Shelf allocation support device, shelf allocation support method, and recording medium

Abstract

This shelf allocation support device is provided with: a determination unit that determines candidate shelves for displaying a target product on the basis of human flow information including at least one of the number of customers passing by each of a plurality of shelves in a store where products are displayed, and the dwell time of the customers; a prediction unit that uses a model trained on a correspondence relationship between shelves on which products are displayed and the sales quantities of the products to predict an expected sales quantity on the assumption that the target product is displayed on each of the candidate shelves; and an output unit that outputs shelf information indicating a shelf predicted to achieve the target sales quantity of the target product on the basis of the expected sales quantities.

Description

Shelf allocation support device, shelf allocation support method, and recording medium

[0001] The present disclosure relates to a shelf layout support device and the like.

[0002] In stores, shelf layouts are often manually determined by store or manufacturer personnel, which can be a time-consuming process. For example, there is a technology that generates multiple shelf layout candidates that represent effective display states for sales of specific products based on analysis results of specific products and sales data of the specific products (see, for example, Patent Document 1).

[0003] International Publication No. 2016 / 199405

[0004] However, some products have a predetermined target sales volume. The products must be displayed in a store so as to meet the target sales volume. Therefore, it can still be difficult to determine the display positions of the products.

[0005] An example of an object of the present disclosure is to provide a shelf allocation support device that facilitates determining product display positions in a store.

[0006] A shelf allocation support device in one aspect of the present disclosure includes a determination means for determining candidate shelves for displaying a target product based on people flow information including at least one of the number of customers passing by each of a plurality of shelves in a store where the product is displayed and the length of time that customers stay at each shelf; a prediction means for predicting the predicted sales volume when the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the product is displayed and the sales volume of the product; and an output means for outputting shelf information indicating the shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume.

[0007] A shelf allocation support device in one aspect of the present disclosure includes a prediction means for predicting a predicted sales volume when a target product is displayed on each of the multiple shelves using a model that has learned a correspondence between multiple shelves in a store on which the product is displayed, and people flow information including at least one of the number of customers passing in front of each of the multiple shelves and the length of time that customers stay in front of the multiple shelves, and the sales volume of the product, and an output means for outputting shelf information indicating a shelf in the store from the multiple shelves that will meet the target sales volume of the target product, based on the predicted sales volume.

[0008] In one aspect of the present disclosure, a shelf allocation support method is implemented by a computer that determines candidate shelves for displaying a target product based on people flow information including at least one of the number of customers passing by each of a plurality of shelves in a store where the product is displayed and the length of time that customers stay there, predicts the predicted sales volume if the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the product is displayed and the sales volume of the product, and outputs shelf information that indicates the shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume.

[0009] In one aspect of the present disclosure, a shelf allocation support method is implemented by a computer that uses a model that has learned a correspondence between multiple shelves in a store where products are displayed and people flow information including at least one of the number of customers passing in front of each of the multiple shelves and the length of time customers stay in front of each of the multiple shelves, and the sales quantity of the product, to predict a predicted sales quantity when a target product is displayed on each of the multiple shelves, and performs a process of outputting shelf information from the multiple shelves that indicates a shelf in the store that will meet the target sales quantity of the target product based on the predicted sales quantity.

[0010] A program in one aspect of the present disclosure causes a computer to perform a process of determining candidate shelves for displaying a target product based on people flow information including at least one of the number of customers passing by each of multiple shelves in a store where the product is displayed and the length of time that customers stay there, predicting the predicted sales volume if the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the product is displayed and the sales volume of the product, and outputting shelf information indicating the shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume.

[0011] A program in one aspect of the present disclosure causes a computer to execute a process that predicts a predicted sales volume when a target product is displayed on each of multiple shelves using a model that has learned a correspondence between multiple shelves in a store on which the product is displayed, and people flow information including at least one of the number of customers passing in front of each of the multiple shelves and the length of time that customers stay in front of each of the multiple shelves, and the sales volume of the product, and outputs shelf information from the multiple shelves that indicates which shelves in the store will meet the target sales volume of the target product based on the predicted sales volume.

[0012] Each program may be stored in a non-transitory computer-readable recording medium.

[0013] According to the present disclosure, it is possible to facilitate the determination of product display positions in a store.

[0014] 1. An example of a connection between a shelf allocation support device and other devices will be briefly described. An explanatory diagram showing an example of a plurality of shelves in a store. A block diagram showing an example of the configuration of a shelf allocation support device. An explanatory diagram showing an example of predicted sales quantities for each candidate shelf. An explanatory diagram showing an example of display of shelf information. A flowchart showing an example of an operation of the shelf allocation support device. A block diagram showing an example of the configuration of a shelf allocation support device. An explanatory diagram showing an example of a product DB. An explanatory diagram showing an example of a model for each sales period. An explanatory diagram showing an example of learning data for year-round products. An explanatory diagram showing an example of learning data for seasonal products. An explanatory diagram showing an example of a model for each product category. A flowchart showing an example of an operation of the shelf allocation support device. A block diagram showing an example of the configuration of a shelf allocation support device. An explanatory diagram showing an example of predicted sales quantities for each shelf. A flowchart showing an example of an operation of the shelf allocation support device. A block diagram showing an example of the configuration of a shelf allocation support device ...

[0015] Hereinafter, with reference to the drawings, embodiments of a shelf allocation support device, a shelf allocation support method, a program, and a non-transitory recording medium for recording the program according to the present disclosure will be described in detail. The disclosed technology is not limited to these embodiments.

[0016] The shelf allocation support device assists in determining the display positions of products in a store. In particular, the shelf allocation support device uses, for example, information about the flow of people in the store to determine on which shelf in the store a product should be placed.

[0017] First Embodiment A first embodiment will be described in detail with reference to the drawings. In the first embodiment, an example will be described in which candidate shelves for displaying target products are determined based on store traffic information, and then a learning model is used to identify from the candidate shelves those shelves that are expected to achieve a target sales volume.

[0018] FIG. 1 briefly illustrates an example of a connection between a shelf allocation support device and another device. For example, the shelf allocation support device 10 is connected to a terminal device 11 used by a user via a communication network. The type of terminal device 11 is not particularly limited, and may be a personal computer (PC), a smartphone, a tablet device, or the like. Note that the terminal device 11 may be pre-installed with an application program that is capable of transmitting information to the shelf allocation support device 10 and outputting information from the shelf allocation support device 10. Alternatively, the terminal device 11 may access a website connected to the shelf allocation support device 10 and transmit and receive information to and from the shelf allocation support device 10 via the website.

[0019] Next, for ease of explanation, an example of multiple shelves in a store is shown in Figure 2. Figure 2 is an explanatory diagram showing an example of multiple shelves in a store. In Figure 2, for example, multiple shelves S1 to S19 are installed in the store.

[0020] 3 is a block diagram showing an example of the configuration of a shelf allocation support device 10. In FIG. 3, the shelf allocation support device 10 includes a determination unit 102, a prediction unit 104, and an output unit 106.

[0021] The determining unit 102 determines candidate shelves for displaying target products based on the people flow information. The target products are not particularly limited as long as they are products sold in the store. For example, the target products may be specific products such as sale products, featured products, or seasonal products.

[0022] The people flow information includes at least one of the number of customers passing by each of multiple shelves in a store where products are displayed and the length of time the customers stay. The products may be, for example, products that have been sold in the past or products that are currently displayed on the shelves. For example, the people flow information is obtained from flow line information that represents the movement lines of customers based on images captured by a camera installed in the store. For example, the number of customers may be expressed by the number of customers or the degree of the number of customers, and is not particularly limited. The length of time that customers stay may be expressed by the duration of stay or the degree of stay, and is not particularly limited. The people flow information may be statistical people flow information for a predetermined period. The predetermined period may be one hour, one day, one week, etc. The statistical people flow information may be, but is not limited to, maximum values, minimum values, average values, median values, modes, etc. The people flow information may also include the direction of customer movement. For example, the direction of customer movement may be a left turn, a right turn, etc., and the method of expressing the direction of customer movement is not particularly limited.

[0023] Here, it is expected that product sales will change depending on the movement of customers within the store. In particular, for example, it is expected that products will be more likely to catch customers' eyes on shelves that are passed by many customers or on shelves where customers tend to stay. Therefore, if a product is on a shelf that is more likely to catch customers' eyes, it is more likely that the target sales volume for the product will be achieved.

[0024] Therefore, as a specific method for determining shelf candidates, the determination unit 102 determines, for example, based on people flow information, shelves among a plurality of shelves that are passed by a large number of customers as shelf candidates. More specifically, the determination unit 102 may determine shelves among which a predetermined number of customers pass as shelf candidates. Alternatively, the determination unit 102 may determine a predetermined number of shelves in descending order of the number of customers passing as shelf candidates. Note that the determination unit 102 may determine, among shelves among which a predetermined number of customers pass as shelf candidates, a predetermined number of shelves in descending order of the number of customers passing as candidate shelves.

