Model generation method, evaluation method, evaluation model, model generation device, evaluation device, and program

Abstract

To provide a model generation method and device and an evaluation method, model, device, and program for grasping a sales promotion effect of each sales promotion plan, each commodity, and each day to increase commodity sales volume.SOLUTION: A model generation method in a sales promotion effect evaluation system 1 executes steps in which: an evaluation device obtains, from store terminals 10 installed in stores A, B, and C, sales transition data regarding the transition of sales quantity of each commodity regarding a plurality of commodities sold in each store, and from a model generation device, selling data including commodity attribute data regarding attributes of each commodity and sales promotion plan data regarding an implemented sales promotion plan; and the model generation device calculates sales quantity characteristics with respect to the number of unit periods in which the sales quantity was achieved relative to the sales quantity per unit period for each commodity, classifies each commodity into commodity groups according to the similarity of sales quantity characteristics, and generates an evaluation model using the sales transition data and the sales promotion plan data relating to commodities classified in each commodity group.SELECTED DRAWING: Figure 1A

Description

[Technical field]

[0001] The present invention relates to a model generation method, an evaluation method, an evaluation model, a model generation device, an evaluation device, and a program for evaluating a promotion effect in terms of product sales volume. [Background technology]

[0002] In a store that sells products, various plans (called sales promotion plans) are implemented to promote sales. In order to increase the sales volume of products, it is necessary to implement the sales promotion plans efficiently, and for this purpose, it is necessary to understand the effect of each sales promotion plan on the sales volume of each product, and to select the sales promotion plan to be implemented and the products to be handled in the sales promotion plan. For example, Patent Document 1 discloses a prediction method for predicting future demand using POS data in order to procure raw materials and supplies and to secure human resources in the business of selling products and providing product services. However, with the conventional method, although it is possible to grasp the sales promotion effect on a daily basis, it is not possible to grasp the sales promotion effect for each sales promotion plan and each product, which is necessary for selecting a sales promotion plan. [Prior art document] [Patent documents] [Patent Document 1] JP 2022-154885 A Summary of the Invention [Means for solving the problem]

[0003] In a first aspect of the present invention, there is provided a model generation method for generating an evaluation model for evaluating a promotional effect on the sales volume of a product, the model generation method comprising the steps of: acquiring, from a store terminal installed in at least one store, sales data regarding a plurality of products sold at the at least one store, the sales data including sales trend data regarding the trends in sales volume of each of the plurality of products, product attribute data regarding attributes of each of the plurality of products, and promotional plan data regarding implemented promotional plans; calculating, for each of the plurality of products, sales volume characteristics regarding the number of unit periods in which the sales volume was achieved relative to the sales volume per unit period, classifying the plurality of products into at least one product group according to the similarity of the sales volume characteristics; and generating the evaluation model using the sales trend data and promotional plan data for products classified into the at least one product group from the sales data.

[0004] In a second aspect of the present invention, there is provided an evaluation method for evaluating a promotional effect in terms of sales volume of a product, comprising the steps of: acquiring sales data from a store terminal installed in at least one store regarding at least one product sold in the at least one store, the sales data including sales trend data regarding trends in sales volume of the at least one product, product attribute data regarding attributes of the at least one product, and promotional plan data regarding implemented promotional plans; and inputting the sales data into an evaluation model generated by the model generation method of the first aspect, to evaluate the promotional effect in terms of sales volume of the at least one product.

[0005] In a third aspect of the present invention, there is provided an evaluation model generated by the model generation method of the first aspect.

[0006] In a fourth aspect of the present invention, there is provided a model generation device for generating an evaluation model for evaluating a promotional effect on the sales volume of a product, the model generation device comprising: a first acquisition unit that acquires, from a store terminal installed in at least one store, sales data regarding a plurality of products sold at the at least one store, the sales data including sales trend data regarding the trends in sales volume of each of the plurality of products, product attribute data regarding attributes of each of the plurality of products, and promotional plan data regarding implemented promotional plans; and a generation unit that calculates, for each of the plurality of products, sales volume characteristics regarding the number of unit periods in which the sales volume was achieved relative to the sales volume per unit period, classifies the plurality of products into at least one product group according to the similarity of the sales volume characteristics, and generates the evaluation model using the sales trend data and promotional plan data for products classified into the at least one product group from the sales data.

