related map. The association map provides an intuitive graphical interface and is highly consistent with touch operations. Therefore, the related map is used as a main operation target, and the touch function implemented in the multi-touch display is used as an interface.

The drag operation can be performed as described above and functions as described above. That is, when “postcard” is dragged to the center with a finger in FIG. 4, “postcard” is moved to the center position as shown in FIG. Then, this time, the related map is reconstructed around “postcard”. This makes it possible to narrow down the products, but it takes time and effort when the displayed information is wide open with the necessary information. In order to access necessary information more quickly, the following operations are possible.

First of all, by pinching in (operation to place two fingers on the screen to reduce the interval) and pinch out (operation to increase the interval by placing two fingers on the screen) To do. That is, in the state shown in FIG. 19, first, a postcard bubble is selected by pushing it with two fingers. If you hold the two fingers away from the display and pinch out to spread them, the postcard information is “enlarged”. That is, as shown in FIG. 20, information on only the products related to the postcard is developed in detail. This product selection and display method is performed according to the above algorithm. Further, when the bubble with “picture” is pinched out, a related map developed in the information of the postcard with picture is displayed.

On the other hand, if the entire product is pinched in the state shown in FIG. 20 where products related to postcards are displayed, a related map centering on postcards is obtained. However, this time, the merchandise shown in FIG. 19 is not displayed, and as shown in FIG. 21, stationery centered on postcards is displayed. Further, when pinch-in is performed, as shown in FIG. 22, the category item one level higher is displayed.

Also, to return to the previous display, it is only necessary to perform a flick (an action to move the screen with your finger). In other words, the screen is flipped to the right as if there is a previous display on the left side of the current screen. By this operation, for example, the display of FIG. 22 returns to the display of FIG. Conversely, to display before returning, flick leftward. By this operation, for example, it is possible to proceed again from the display of FIG. 21 to the display of FIG.

In addition, the function key consisting of task, category, industry / job type and manufacturer is displayed at the bottom of the screen, and by tapping this, the corresponding items are displayed as shown in FIG. The filtering operation as described above is also possible.

Also, when you tap the search window, a software keyboard appears at the bottom of the display, so you can enter search keywords. For example, as shown in FIG. 24, when a search is performed by inputting paint, “paint” is displayed at the center position. Then, the relation map as shown in FIG. 25 is reconstructed with “paint” as the center. Thereafter, necessary products can be traced by repeating the above-described operation.

As described above, the present invention has been described in detail according to the embodiments. However, it is obvious for those skilled in the art that the present invention is not limited to the embodiments described in the present application. The present invention can be implemented as modified and changed modes without departing from the spirit and scope of the present invention defined by the description of the scope of claims. Therefore, the description of the present application is for illustrative purposes and does not have any limiting meaning to the present invention.

As described above, the merchandise purchase assistance system according to the present invention is effective when it is desired to efficiently buy and sell merchandise through a network, and has great industrial applicability.

1 Server computer 3-1 ... 3-N Client computer

Claims

A product purchase assistance system that transmits data necessary for displaying a product introduction screen to a user terminal via a network, and enables the user to select and purchase a product through the screen display. , A database for managing the history of product selection and purchase by users in association with each user, and the screen display includes a plurality of filter items that can be arbitrarily selected by the user,
When the user selects one filter item, by referring to the database, the purchase tendency peculiar to the user is estimated from the history corresponding to the filter item selected by the user, and the probability of the user purchasing is estimated. Prioritize expensive products,
Further, from the history, the probability of purchasing before and after the preferentially displayed product is calculated for the user and each product from the history, and the product having a high purchase probability is displayed simultaneously. Product purchase assistance system.
The commodity purchase assisting system according to claim 1, wherein the preferentially displayed commodity and the commodity having a high purchase probability are connected by an arm emphasized according to the purchase probability.
The said arm calculates the probability that the goods connected with this arm will be purchased one after another for the user, and the higher the purchase probability, the thicker the item is displayed. Product purchase assistance system.
3. The arm according to claim 2, wherein the arm calculates an average gap in purchase time of the products connected by the arm for the user, and the longer the average gap is, the longer it is displayed. Product purchase assistance system.
2. The product purchase assistance system according to claim 1, wherein the display of the product is displayed larger as the purchase probability of the product is higher.
The commodity purchase assistance system according to claim 1, wherein the plurality of filter items include an item representing a task related to a specific purpose or application.
The product purchase assisting system according to claim 1, wherein the plurality of filter items include an item indicating a product category.
The plurality of filter items include an item indicating a time, and when a user specifies a specific time, products to be displayed based on historical data associated with the user before the specified time The commodity purchase assisting system according to claim 1, wherein: is determined.
The product which is regularly purchased for the user is extracted from the history, the period is estimated, and the product is displayed according to the next purchase date estimated therefrom. Product purchase assistance system described in 1.
2. The commodity purchase assistance system according to claim 1, wherein the purchase history stored in the database includes data relating to external factors at the time of the purchase.
11. The commodity purchase assistance system according to claim 10, wherein the external factor includes at least one of weather, season, and IP address of the access source.
If any of the above preferentially displayed products or products that have a high probability of being purchased before or after is included in a product that has only a short time elapsed since the previous purchase, the user notices that fact. The product purchase assisting system according to claim 10, wherein the display is different from other products.
The purchase probability and the probability of purchases in succession use the history of selection and purchase of products by the entire user using this product purchase assistance system in addition to the history of selection and purchase of products by the user. The commodity purchase assistance system according to claim 1, wherein the commodity purchase assistance system is calculated as follows.