[0025] As another method for determining specific shelf candidates, the determination unit 102 determines, for example, a shelf among a plurality of shelves where customers stay for the longest as a shelf candidate based on people flow information. More specifically, the determination unit 102 may determine, as a shelf candidate, a shelf where passing customers stay for a predetermined length or more. Alternatively, the determination unit 102 may determine, as shelf candidate, a predetermined number of shelves in descending order of the length of customer stay. Note that the determination unit 102 may determine, as shelf candidate, a predetermined number of shelves in descending order of the length of customer stay among shelves where passing customers stay for a predetermined length or more.

[0026] Furthermore, the determining unit 102 may determine, as a shelf candidate, a shelf among a plurality of shelves where many customers pass by and where customers stay for a long time, based on the people flow information.

[0027] Furthermore, if the target sales volume is small, the target sales volume may be achieved even if the product does not catch the eye of many customers. The target sales volume may be determined according to the product category. For this reason, the determination unit 102 may, for example, determine, as shelf candidates, a shelf among multiple shelves that is passed by few customers, based on people flow information. Note that, like shelves passed by many customers, a shelf among multiple shelves may be selected based on relative evaluation, absolute evaluation, or a combination thereof. Furthermore, the determination unit 102 may, for example, determine, as shelf candidates, a shelf among multiple shelves that is passed by customers for a short period of time, based on people flow information. Note that, like shelves passed by many customers, a shelf among multiple shelves that is passed by customers for a long period of time may be selected based on relative evaluation, absolute evaluation, or a combination thereof.

[0028] The prediction unit 104 predicts the predicted sales volume when the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the products are displayed and the sales volume of the products. For example, the model is one that has learned the correspondence using the shelves on which the products are displayed as explanatory variables and the sales volume of the products as the target variable. Using the shelves as explanatory variables may mean, for example, using the location information of the shelves in the store as explanatory variables. Note that the target sales volume and the sales volume are volumes per predetermined period, such as one day, one week, or one month. The predetermined period is not particularly limited. The prediction unit 104 inputs the candidate shelves into the model and obtains from the model the predicted sales volume when the target product is displayed on the candidate shelves.

[0029] Fig. 4 is an explanatory diagram showing an example of predicted sales quantities for each shelf candidate. For example, in Fig. 4, there are five shelf candidates: shelf S2, shelf S8, shelf S9, shelf S18, and shelf S19. The predicted sales quantity for shelf S2 is 200. The predicted sales quantity for shelf S8 is 180. The predicted sales quantity for shelf S9 is 120. The predicted sales quantity for shelf S18 is 110. The predicted sales quantity for shelf S19 is 100.

[0030] Furthermore, a plurality of models may be prepared. For example, a model may be prepared for each type of product information.

[0031] For example, a model may be prepared for each product category. Product categories are classifications that group together items of the same type or nature. For example, product categories may be broad categories such as daily necessities, food, and clothing. Taking clothing as an example, product categories may be subcategorised into shirts, pants, skirts, and so on. The prediction unit 104 uses a model corresponding to the category of the target product to predict the predicted sales volume when the target product is displayed on each of the candidate shelves.

[0032] Furthermore, specifically, for example, models may be prepared for each sales period of a product. The sales period here indicates whether the product is seasonal or year-round. Year-round products are products that are sold throughout the year, such as detergent. For example, seasonal products are products that are sold in larger quantities at specific times. For example, sunscreen is a seasonal product that is sold in larger quantities during hot seasons. The prediction unit 104 uses a model corresponding to the sales period of the target product to predict the predicted sales quantity when the target product is displayed on each of the candidate shelves.

[0033] Furthermore, models may be prepared for each product category and each sales period. An example in which multiple models are prepared will be described in detail using a second embodiment.

[0034] The output unit 106 outputs shelf information indicating shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume of the candidate shelves. This shelf may also be called a recommended shelf. If there are multiple shelves that are predicted to achieve the target sales volume of the target product, the output unit 106 may output shelf information indicating at least one of the shelves. For example, the output unit 106 may output shelf information indicating the shelf with the highest predicted sales volume. For example, the output unit 106 may output shelf information indicating the shelf with the predicted sales volume closest to the target sales volume. Note that in FIG. 4, for example, if the target sales volume of the target product is 150, the shelves that are predicted to achieve the target sales volume of the target product are shelf S2 and shelf S8.

[0035] Here, the output format of the output unit 106 is not particularly limited. For example, the output method may be recording in a database or the like, screen output to a display device, audio output to an audio output device, or print output to a printer. The display device and audio output device may be provided in the terminal device.

[0036] The output unit 106 may, for example, display the recommended shelves in an identifiable manner on a map of the store. The output unit 106 may also highlight the recommended shelves on the map of the store.

[0037] Furthermore, the output unit 106 may output the shelf information and the product information of the target product in association with each other.

[0038] An example of the output unit 106 displaying shelf information on a terminal device will be described using Fig. 5. Fig. 5 is an explanatory diagram showing an example of shelf information display. In Fig. 5, recommended shelves are highlighted on a shelf map on the screen of the terminal device. In Fig. 5, the screen displays product information such as the manufacturer and product name of the target product, an image of the product, shelf information indicating the recommended shelf, the target sales quantity, and the predicted sales quantity if the product is displayed on the recommended shelf. The user of the terminal device can confirm that the recommended shelf for the target product is S8.

[0039] The screen further displays a "Go to shelf allocation determination" button. For example, when the "Go to shelf allocation determination" button is pressed, the display position of the target product on the recommended shelf may be determined.

[0040] (Flowchart) Figure 6 is a flowchart showing an example of the operation of the shelf allocation support device 10. First, the determination unit 102 determines, from among a plurality of shelves, shelf candidates for displaying target products based on people flow information (step S101). For example, the determination unit 102 determines, from among the plurality of shelves, a shelf that is passed by many customers as a shelf candidate based on the people flow information. For example, the determination unit 102 determines, from among the plurality of shelves, a shelf where customers stay for a long time as a shelf candidate based on the people flow information.

[0041] Next, the prediction unit 104 predicts the predicted sales volume of the target product for each candidate shelf using a model that has learned the correspondence between the shelves where the products are displayed and the sales volume of the product (step S102).The output unit 106 outputs shelf information indicating the shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume of the target product for each candidate shelf (step S103).

[0042] Store shelf layouts are determined manually by store staff or manufacturers, which can be time-consuming. For example, manufacturers may determine shelf layouts arbitrarily in order to sell more of their products. As a result, the appropriateness of the determined shelf layouts may not be clear. Also, as mentioned above, when selling a product, a target sales volume may be predetermined. Furthermore, it may not be clear on which shelf the product should be displayed to meet the target sales volume.

[0043] As mentioned above, product sales are expected to change depending on customer movement within a store. Therefore, in this embodiment, the shelf allocation support device 10 determines candidate shelves on which to display target products based on people flow information. This makes it possible to narrow down the candidate shelves using the people flow information. The shelf allocation support device 10 then predicts the expected sales volume of the target product for each candidate shelf using a model that has learned the correspondence between the shelves on which the products are displayed and the sales volume of the products. The shelf allocation support device 10 outputs shelf information that indicates shelves that will meet the target sales volume of the target product. This makes it easier to determine the display positions of products in a store.

[0044] Furthermore, the shelf allocation support device 10 may determine, based on people flow information, a shelf among a plurality of shelves that has a large number of customers as a shelf candidate. The shelf allocation support device 10 may also determine, among a plurality of shelves, a shelf on which customers spend a long time as a shelf candidate. In this way, it is possible to narrow down the shelf candidates to shelves that are likely to catch the eye. This makes it easier to determine shelves on which to display products in a store so that the target sales volume of the products can be achieved.

[0045] (Second embodiment) A second embodiment will be described in detail with reference to the drawings. In the second embodiment, a learning stage in which a model is learned will be described. In the second embodiment, an example of identifying a target product will also be described. Below, to the extent that the description of the second embodiment is not unclear, descriptions of content that overlaps with the above description will be omitted.

[0046] An example of the connection between the shelf allocation support device and other devices may be the same as the example described with reference to FIG. 1, and therefore a detailed description thereof will be omitted.

[0047] 7 is a block diagram showing an example of the configuration of the shelf allocation support device 20. The shelf allocation support device 20 includes a determination unit 202, a prediction unit 204, an output unit 206, a learning unit 208, a target product identification unit 210, an acquisition unit 212, a shelf allocation information generation unit 214, and a product determination unit 216.

[0048] The determination unit 202 may have the determination unit 102 shown in Fig. 3 as a basic function. The prediction unit 204 may have the prediction unit 104 shown in Fig. 3 as a basic function. The output unit 206 may have the output unit 106 shown in Fig. 3 as a basic function.

[0049] The shelf allocation support device 20 may include a product DB 2000. The product DB 2000 manages product information for each product.