[0007] In a fifth aspect of the present invention, there is provided an evaluation device for evaluating a promotional effect in terms of sales volume of a product, the evaluation device comprising: a second acquisition unit that acquires sales data from a store terminal installed in at least one store regarding at least one product sold in the at least one store, the sales data including sales trend data regarding trends in sales volume of the at least one product, product attribute data regarding attributes of the at least one product, and promotional plan data regarding implemented promotional plans; and an evaluation unit that inputs the sales data into an evaluation model generated by the model generation device of the fourth aspect to evaluate the promotional effect in terms of sales volume of the at least one product.

[0008] In a sixth aspect of the present invention, in order to generate an evaluation model for evaluating a promotional effect on the sales volume of a product, a program is provided that causes a computer to execute the steps of: acquiring sales data from a store terminal installed in at least one store regarding a plurality of products sold at the at least one store, the sales data including sales trend data regarding the trends in sales volume of each of the plurality of products, product attribute data regarding attributes of each of the plurality of products, and promotional plan data regarding implemented promotional plans; calculating, for each of the plurality of products, sales volume characteristics regarding the number of unit periods in which the sales volume was achieved relative to the sales volume per unit period, classifying the plurality of products into at least one product group according to the similarity of the sales volume characteristics; and generating the evaluation model using the sales trend data and promotional plan data for products classified into the at least one product group from the sales data.

[0009] In a seventh aspect of the present invention, there is provided a program for causing a computer to execute the steps of: acquiring sales data from a store terminal installed in at least one store regarding at least one product sold in the at least one store, the sales data including sales trend data regarding trends in sales volume of the at least one product, product attribute data regarding attributes of the at least one product, and promotional plan data regarding implemented promotional plans; and inputting the sales data into an evaluation model generated by executing the program of the sixth aspect, in order to evaluate the promotional effect in terms of sales volume of the at least one product.

[0010] The above summary of the invention does not list all of the features of the present invention. Also, subcombinations of these features may also be inventions. [Brief description of the drawings]

[0011] [Figure 1A] 1 shows a configuration of a sales promotion effect evaluation system according to an embodiment of the present invention. [Figure 1B] 2 shows the functional configuration of a model generating device. [Figure 1C] 2 shows the functional configuration of an evaluation device. [Figure 2A] The various data acquired during the data acquisition stage are shown below. [Figure 2B] An example of data to be input for model generation is shown below. [Figure 2C] 4 shows an example of evaluation model parameters. [Figure 3A] The data to be additionally input to the evaluation device is shown. [Figure 3B] 1 shows an example of data output from the evaluation device. [Figure 4A] 4 shows the flow of generating an evaluation model. [Figure 4B] 1 shows a sub-flow of the evaluation model generation stage. [Diagram 5] This shows a flow specific to the promotional period. [Figure 6A] An outline of discount period specifications is given below. [Figure 6B] An outline of the specific promotional period is given below. [Figure 7] This shows the flow for determining standard selling prices. [Figure 8] An outline of standard selling prices is shown below. [Figure 9] The flow of sales promotion planning grouping is shown below. [Figure 10A] 13 shows an example of aggregated data of daily sales volume for sales promotion plans. [Figure 10B] An example of sales volume characteristics of a promotional plan is shown below. [Figure 10C] 1 shows an example of the similarity of sales volume characteristics between promotional plans. [Figure 10D] An example of promotional project grouping is shown below. [Figure 11] The flow of store grouping is shown below. [Figure 12A] 13 shows an example of aggregated data on the implementation status of sales promotion projects in stores. [Figure 12B] 13 shows an example of a similarity coefficient of the implementation status of a sales promotion plan between stores. [Figure 12C] An example of store grouping is shown. [Figure 13] 1 shows the flow of product grouping. [Figure 14A] An example of aggregated data of daily sales quantities of products is shown below. [Figure 14B] An example of a product sales quantity characteristic is shown below. [Figure 14C] 1 shows an example of the similarity of sales volume characteristics between products. [Figure 14D] An example of product grouping is shown below. [Figure 15] The flow of new product classification is shown below. [Figure 16] An example of a new product classification chart is shown below. [Figure 17] The model construction flow is shown below. [Figure 18A] The flow of evaluating the effectiveness of sales promotion is shown below. [Figure 18B] 13 shows a subflow of a promotion effect evaluation. [Figure 19] An example of an improvement priority (tier) determination table for each promotional plan is shown below. [Figure 20] 1 shows an example of a computer configuration. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the scope of the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0013] 1A shows the configuration of a sales promotion effect evaluation system 1 according to this embodiment. The sales promotion effect evaluation system 1 is a system that analyzes the effects of sales promotion plans implemented to promote product sales in each store, and includes a store terminal 10, a model generation device 20, an evaluation device 30, an external database 40, and an external storage device 50. These are connected to each other so as to be able to communicate with each other via a network 60 such as the Internet.