[0050] FIG. 8 is an explanatory diagram showing an example of product DB 2000. In FIG. 8, product DB 2000 stores, for each product, information such as a product ID (identifier) ​​that identifies the product, product name, product price, product category, sales period, manufacturer, and target sales quantity. The product ID is an example of identification information for identifying a product. Also, for example, if a product in product DB 2000 is a year-round product, the sales period is set to be year-round. For example, if a product in product DB 2000 is a seasonal product, the sales period is set to be a season or period when the product is more popular.

[0051] In FIG. 8, the product name of the product identified by product ID "P001" is "sunscreen zz." The product price is 700 yen. The product category is daily necessities. The sales period is from June to September or a season such as summer. The manufacturer is Company A. The target sales quantity is 150 units.

[0052] In FIG. 8, the product name of the product identified by the product ID "P002" is eye drops yz. The product price is 500 yen. The product category is daily necessities. Eye drops yz are available all year round, so the sales period is all year round. The manufacturer is Company B. The target sales quantity is 100 units.

[0053] Next, the learning stage and the prediction stage will be described for each functional unit.

[0054] <Learning Stage> First, the learning stage will be described.

[0055] The learning unit 208 generates a learning model by learning the correspondence between the shelves on which products are displayed and the sales quantities of those products. More specifically, the learning unit 208 learns the correspondence between the shelves and the sales quantities by using the shelves on which products are displayed as explanatory variables and the sales quantities of those products as objective variables. This explanatory variable may be information about the location of the shelves in the store.

[0056] A plurality of models may be trained. For example, as described in the first embodiment, a model may be trained for each product sales period.

[0057] 9 is an explanatory diagram showing an example of models for different sales periods. For example, in FIG. 9, the models are a learning model for year-round products, a learning model for spring, a learning model for summer, a learning model for autumn, and a learning model for winter. Note that there may also be a learning model for spring / summer, a learning model for autumn / winter, a learning model for winter / spring, and a learning model for summer / winter.

[0058] 10A is an explanatory diagram showing an example of learning data for year-round products. For example, in FIG. 10A, the horizontal axis represents time and the vertical axis represents sales volume. Data for the year may be used as the learning data for a learning model for year-round products.

[0059] 10B is an explanatory diagram showing an example of training data for seasonal products. For example, in FIG. 10B, the horizontal axis represents time and the vertical axis represents sales volume. As an example of seasonal products, spring and summer data may be used as training data for a learning model for spring and summer.

[0060] The example of the period of data used as learning data is merely an example and is not particularly limited.

[0061] The learning unit 208 may learn the correspondence between the shelves on which the year-round products are displayed and the sales volume of the year-round products for the year-round products. The learning unit 208 may learn the correspondence between the shelves on which the seasonal products are displayed and the sales volume of the seasonal products for the year-round products for the year-round products. The learning unit 208 may also learn the correspondence between the shelves on which the seasonal products are displayed and the sales volume of the seasonal products for the year-round products for the year-round products. The learning unit 208 may identify the sales period of each product using the product DB 2000.

[0062] As another example of training a plurality of models, a model may be trained for each product category, as described in the first embodiment.

[0063] 11 is an explanatory diagram showing examples of models for each product category. For example, in FIG. 11, the models include a learning model for food, a learning model for daily necessities, and a learning model for cosmetics.

[0064] The learning unit 208 learns the correspondence between the shelves on which products classified in a category are displayed and the sales quantities of the products, for each category. The learning unit 208 can identify the category of each product using the product DB 2000.

[0065] In this way, models may be prepared for each item of product information, not just for the item category or the item sales period. Also, models may be prepared for each item category and item sales period.

[0066] The learning unit 208 may also perform learning using product information such as the product sales period and product category as explanatory variables and the product sales quantity as a response variable.

[0067] For example, the model may be trained using the sales period of the product as an explanatory variable. Specifically, for example, the learning unit 208 may train the learning model using the sales period of the product and the shelf location of the product as explanatory variables and the sales quantity of the product as a target variable.

[0068] The model may also be trained using the product category as an explanatory variable. Specifically, for example, the learning unit 208 may train the learning model using the product category and the product shelf as explanatory variables and the product sales quantity as a target variable.

[0069] Furthermore, there are multiple positions within a shelf. In some cases, the display position of a target product on the shelf needs to be determined. Therefore, the learning unit 208 may generate a learning model for determining the position on the shelf by learning the correspondence between each position on the shelf and the sales quantity of the product displayed at that position for each shelf.

[0070] In addition, there may be cases where products are determined to be displayed on the same shelf as the target product. Therefore, for example, the learning unit 208 may generate a learning model for determining products to be displayed on the same shelf by learning the correspondence between the product information and sales quantity of the product and the product information of other products displayed on the same shelf as the product. The product information and sales quantity of the product are explanatory variables, and the product information of the other products displayed on the same shelf as the product is the objective variable.

[0071] <Prediction Stage> Next, the prediction stage will be described.

[0072] The products for which the shelf allocation is determined may be designated by a user, or may be identified based on a future sales schedule, etc. For example, the products may be designated by a user using an input device such as a terminal device.

[0073] A store may have an annual sales promotion schedule determined by month, week, season, or the like. For example, seasonal products may include cold weather products in February, pollen products in March, and heat protection products, insect repellent products, and sunscreen products in summer. In such cases, a sales shelf may be prepared that collects seasonal products, featured products, and sale products. Therefore, the target product identification unit 210 refers to sales promotion schedule information that defines sales shelves by season, and identifies products to be promoted as target products. For example, the sales promotion schedule information may define a theme for the sales shelves or a type of product, such as sunscreen. For example, the target product identification unit 210 may identify products that fit the theme as target products. Furthermore, for example, the target product identification unit 210 may identify products that are the same as or similar to products that have been displayed on shelves with the same theme in the past as target products. Whether products are similar may be determined, for example, based on product features or product type. The product DB 2000 may store information for each product that can identify the characteristics and type of the product. For example, if the theme of the current sale is sunscreen, the target product identification unit 210 may identify, as the target product, the same product as a product previously displayed under the same theme in the product DB 2000 or a different sunscreen product.

[0074] Next, the target sales volume may be determined at the headquarters or at the store.

[0075] For example, when a target sales volume is determined for a store, the target sales volume may be the number of items that each store wants to sell, the number of items to order, or the like.

[0076] On the other hand, when the headquarters decides on the target sales volume, the headquarters may have agreed with the manufacturer on the sales volume. For example, if the target product is a promotional product decided by the headquarters, the headquarters is expected to decide the target sales volume. The target sales volume may be the sales volume that serves as the basis for a rebate from the manufacturer. The rebate is the amount paid by the manufacturer to the retailer.

[0077] The acquisition unit 212 acquires the target sales quantity of the target product. As a specific example of acquisition, the acquisition unit 212 acquires the target sales quantity by accepting input of the target sales quantity. For example, the acquisition unit 212 acquires the target sales quantity by accepting input of the target sales quantity via a terminal device. Note that at this time, the acquisition unit 212 may accept input of both the designation of the target product and the target sales quantity.

[0078] As another specific example of acquisition, the acquisition unit 212 acquires the target sales quantity of the target product from the product DB 2000. The target sales quantity in the product DB 2000 may be determined by the headquarters or the store.

[0079] As another specific example of acquisition, the acquisition unit 212 may acquire a target sales quantity in cooperation with order information or purchase information. For example, the acquisition unit 212 may acquire the order quantity of the target product as the target sales quantity. For example, the acquisition unit 212 may acquire the number obtained by subtracting a predetermined number from the order quantity as the target sales quantity.

[0080] As described in the first embodiment, the determining unit 202 determines candidates for shelves on which to display target products based on people flow information.

[0081] As described in the first embodiment, the prediction unit 204 uses a model that has learned the correspondence between the shelf on which the product is displayed and the sales volume of the product, to predict the expected sales volume when the target product is displayed on each of the candidate shelves.

[0082] For example, if there are models for different product sales periods, the prediction unit 204 may predict the predicted sales quantity for each candidate shelf using a model corresponding to the sales period of the target product in the product DB 2000. For example, if the sales period of the target product is spring / summer, the prediction unit 204 may acquire the predicted sales quantity for each candidate shelf from the learning model for spring / summer.

[0083] For example, if there is a model for each product category, the prediction unit 204 may predict the predicted sales quantity for each shelf candidate using a model corresponding to the category of the target product in the product DB 2000. For example, if the category of the target product is daily necessities, the prediction unit 204 may predict the predicted sales quantity for each shelf candidate from a learning model for daily necessities.

[0084] In this way, when there is a model for each item of product information, the prediction unit 204 can predict the predicted sales quantity for each shelf candidate using a model corresponding to the product information of the target product in the product DB 2000.

[0085] For example, if the sales period is included as an explanatory variable, the prediction unit 204 may input the sales period of the target product in the product DB 2000 into the model and predict the predicted sales quantity for each candidate shelf. For example, if the sales period of the target product is spring / summer, the prediction unit 204 may input spring / summer as the sales period into the model and obtain the predicted sales quantity for each candidate shelf from the model.

[0086] For example, if a category is used as an explanatory variable, the prediction unit 204 may input the category of the target product in the product DB 2000 into the model and predict the predicted sales quantity for each candidate shelf. For example, if the category of the target product is daily necessities, the prediction unit 204 may input the daily necessities as a category into the model and obtain the predicted sales quantity for each candidate shelf from the model.

[0087] For example, when product information is used as an explanatory variable, the prediction unit 204 may input product information of the target product in the product DB 2000 into the model and predict the predicted sales quantity for each candidate shelf.

[0088] As explained in the first embodiment, the output unit 206 outputs shelf information indicating shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume of the candidate shelves.

[0089] The shelf allocation information generating unit 214 also generates shelf allocation information for when the target product is displayed on a shelf that is predicted to achieve the target sales volume of the target product. For example, when creating a shelf allocation for a themed shelf, products that fit the shelf theme, like the target product, are displayed on the same shelf as the target product.

[0090] Therefore, the shelf allocation information generating unit 214 generates shelf allocation information for the case where the target product is displayed on a shelf that is predicted to achieve the target sales volume of the target product.

[0091] Furthermore, there are multiple positions within the recommended shelf. Therefore, the prediction unit 204 predicts the predicted sales volume when the target product is displayed at each position on the recommended shelf. For example, the prediction unit 204 predicts the predicted sales volume of the target product when it is displayed at each position on the recommended shelf using a learning model that has learned the correspondence between each position on the recommended shelf and the sales volume of the product displayed at that position.

[0092] Then, the shelf allocation information generating unit 214 generates shelf allocation information when the target product is displayed on the recommended shelf, based on the predicted sales volume at each position on the recommended shelf.

[0093] Furthermore, a priority may be assigned to each product. Therefore, the prediction unit 204 predicts the predicted sales volume of other products to be displayed on the recommended shelf, based on their positions on the recommended shelf. The shelf allocation information generation unit 214 then generates shelf allocation information based on the predicted sales volume of each position on the recommended shelf for the target product and other products, and the priorities set for the target product and other products, so that products with higher priorities are placed in positions on the recommended shelf where they will have the highest predicted sales volume. Note that the method for setting priorities is not particularly limited, and priorities may be set, for example, by a manufacturer.

[0094] Furthermore, the other products displayed on the same shelf as the target product may be products similar to the target product. Furthermore, the other products may be products that fall under the same theme as the target product. Furthermore, the other products may be products that sell better when displayed together with the target product. The product determination unit 216 may determine, as the product to be displayed together with the target product, products that sell better when displayed together with the target product. The product determination unit 216 may acquire product information about the other products by inputting the product information and the target sales quantity of the target product into a model in which the product information and the product sales quantity of the product are used as explanatory variables and the product information of the other products displayed on the same shelf as the product is used as a target variable.

[0095] The output unit 206 outputs the shelf allocation information. The output format may be the same as the example of outputting shelf information, and is not particularly limited.

[0096] Furthermore, the output unit 206 may output shelf information indicating shelves in a store and shelf allocation information indicating the position of each product on the shelf in association with each other. For example, the output unit 206 may highlight a shelf in the store on which the target product is to be displayed on a map of the store, and output shelf allocation information for the shelf when a shelf on the map is selected.

[0097] In addition to the shelf information and shelf allocation information, the output unit 206 may also output at least some of information such as people flow information, sales volume of the target product when it was sold in the past, sales promotion schedule, rebate information, and priority.

[0098] 12 is a flowchart showing an example of the operation of the shelf allocation support device 20. The target product identification unit 210, for example, identifies a target product (step S201).

[0099] The acquisition unit 212 acquires the target sales quantity (step S202). The determination unit 202 determines candidates for shelves on which to display the target products based on the people flow information (step S203).

[0100] The prediction unit 204 predicts the predicted sales quantity of the target product for each shelf candidate (step S204). The prediction unit 204 determines whether there is a shelf that is predicted to achieve the target sales quantity of the target product (step S205). If there is no shelf that is predicted to achieve the target sales quantity of the target product (step S205: No), the determination unit 202 returns to step S203. At this time, the determination unit 202 determines a new shelf candidate for the target product from shelves that are not candidate shelves based on people flow information.

[0101] If all shelves are candidate shelves and no shelf satisfies the predicted sales volume, the output unit 206 may output that no shelf satisfies the predicted sales volume. Also, the target sales volume may be reset.

[0102] If there is a shelf that is predicted to achieve the target sales quantity of the target product (Step S205: Yes), the prediction unit 204 predicts the predicted sales quantity when the target product is displayed at each position on the recommended shelf (Step S206). Then, the shelf allocation information generation unit 214 generates shelf allocation information (Step S207). The output unit 206 outputs the shelf allocation information (Step S208), and the shelf allocation support device 20 ends the series of processes shown in FIG.

[0103] In step S206, the prediction unit 204 may predict the predicted sales volume of other products to be displayed together with the target product on the recommended shelf when they are displayed at each position on the recommended shelf. For example, the other products are products similar to the target product. In step S207, for example, the shelf allocation information generation unit 214 generates shelf allocation information based on the predicted sales volume of each position on the recommended shelf and the priorities set for the target product and similar products to the target product, so that products with higher priorities are placed at positions on the shelf where they will have the highest predicted sales volume.

[0104] As described above, in this embodiment, the shelf allocation support device 20 generates a learning model by learning the correspondence between the shelf on which a product is displayed and the sales volume of that product. This makes it possible to predict the sales volume of a target product based on the sales volume of products sold on each shelf in the past.

[0105] Furthermore, multiple models may be prepared for determining shelf positions in a store. For example, the models may be prepared by product category. The shelf allocation support device 20 predicts the predicted sales quantity for each shelf candidate using one of the multiple models that corresponds to the category of the target product. Furthermore, the models may be prepared by product sales period. The shelf allocation support device 20 predicts the predicted sales quantity for each shelf candidate using one of the multiple models that corresponds to the sales period of the target product. Furthermore, models may be prepared by product category and sales period.

[0106] The model may also be trained using product information as an explanatory variable. For example, the model may also be trained using the product category as an explanatory variable. The shelf allocation support device 20 inputs the category of the target product into the model, and predicts the predicted sales quantity when the target product is displayed on each of the candidate shelves. The model may also be trained using the product sales period as an explanatory variable. The shelf allocation support device 20 inputs the target product sales period into the model, and predicts the predicted sales quantity when the target product is displayed on each of the candidate shelves.

[0107] In order to sell a variety of products, it is necessary not only to focus on the target product and increase its sales volume, but also to consider the sales volumes of other products. Therefore, in this embodiment, the shelf allocation support device 20 predicts the predicted sales volume when the target product is displayed at each position on the recommended shelf, and generates shelf allocation information such that target products with high priority are placed at positions on the recommended shelf where their predicted sales volume will be higher, based on the predicted sales volume of each recommended position and the priorities set for the target product and similar products similar to the target product. In this way, shelf allocation information can be generated such that target products with high priority have a higher predicted sales volume.

[0108] (Third Embodiment) A third embodiment will be described in detail with reference to the drawings. In the first embodiment, an example will be described in which shelves that are expected to achieve a target sales volume are identified using a learning model that has learned information about the flow of people at each shelf in a store and the sales volume of products displayed on each shelf. Below, explanations that overlap with the above explanations will be omitted to the extent that the explanation of the third embodiment is not unclear.

[0109] In this embodiment, an example of connection between the shelf allocation support device and other devices may be the same as the example described with reference to FIG. 1, and therefore detailed description thereof will be omitted.

[0110] 13 is a block diagram showing an example of the configuration of the shelf allocation support device 30. In FIG. 13, the shelf allocation support device 30 includes a prediction unit 304 and an output unit 306.

[0111] The prediction unit 304 predicts the predicted sales volume of a target product when it is displayed on each of multiple shelves, using a model that has learned the correspondence between the sales volume of the product and the multiple shelves and people flow information in the store where the product is displayed. The people flow information includes at least one of the number of customers passing in front of each of the multiple shelves, the length of time that customers stay in front of each shelf, and the direction of customer movement in the store. The model has learned the correspondence using the location information of the shelf where the product is displayed and the people flow information at that shelf location as explanatory variables, and the sales volume of the product displayed on that shelf as a target variable. Note that using the shelf as an explanatory variable may also mean using the shelf location information as an explanatory variable.

[0112] Specifically, for example, the prediction unit 304 inputs planned people flow information into the model and predicts the predicted sales volume when the target product is placed on each of multiple shelves. The planned people flow information may be people flow information determined by a user. For example, the planned people flow information may be statistical people flow information for the same period as the sales period of the target product.