[0014] The store terminal 10 is a computer device that is installed in each store and that collects and records data (called sales data) 100 related to the sale of goods at each store. The store terminal 10 may be, for example, a POS terminal or a terminal device that manages a plurality of POS terminals. The store terminal 10 includes a central processing unit (CPU), a communication device, and a storage device (all not shown). The CPU executes a dedicated program to realize a function of recording goods transaction information. The dedicated program is stored, for example, in a ROM (not shown) and read by the CPU, or stored in a storage medium such as a CD-ROM and read by the CPU using a reading device (not shown), or stored in a cloud (multiple distributed servers or multiple subsystems, etc.), and is started by the CPU reading from the cloud via the communication device and expanding it in RAM. The communication device is a means for communicating with the model generating device 20, the evaluation device 30, and the external storage device 50 mutually via a network 60, and can communicate using a protocol such as TCP / IP, for example. The communication device may adopt either a wired communication method or a wireless communication method. The storage device is a storage device such as a hard disk drive (HDD) that stores sales data 100, which will be described later. In this embodiment, the storage device is provided within the store terminal 10. The store terminal 10 communicates with the model generation device 20 via a network 60 and can provide the model generation device 20 with the sales data 100 required for generating an evaluation model. An external storage device 50 may also be used as the storage device.

[0015] The store terminal 10 records, as sales data 100 (see FIG. 2A), product information including the list price, which is the standard selling price of the product, the sales points, which is the number of units sold of the product, the transaction date when the product was traded, and sales calculated by multiplying the sales points by the selling price, etc. The store terminal 10 can calculate the number of units sold and / or sales for each day and provide it to the system 1 as sales trends.

[0016] The model generating device 20 is a computer device that generates an evaluation model that evaluates the sales promotion effect in the sales volume of a product. The model generating device 20 includes a central processing unit (CPU), a communication device, and a storage device (all not shown). The model generating function is realized by executing a dedicated program by the CPU. The dedicated program is stored in, for example, a ROM (not shown), and is read by the CPU, or is stored in a storage medium such as a CD-ROM, and is read by the CPU using a reading device (not shown), or is stored in a cloud (multiple distributed servers or multiple subsystems, etc.), and is read by the CPU from the cloud via a communication device and deployed in RAM, thereby being started. The communication device is a means for mutually communicating with the store terminal 10, the evaluation device 30, the external database 40, and the external storage device 50 via the network 60, and can communicate using a protocol such as TCP / IP, for example. The storage device is a storage device such as a hard disk drive (HDD) that stores various information such as the created evaluation model and the sales data 100. In this embodiment, the storage device is provided in the model generating device 20. The model generating device 20 communicates with the evaluation device 30 via a network 60, and can provide the evaluation model that has been created to the evaluation device 30. The storage device may be disposed independently of the network 60, and connected to the evaluation device 30 via an interface such as SCSI or SATA. An external storage device 50 may be used as the storage device.

[0017] 1B shows a functional configuration of the model generating device 20. The model generating device 20 includes a first acquiring unit 21, a generating unit 22, and a storage unit .

[0018] The first acquisition unit 21 acquires sales data 100 related to a plurality of products sold in a store from a store terminal 10 installed in the store. The first acquisition unit 21 also connects to an external database 40 to acquire external data 105 including information supplementing the sales data 100.