[0113] 14 is an explanatory diagram showing an example of predicted sales quantities by shelf. The predicted sales quantities for shelf S1 and shelf S2 are 210. The predicted sales quantities for shelves S3 to S7, shelves S10 to S15, and shelf S19 are each 100. The predicted sales quantity for shelf S8 is 170. The predicted sales quantity for shelf S9 is 120. The predicted sales quantity for shelf S18 is 110.

[0114] 14, an example has been described in which the prediction unit 304 predicts the predicted sales quantity for all shelves. For example, the prediction unit 304 may select some of the shelves as shelf candidates and predict the predicted sales quantity for each of the shelf candidate. The selected shelves may be predetermined based on the product category.

[0115] Furthermore, as in the first and second embodiments, a model may be prepared for each item of product information. Specifically, for example, a model may be prepared for each item category. The prediction unit 304 uses a model corresponding to the category of the target item to predict the predicted sales quantity when the target item is displayed on each of a plurality of shelves. Furthermore, for example, a model may be prepared for each item sales period. The prediction unit 304 uses a model corresponding to the sales period of the target item to predict the predicted sales quantity when the target item is displayed on each of a plurality of shelves. Furthermore, a model may be prepared for each item category and sales period.

[0116] The output unit 306 outputs shelf information indicating shelves in the store that will satisfy the target sales quantity of the target product from among multiple shelves based on the predicted sales quantity. In Fig. 14, when the target sales quantity is 150, the shelves that will satisfy the target sales quantity of the target product are shelf S1, shelf S2, and shelf S8. The output format by the output unit 306 may be similar to that of the output unit 306 shown in Fig. 3 and the output unit 306 shown in Fig. 7, and is not particularly limited. For example, an example output screen on which the output unit 306 outputs shelf information to a terminal device may be the same as the example shown in Fig. 5.

[0117] As in the first and second embodiments, when there are multiple shelves that are predicted to achieve the target sales volume of the target product, the output unit 306 may output shelf information that indicates at least one of the shelves. For example, the output unit 306 may output shelf information that indicates the shelf with the highest predicted sales volume. For example, the output unit 306 may output shelf information that indicates the shelf with the predicted sales volume that is close to the target sales volume.

[0118] 15 is a flowchart showing an example of the operation of the shelf allocation support device 30. The prediction unit 304 predicts the predicted sales quantity of a target product for each shelf using a model that has learned the correspondence between the sales quantity of the product and people flow information, which includes at least one of a plurality of shelves in a store where the products are displayed and the number of customers passing in front of each of the plurality of shelves and the length of time that customers stay in front of each shelf (step S301).

[0119] The output unit 306 outputs shelf information indicating shelves that are predicted to achieve the target sales volume of the target product (step S302), and the shelf allocation support device 30 ends the series of processes shown in FIG.

[0120] It is expected that the sales volume of a product will vary depending on the movement of customers in a store. In particular, it is expected that the sales volume will vary depending on the number of customers passing in front of a shelf, the length of time customers stay in front of the shelf, the direction of customer movement in the store, and other factors. For example, taking the number of customers as an example, the shelf allocation support device 30 may predict how many products will be sold if a certain number of people pass by each shelf. Therefore, the shelf allocation support device 30 uses a model in which shelf and shelf traffic information are used as explanatory variables and the sales volume of products displayed on the shelf is used as a target variable to predict the expected sales volume when target products are displayed on each shelf, and outputs shelf information indicating the shelves that are expected to achieve the target sales volume of the target products. This makes it possible to predict sales volumes based on the traffic flow information. It is also possible to propose product displays that will more likely achieve the target sales volume. This makes it easier to determine the display positions of products in a store.

[0121] (Fourth embodiment) A fourth embodiment will be described in detail with reference to the drawings. In the fourth embodiment, a learning stage in which a model is further learned in addition to the third embodiment will be described. Below, descriptions of content that overlaps with the above description will be omitted to the extent that the description of the fourth embodiment is not unclear.

[0122] In this embodiment, an example of connection between the shelf allocation support device and other devices may be the same as the example described with reference to FIG. 1, and therefore detailed description thereof will be omitted.

[0123] 16 is a block diagram showing an example of the configuration of the shelf allocation support device 40. The shelf allocation support device 40 includes a determination unit, a prediction unit 404, an output unit 406, a learning unit 408, a target product identification unit 410, an acquisition unit 412, a shelf allocation information generation unit 414, and a product determination unit 416.

[0124] The prediction unit 404 may have the basic function of the prediction unit 304 shown in Fig. 13. The output unit 406 may have the basic function of the output unit 306 shown in Fig. 13.

[0125] The target product identification unit 410 may have the target product identification unit 210 shown in Fig. 7 as a basic function. The acquisition unit 412 may have the acquisition unit 212 shown in Fig. 7 as a basic function. The shelf allocation information generation unit 414 may have the shelf allocation information generation unit 214 shown in Fig. 7 as a basic function. The product determination unit 416 may have the product determination unit 216 shown in Fig. 7 as a basic function.

[0126] The shelf allocation support device 40 may further include a product DB 4000. The product DB 4000 may be similar to the product DB 2000 shown in FIG.

[0127] Next, the learning stage and the prediction stage will be described for each functional unit.

[0128] <Learning Stage> First, the learning stage will be described.

[0129] The learning unit 408 learns the correspondence between the shelves where products are displayed, the people flow information, and the sales volume of the products. More specifically, the learning unit 408 learns the correspondence between the shelves where products are displayed, the people flow information, and the sales volume, using the shelves where products are displayed and the people flow information as explanatory variables and the sales volume of the products as a target variable. This explanatory variable may be the position information of the shelves in the store.

[0130] A plurality of models may be trained. For example, as described in the first, second, and third embodiments, a model may be trained for each item of product information. A model may be trained for each item category. A model may be trained for each item sales period. A model may be trained for each item category and each item sales period. For this reason, the training unit 408 trains a plurality of models.

[0131] Furthermore, as described in the first, second, and third embodiments, for example, the model may further learn product information as an explanatory variable. For example, the learning unit 408 may learn the correspondence between a plurality of shelves, people flow information, and product information, and the sales quantity of a product. If the product information is a product category, the learning unit 408 learns the correspondence between a plurality of shelves, people flow information, and product category as explanatory variables and the sales quantity of the product as the objective variable. If the product information is a product sales period, the learning unit 408 learns the correspondence between a plurality of shelves, people flow information, and product sales period as explanatory variables and the sales quantity of the product as the objective variable.

[0132] As described in the second embodiment, when determining the display position of a target product on a shelf, the learning unit 408 may generate a learning model for determining the position on the shelf by learning the correspondence between each position on the shelf and the sales quantity of the product displayed at that position, for each shelf.

[0133] Furthermore, as described in the second embodiment, the learning unit 408 may generate a learning model for determining products to be displayed on the same shelf by learning the correspondence between the product information and sales volume of a product and the product information of other products displayed on the same shelf as the product. The product information and sales volume of a product are explanatory variables, and the product information of the other products displayed on the same shelf as the product is a target variable.

[0134] <Prediction Stage> Next, the prediction stage will be described.

[0135] The example in which the target product specification unit 410 specifies the target product is as described in the second embodiment.

[0136] The example in which the acquisition unit 412 acquires the target sales quantity of the target product is as described in the second embodiment.

[0137] As explained in the third embodiment, the prediction unit 404 uses a model to predict the expected sales volume when the target products are displayed on each shelf.

[0138] For example, if there are models for different product sales periods, the prediction unit 404 can predict the predicted sales quantity for each shelf using a model corresponding to the sales period of the target product in the product DB 4000. For example, if the sales period of the target product is spring / summer, the prediction unit 404 can obtain the predicted sales quantity for each shelf from the learning model for spring / summer.

[0139] For example, if there is a model for each product category, the prediction unit 404 may predict the predicted sales quantity for each shelf using a model corresponding to the category of the target product in the product DB 4000. For example, if the category of the target product is daily necessities, the prediction unit 404 may acquire the predicted sales quantity for each shelf from the learning model for daily necessities for each candidate shelf.

[0140] In this way, when there is a model for each item of product information, the prediction unit 404 can predict the predicted sales quantity for each shelf using a model corresponding to the product information of the target product in the product DB 4000.

[0141] For example, if the sales period is an explanatory variable, the prediction unit 404 may input the sales period of the target product in the product DB 4000 into the model and predict the predicted sales quantity of the target product by shelf. For example, if the sales period of the target product is spring / summer, the prediction unit 404 may input spring / summer as the sales period into the model and obtain the predicted sales quantity by shelf from the model.

[0142] For example, if a category is used as an explanatory variable, the prediction unit 404 may input the category of the target product in the product DB 4000 into the model and predict the predicted sales quantity of the target product by shelf. For example, if the category of the target product is daily necessities, the prediction unit 404 may input the daily necessities as a category into the model and obtain the predicted sales quantity of the target product by shelf.

[0143] As explained in the third embodiment, the output unit 406 outputs, from among a plurality of shelves, shelf information indicating shelves in the store that will meet the target sales volume of the target product, based on the predicted sales volume.

[0144] Furthermore, if there is no shelf that satisfies the predicted sales quantity, the output unit 406 outputs a message indicating that there is no shelf that satisfies the predicted sales quantity of the target product. The target sales quantity of the target product may also be reset. The acquisition unit 412 may acquire the newly set target sales quantity. Then, the output unit 406 may output shelf information indicating shelves in the store that satisfy the new target sales quantity from among multiple shelves based on the predicted sales quantity. The output unit 406 may also output a message indicating that there is no shelf that satisfies the predicted sales quantity of the target product, along with planned people flow information. For example, the planned people flow information may be changed, and the prediction unit 404 may predict the predicted sales quantity using the changed planned people flow information.

[0145] As explained in the second embodiment, the shelf allocation information generating unit 414 generates shelf allocation information for when the target product is displayed on a shelf that is predicted to achieve the target sales volume of the target product.

[0146] The prediction unit 404 also predicts the predicted sales volume when the target product is displayed at each position on the recommended shelf that is predicted to achieve the target sales volume of the target product, and the predicted sales volume when other products to be displayed on the shelf together with the target product are displayed at each position on the recommended shelf.Then, the shelf allocation information generation unit 414 generates shelf allocation information based on the priorities set for the target product and the other products, such that products with higher priorities are displayed at positions on the recommended shelf that will result in higher predicted sales volumes.This makes it possible to generate shelf allocation information that will result in higher predicted sales volumes for target products with higher priorities.

[0147] Furthermore, other products displayed on the same shelf as the target product may be products similar to the target product, products that fall under the same theme, or products determined by the product determination unit 416, as described in the second embodiment.

[0148] 17 is a flowchart showing an example of the operation of the shelf allocation support device 40. First, the target product identification unit 410 identifies, for example, a target product (step S401).

[0149] The acquisition unit 412 acquires the target sales quantity (step S402). The prediction unit 404 predicts the predicted sales quantity of the target product for each shelf (step S203). The prediction unit 404 determines whether there is a shelf that is predicted to achieve the target sales quantity of the target product (step S403).

[0150] The prediction unit 404 predicts the predicted sales volume when the target product is displayed at each position on the recommended shelf that is predicted to achieve the target sales volume of the target product (step S404). Then, the shelf allocation information generation unit 414 generates shelf allocation information (step S405). The output unit 406 outputs the shelf allocation information (step S406), and the shelf allocation support device 40 ends the series of processes shown in FIG.

[0151] As described above, in this embodiment, the shelf allocation support device 40 generates a model by learning the correspondence between a plurality of shelves and people flow information and the sales quantity of a product. This makes it possible to predict the sales quantity of a target product based on the people flow information on each shelf and the sales quantity of a product sold on each shelf in the past.

[0152] The above is the description of each embodiment. The embodiments may be combined or modified.

[0153] Furthermore, the shelf allocation support devices 10, 20, 30, and 40 may be configured to include some of the functional units and information.

[0154] Furthermore, the embodiments are not limited to the examples described above and can be modified in various ways. Furthermore, the configuration of the shelf allocation support devices 10, 20, 30, and 40 is not particularly limited. For example, the functional units of the shelf allocation support devices 10, 20, 30, and 40 may be implemented by a single device. Alternatively, for example, each functional unit or database of the shelf allocation support devices 10, 20, 30, and 40 may be implemented by a different device and configured as a system. For example, each functional unit of the shelf allocation support device 20 may be configured by a plurality of servers and configured as a system. For example, it may be implemented by a database server including each database and a server having each functional unit.

[0155] In addition, in the embodiment, each database may include part of the information described above. Also, each piece of information may include information other than the information described above.

[0156] Furthermore, the process of generating information to be displayed on the terminal device 11 may be performed by a functional unit such as an output unit provided in the shelf allocation support devices 10, 20, 30, and 40. This process may also be performed by the terminal device 11. That is, the terminal device 11 may generate information for a screen to be displayed on the terminal device 111 based on data received from the shelf allocation support devices 10, 20, 30, and 40, and display the screen.

[0157] (Example of Computer Hardware Configuration) Next, an example of a hardware configuration in the case where each device such as the shelf allocation support devices 10, 20, 30, and 40 and the terminal device 11 is realized by a computer will be described.

[0158] 18 is an explanatory diagram showing an example of the hardware configuration of a computer. For example, some or all of the devices can be realized using any combination of a computer 80 and a program as shown in FIG.

[0159] The computer 80 includes, for example, a processor 801, a ROM (Read Only Memory) 802, a RAM (Random Access Memory) 803, and a storage device 804. The computer 80 also includes a communication interface 805 and an input / output interface 806. The components are connected to each other, for example, via a bus 807. The number of each component is not particularly limited, and there may be one or more of each component.

[0160] The processor 801 controls the entire computer 80. The processor 801 may be, for example, a central processing unit (CPU), a digital signal processor (DSP), a graphics processing unit (GPU), a physics processing unit (PPU), a tensor processing unit (TPU), a quantum processor, or a combination thereof, and is not particularly limited.

[0161] The computer 80 also includes a ROM 802, a RAM 803, and a storage device 804. Examples of the storage device 804 include semiconductor memory such as flash memory, a hard disk drive (HDD), and a solid state drive (SSD). For example, the storage device 804 stores an operating system (OS) program, application programs, and programs according to the embodiments. Alternatively, the ROM 802 stores application programs and programs according to the embodiments. The RAM 803 is used as a work area for the processor 801.

[0162] The processor 801 also loads programs stored in the storage device 804, ROM 802, etc. The processor 801 then executes each process coded in the program. The processor 801 may also download various programs via the communication network NT. The processor 801 also functions as a part or all of the computer 80. The processor 801 may then execute the processes or instructions in the illustrated flowchart based on the program.

[0163] The communication interface 805 is connected to a communication network NT such as a LAN (Local Area Network) or a WAN (Wide Area Network) via a wireless or wired communication line. The communication network NT may be composed of multiple communication networks NT. As a result, the computer 80 is connected to external devices and external computers 80 via the communication networks NT. The communication interface 805 serves as an interface between the communication network NT and the inside of the computer 80. The communication interface 805 also controls the input and output of data from external devices and external computers 80.

[0164] Furthermore, the input / output interface 806 is connected to at least one of an input device, an output device, and an input / output device. The connection method may be wireless or wired. Examples of the input device include a keyboard, a mouse, and a microphone. Examples of the output device include a display device, a lighting device, and an audio output device that outputs audio. Examples of the input / output device include a touch panel display. Note that the input device, output device, and input / output device may be built into the computer 80 or may be external.

[0165] The hardware configuration of the computer 80 is an example. The computer 80 may have some of the components shown in FIG. 18 . The computer 80 may have components other than those shown in FIG. 18 . For example, the computer 80 may have a drive device or the like. The processor 801 may then read programs and data stored on a recording medium attached to the drive device or the like into the RAM 803. Examples of non-transitory tangible recording media include optical disks, flexible disks, magneto-optical disks, and USB (Universal Serial Bus) memories. As described above, the computer 80 may have input devices such as a keyboard and a mouse. The computer 80 may have an output device such as a display. The computer 80 may also have an input device, an output device, and an input / output device.

[0166] The computer 80 may also include various sensors (not shown). The types of sensors are not particularly limited. The computer 80 may also include an imaging device capable of capturing images or videos.

[0167] This concludes the description of the hardware configuration of each device. There are various variations in the method of realizing each device. For example, each device may be realized by any combination of a different computer and program for each component. Furthermore, multiple components of each device may be realized by any combination of a single computer and program.

[0168] Furthermore, some or all of the components of each device may be realized by circuits for specific applications. Furthermore, some or all of the components of each device may be realized by general-purpose circuits such as FPGAs (Field Programmable Gate Arrays). Furthermore, some or all of the components of each device may be realized by a combination of circuits for specific applications and general-purpose circuits. These circuits may be a single integrated circuit. Alternatively, these circuits may be divided into multiple integrated circuits. The multiple integrated circuits may be connected via a bus or the like.

[0169] Furthermore, when some or all of the components of each device are realized by a plurality of computers, circuits, etc., the plurality of computers, circuits, etc. may be centrally located or distributed.

[0170] The shelf allocation support methods described in the respective embodiments may be realized by being executed by a computer such as the shelf allocation support devices 10, 20, 30, and 40.

[0171] Each program described in each embodiment is recorded on a computer-readable recording medium such as a HDD, SSD, flexible disk, optical disk, magneto-optical disk, or USB memory. Each program is executed by being read from the recording medium by a computer. Each program may also be distributed via a communication network NT.

[0172] The functions of each of the components of the shelf allocation support device 10 and the shelf allocation support device 20 described above may be realized by dedicated hardware, such as a computer. Alternatively, each component may be realized by software. Alternatively, each component may be realized by a combination of hardware and software.

[0173] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above embodiments. The configuration and details of each of the present disclosures may include embodiments to which various modifications that would be apparent to those skilled in the art are applied within the scope of the present disclosure. The present disclosure may include embodiments in which the features described herein are appropriately combined or substituted as necessary. For example, features described using a particular embodiment may also be applied to other embodiments to the extent that no contradiction occurs. For example, although multiple operations are described in sequence in the form of a flowchart, the order of description does not limit the order in which the multiple operations are performed. Therefore, when implementing the embodiments, the order of the multiple operations may be changed as long as the content is not affected.

[0174] Some or all of the above-described embodiments can be described as follows: However, some or all of the above-described embodiments are not limited to the following.

[0175] (Supplementary Note 1) A shelf allocation support device comprising: a determination means for determining candidate shelves for displaying a target product based on people flow information including at least one of the number of customers passing by each of a plurality of shelves in a store where the product is displayed and the length of time that customers stay; a prediction means for predicting the predicted sales volume when the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the product is displayed and the sales volume of the product; and an output means for outputting shelf information indicating shelves that are predicted to achieve the target sales volume of the target product, based on the predicted sales volume.

[0176] (Supplementary Note 2) A shelf allocation support device as described in Supplementary Note 1, comprising: a shelf allocation information generation means for generating shelf allocation information when the target product is displayed on the predicted shelf, wherein the prediction means predicts a predicted sales quantity when the target product is displayed at each position on the shelf, and a predicted sales quantity when other products are displayed together with the target product at each position on the shelf, and the shelf allocation information generation means generates shelf allocation information for displaying products with higher priorities on the shelf at positions that will result in higher predicted sales quantities, based on priorities set for the target product and the other products.

[0177] (Supplementary Note 3) The shelf allocation support device according to Supplementary Note 1 or 2, wherein the models are for each product category, and the prediction means predicts the predicted sales quantity for each of the shelf candidates using a model corresponding to the category of the target product from among the plurality of models.

[0178] (Supplementary Note 4) The shelf allocation support device according to any one of Supplementary Notes 1 to 3, wherein the models are divided by product sales period, and the prediction means predicts the predicted sales quantity for each of the candidate shelves using a model corresponding to the sales period of the target product from among the plurality of models.

[0179] (Supplementary Note 5) The shelf allocation support device described in any of Supplementary Notes 1 to 4, wherein the model is further trained using a product category as an explanatory variable, and the prediction means inputs the category of the target product into the model and predicts the predicted sales quantity when the target product is displayed on each of the candidate shelves.

[0180] (Supplementary Note 6) The shelf allocation support device described in any of Supplementary Notes 1 to 5, wherein the model is further trained using the sales period of the product as an explanatory variable, and the prediction means inputs the sales period of the target product into the model and predicts the predicted sales quantity when the target product is displayed on each of the candidate shelves.

[0181] (Supplementary Note 7) The shelf allocation support device according to any one of Supplementary Notes 1 to 6, wherein the determining means determines a shelf among the plurality of shelves that has a large number of customers as the shelf candidate.

[0182] (Supplementary Note 8) The shelf allocation support device according to any one of Supplementary Notes 1 to 7, wherein the determining means determines, as the shelf candidate, a shelf on which the customer has been staying for a long time among the plurality of shelves.

[0183] (Supplementary Note 9) The shelf allocation support device according to any one of Supplementary Notes 1 to 8, further comprising: a learning means that generates the model by learning a correspondence relationship between a shelf on which the product is arranged and a sales quantity of the product.

[0184] (Supplementary Note 10) A shelf allocation support device comprising: a prediction means for predicting a predicted sales volume when a target product is displayed on each of the plurality of shelves, using a model that has learned a correspondence between a plurality of shelves in a store on which the product is displayed, and people flow information including at least one of the number of customers passing in front of each of the plurality of shelves and the length of time that customers stay in front of the plurality of shelves, and the sales volume of the product; and an output means for outputting shelf information indicating a shelf in the store that meets a target sales volume for the target product from the plurality of shelves, based on the predicted sales volume.

[0185] (Supplementary Note 11) The shelf allocation support device according to Supplementary Note 10, wherein the prediction means inputs expected customer flow information into the model and predicts a predicted sales quantity when the target product is placed on each of the plurality of shelves.

[0186] (Supplementary Note 12) The shelf allocation support device according to Supplementary Note 11, wherein the expected people flow information is people flow information for a period that is the same as a period when the target product is on sale.

[0187] (Supplementary Note 13) A shelf allocation support device as described in any of Supplementary Notes 10 to 12, comprising: a shelf allocation information generation means for generating shelf allocation information when the target product is displayed on the predicted shelf, wherein the prediction means predicts a predicted sales quantity when the target product is displayed at each position on the shelf, and a predicted sales quantity when other products are displayed together with the target product at each position on the shelf, and the shelf allocation information generation means generates shelf allocation information for displaying products with higher priorities on the shelf at positions that will result in higher predicted sales quantities, based on priorities set for the target product and the other products.

[0188] (Appendix 14) The shelf allocation support device described in any of Appendices 10 to 13, wherein the models are for each product category, and the prediction means predicts the predicted sales quantity when the target product is displayed on each of the plurality of shelves using a model corresponding to the category of the target product from among the plurality of models.

[0189] (Appendix 15) The shelf allocation support device described in any of Appendices 10 to 14, wherein the models are for each sales period of the product, and the prediction means predicts the predicted sales quantity when the target product is displayed on each of the plurality of shelves using a model corresponding to the sales period of the target product from among the plurality of models.

[0190] (Supplementary Note 16) The shelf allocation support device described in any one of Supplementary Notes 10 to 13, wherein the model is further trained using a product category as an explanatory variable, and the prediction means inputs the category of the target product into the model and predicts the predicted sales quantity when the target product is displayed on each of the plurality of shelves.

[0191] (Supplementary Note 17) The shelf allocation support device described in any one of Supplementary Notes 1 to 13 and 16, wherein the model is further trained using a sales period of the product as an explanatory variable, and the prediction means inputs the sales period of the target product into the model and predicts a predicted sales quantity when the target product is displayed on each of the plurality of shelves.

[0192] (Supplementary Note 18) The shelf allocation support device according to any one of Supplementary Notes 10 to 17, further comprising: a learning means for generating the model by learning a correspondence between the plurality of shelves and the people flow information, and the sales quantity of the product.

[0193] (Supplementary Note 19) The shelf allocation support device according to any one of Supplementary Notes 1 to 18, further comprising: an acquisition unit that acquires the target sales quantity of the target product.

[0194] (Supplementary Note 20) The shelf allocation support device according to Supplementary Note 19, wherein the acquisition unit acquires the target sales quantity by accepting an input of the target sales quantity.

[0195] (Supplementary Note 21) The shelf allocation support device according to Supplementary Note 19, wherein the acquisition unit acquires an order quantity of the target product as the target sales quantity.

[0196] (Supplementary Note 22) A shelf allocation support method in which a computer performs the following processes: determines candidate shelves for displaying a target product based on people flow information including at least one of the number of customers passing by each of a plurality of shelves in a store where the product is displayed and the length of time that customers stay there; predicts the predicted sales volume if the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the product is displayed and the sales volume of the product; and outputs shelf information indicating the shelves that are predicted to achieve the target sales volume of the target product, based on the predicted sales volume.

[0197] (Supplementary Note 23) A shelf allocation support method in which a computer executes a process of: predicting a predicted sales volume when a target product is displayed on each of a plurality of shelves in a store on which a product sold in the past is displayed, using a model that has learned a correspondence between a plurality of shelves in the store on which the product has been previously sold, and people flow information including at least one of the number of customers passing in front of each of the plurality of shelves and the length of time that customers stay in front of each of the plurality of shelves, and the sales volume of the product; and outputting shelf information indicating a shelf in the store that will meet the target sales volume of the target product from the plurality of shelves based on the predicted sales volume.

[0198] (Supplementary Note 24) A non-transitory computer-readable recording medium having recorded thereon a program for causing a computer to execute the following processes: determine candidate shelves for displaying a target product based on people flow information including at least one of the number of customers passing by each of a plurality of shelves in a store where the product is displayed and the length of time that customers stay there; predict the predicted sales volume if the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the product is displayed and the sales volume of the product; and output shelf information indicating shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume.

[0199] (Supplementary Note 25) A non-transitory computer-readable recording medium having recorded thereon a program for causing a computer to execute the following process: predicting a predicted sales volume when a target product is displayed on each of a plurality of shelves using a model that has learned a correspondence between a plurality of shelves in a store on which the product is displayed, and people flow information including at least one of the number of customers passing in front of each of the plurality of shelves and the length of time customers stay in front of each of the plurality of shelves, and the sales volume of the product; and outputting shelf information indicating a shelf in the store that will meet the target sales volume of the target product from the plurality of shelves based on the predicted sales volume.

[0200] (Supplementary Note 26) A non-transitory computer-readable recording medium having recorded thereon a program for causing a computer to execute the following processes: determine candidate shelves for displaying a target product based on people flow information including at least one of the number of customers passing by each of a plurality of shelves in a store where the product is displayed and the length of time that customers stay there; predict the predicted sales volume if the target product is displayed on each of the candidate shelves using a model that has learned the correspondence between the shelves on which the product is displayed and the sales volume of the product; and output shelf information indicating shelves that are predicted to achieve the target sales volume of the target product based on the predicted sales volume.

[0201] (Supplementary Note 27) A non-transitory computer-readable recording medium having recorded thereon a program for causing a computer to execute the following process: predicting a predicted sales volume when a target product is displayed on each of a plurality of shelves using a model that has learned a correspondence between a plurality of shelves in a store on which the product is displayed, and people flow information including at least one of the number of customers passing in front of each of the plurality of shelves and the length of time customers stay in front of each of the plurality of shelves, and the sales volume of the product; and outputting shelf information indicating a shelf in the store that will meet the target sales volume of the target product from the plurality of shelves based on the predicted sales volume.

[0202] Furthermore, some or all of the configurations described in Supplementary Notes 2 to 9, Supplementary Notes 19, Supplementary Notes 20, and Supplementary Notes 21, which are dependent on Supplementary Notes 1 described above, may also be dependent on Supplementary Notes 22, Supplementary Notes 24, and Supplementary Notes 26 in the same dependent relationship as Supplementary Notes 2 to 9, Supplementary Notes 19, Supplementary Notes 20, and Supplementary Notes 21. Furthermore, some or all of the configurations described in Supplementary Notes 11 to 21, which are dependent on Supplementary Notes 10 described above, may also be dependent on Supplementary Notes 23, Supplementary Notes 25, and Supplementary Notes 27 in the same dependent relationship as Supplementary Notes 11 to 21. Furthermore, not limited to Supplementary Notes 1, Supplementary Notes 10, Supplementary Notes 22, Supplementary Notes 23, Supplementary Notes 24, Supplementary Notes 25, Supplementary Notes 26, and Supplementary Notes 27, various hardware, software, various recording means for recording software, or systems may also be dependent on some or all of the configurations described as Supplements, within the scope of each of the above-mentioned embodiments.

[0203] 10, 20, 30, 40 Shelf allocation support device 11 Terminal device 80 Computer 102, 202 Determination unit 104, 204, 304, 404 Prediction unit 106, 206, 306, 406 Output unit 208, 408 Learning unit 210, 410 Target product identification unit 212, 412 Acquisition unit 214, 414 Shelf allocation information generation unit 216, 416 Product determination unit 304, 404 Prediction unit 801 Processor 802 ROM 803 RAM 804 Storage device 805 Communication interface 806 Input / output interface 807 Bus 2000, 4000 Product DB NT Communication network S1 to S19 Shelf

Claims

1. A shelf allocation support device comprising: a determination means for determining candidates for a shelf on which a target product is to be displayed based on flow information including at least one of the number of customers passing through each of a plurality of shelves in a store where the product is displayed and the length of customer stay; a prediction means for predicting a predicted sales quantity when the target product is displayed on each of the shelf candidates using a model in which the correspondence between the shelf on which the product is displayed and the sales quantity of the product is learned; and an output means for outputting shelf information indicating a shelf that is predicted to achieve the target sales quantity of the target product based on the predicted sales quantity.

2. The shelf allocation support device according to claim 1, further comprising: a shelf allocation information generation means for generating shelf allocation information when the target product is displayed on the predicted shelf, wherein the prediction means predicts a predicted sales quantity when the target product is displayed at each position on the shelf and a predicted sales quantity when other products to be displayed together with the target product are displayed at each position on the shelf, and the shelf allocation information generation means generates shelf allocation information in which a product with a higher priority is displayed at a position on the shelf where the predicted sales quantity is higher based on the priority determined for each of the target product and the other products.

3. The shelf allocation support device according to claim 1 or 2, wherein the model is categorized by product category, and the prediction means predicts the predicted sales quantity for each of the shelf candidates using a model corresponding to the category of the target product among the plurality of models.

4. The shelf allocation support device according to any one of claims 1 to 3, wherein the model is categorized by product sales period, and the prediction means predicts the predicted sales quantity for each of the shelf candidates using a model corresponding to the sales period of the target product among the plurality of models.

5. The shelf allocation support device according to any one of claims 1 to 4, wherein the determination means determines, as the shelf candidates, the shelves with a large number of customers among the plurality of shelves.

6. The shelf allocation support device according to any one of claims 1 to 5, wherein the determination means determines, as the shelf candidates, the shelves with a long customer stay time among the plurality of shelves.

7. The shelf allocation support device according to any one of claims 1 to 6, further comprising a learning means for generating the model by learning the correspondence between the shelf on which the product is arranged and the sales quantity of the product.

8. Prediction means for predicting the predicted sales quantity when the target product is displayed on each of the plurality of shelves, using a model in which the correspondence between the flow information including at least one of the number of customers passing in front of each of the plurality of shelves and the length of customer stay in the store where the product is displayed and the sales quantity of the product is learned; Output means for outputting shelf information indicating the shelves in the store that satisfy the target sales quantity of the target product from the plurality of shelves based on the predicted sales quantity; A shelf allocation support device comprising:

9. The shelf allocation support device according to claim 8, wherein the prediction means inputs planned flow information into the model to predict the predicted sales quantity when the target product is arranged on each of the plurality of shelves.

10. The shelf allocation support device according to claim 9, wherein the planned flow information is flow information in the same period as the period when the target product is sold.

11. Shelf allocation information generation means for generating shelf allocation information when the target product is displayed on the predicted shelf; The prediction means predicts the predicted sales quantity when the target product is displayed at each position on the shelf and the predicted sales quantity when other products to be displayed together with the target product are displayed at each position on the shelf; The shelf allocation information generation means generates shelf allocation information in which products with higher priority are displayed at positions where the predicted sales quantity is higher on the shelf based on the priority determined for each of the target product and the other products. The shelf allocation support device according to any one of claims 8 to 10.

12. The model exists for each category of products, The prediction means predicts the predicted sales quantity when the target product is displayed on each of the plurality of shelves, using the model corresponding to the category of the target product among the plurality of models. The shelf allocation support device according to any one of claims 10 to 11.

13. The model exists for each sales period of products, The prediction means predicts the predicted sales quantity when the target product is displayed on each of the plurality of shelves, using the model corresponding to the sales period of the target product among the plurality of models. The shelf allocation support device according to any one of claims 10 to 12.

14. A learning means for generating the model by learning the correspondence between the plurality of shelves, the customer flow information, and the sales quantity of the product, The shelf allocation support device according to any one of claims 8 to 13, further comprising.

15. An acquisition means for acquiring the target sales quantity of the target product, The shelf allocation support device according to any one of claims 1 to 14, comprising.

16. The acquisition means acquires the order quantity of the target product as the target sales quantity, The shelf allocation support device according to claim 15.

17. A computer determines candidates for shelves on which to display a target product based on customer flow information including at least either the number of customers passing through each of a plurality of shelves in a store where the product is displayed and the length of customer stay, Using a model in which the correspondence between the shelf on which the product is displayed and the sales quantity of the product is learned, predicts the predicted sales quantity when the target product is displayed on each of the shelf candidates, Based on the predicted sales quantity, outputs shelf information indicating the shelf that is predicted to achieve the target sales quantity of the target product, A shelf allocation support method for executing the process.

18. A computer uses a model in which the correspondence between a plurality of shelves in a store where a product is displayed, customer flow information including at least either the number of customers passing in front of each of the plurality of shelves and the length of customer stay, and the sales quantity of the product is learned, to predict the predicted sales quantity when the target product is displayed on each of the plurality of shelves, Based on the predicted sales quantity, outputs shelf information indicating the shelves in the store that satisfy the target sales quantity of the target product from the plurality of shelves, A shelf allocation support method for executing the process.

19. A non-transitory computer-readable recording medium that records a program for causing a computer to execute a process of determining candidates for shelves on which to display a target product based on customer flow information including at least either the number of customers passing through each of a plurality of shelves in a store where the product is displayed and the length of customer stay, Using a model in which the correspondence between the shelf on which the product is displayed and the sales quantity of the product is learned, predicting the predicted sales quantity when the target product is displayed on each of the shelf candidates, Based on the predicted sales quantity, outputting shelf information indicating the shelf that is predicted to achieve the target sales quantity of the target product.

20. A non-transitory computer-readable recording medium that records a program for causing a computer to execute a process of predicting a predicted sales quantity when a target product is displayed on each of a plurality of shelves in a store where the product is displayed, using a model in which a correspondence relationship between flow information including at least one of the number of customers passing in front of each of the plurality of shelves and the length of customer stay in the store and the sales quantity of the product is learned, and outputting shelf information indicating a shelf in the store that satisfies the target sales quantity of the target product from the plurality of shelves based on the predicted sales quantity.

Images