[0019] 2A shows an example of sales data 100. The sales data 100 includes sales trend data 101 relating to the trend of sales quantities of each of a plurality of products, product attribute data 102 relating to the attributes of each of the plurality of products, promotion plan data 103 relating to promotion plans implemented, and store data 104 relating to the implementation status of the promotion plans in the store. The sales trend data 101, the product attribute data 102, the promotion plan data 103, and the store data 104 are generated and stored in the store terminal 10, and are provided from the store terminal 10 to the model generation device 20. The sales data 100 further includes, as external data 105 for supplementing the information, calendar data 106 relating to the calendar of business days of the store, weather data 107 relating to the weather on the business days of the store, and event data 108 relating to social events on the business days of the store.

[0020] The sales trend data 101 is data related to the trend of the sales points of the product, and is an accumulation of sales statement data recorded each time the product is sold. The sales statement data may be data generally used in a POS terminal, and includes, for example, the date of the sale on which the product was sold, store information including at least one of the store name, the store address or location indicating the area name, and the store number assigned to each store, product information including the product category (large category, medium category, and small category) and the product name, product management number including the POS code assigned to each product for management, the selling price which is the actual selling price including discounts, the sales points which is the number of products traded on the store's business day indicated on the date, and the sales amount calculated by multiplying the sales points by the selling price, the discount points which is the number of products traded at the discounted price, and the discount amount which is the discounted price. In this example, the sales detail data for the first row of the sales trend data 101 indicates that on the sales date "May 2, 2023", at the "Yokohama Store" shown in the store information, "Yamada Nishiki Junmai Daiginjo" from "National Sake" shown in the product information managed by product management number "69832984" was sold for a selling price of "1,300" yen and "1" unit shown in the sales points was discounted by "300" yen, as shown in the discount amount. The sales detail data in the second row of the sales trend data 101 indicates that on the sales date "May 2, 2023", at the "Kawasaki Store" shown in the store information, "Delicious Milk", a "Processed Milk" product shown in the product information managed by product management number "10984958", was sold for a selling price of "190" yen, with "2" units shown in the sales points, resulting in a sales amount of "380" yen, and no discount was applied to the price reduction point of "0".

[0021] The product attribute data 102 is data used for product management related to the name and product features of each product. It includes information such as a product category (which may include a large category, a medium category, and a small category) indicating a classification of the product, a manufacturer name indicating a manufacturer or seller of the product, a brand name that is a name determined by the manufacturer to identify the product, a product name that is a name of the product determined by the manufacturer, a POS code that is a management number assigned to each product, a product DNA that succinctly describes the characteristics of the product, a list price that is a standard selling price, a start date when the product was started, and an end date when the product was ended. The product attribute data 102 may be commonly managed between stores, or may be a minimum management unit set for the purpose of product management and a data group linked to it (SKU data). In this example, the first line of the product attribute data 102 indicates that the product category is "udon," that "Xx udon 250g" is manufactured by "manufacturer A" indicated by the manufacturer name and is of the brand name "brand A," that has a characteristic of "smoothness" indicated by the product DNA, that has a list price of "250" yen, and that it has been on sale since "December 1, 2000" indicated by the sales start date. The second line of the product attribute data 102 indicates that the product category is "Tsuyahime," that "Xxx Yamagata Tsuyahime 5kg" is manufactured by "manufacturer B" indicated by the manufacturer name and is of the brand name "brand B," that has a characteristic of "chewy" indicated by the product DNA, that has a list price of "1900" yen, and that it has been on sale since "April 24, 2012" indicated by the sales start date.

[0022] The sales promotion plan data 103 includes information such as a plan type indicating an overview of the sales promotion plan, a plan name given to the plan, a plan code assigned to each plan, a start date and end date of the implemented sales promotion plan, and a product code indicating the product handled in the sales promotion plan. The sales promotion plan data 103 may be common between stores. The sales promotion plan data 103 is stored and managed on a management server. In this example, the first line of the sales promotion plan data 103 is the plan code "10940", which indicates that a sales promotion plan with the plan name "Saturday and Sunday daily necessities" of the plan type "Saturday and Sunday menu" was implemented for the product with the product code "34395439" from the start date "January 20, 2022" to the end date "February 20, 2022". The second line of the promotional plan data 103 is the plan code "15349", which indicates that the promotional plan with the plan name "Monday, Tuesday, Wednesday Daily Special" of the plan type "Monday, Tuesday, Wednesday Menu" was implemented for the product with the product code "98027485" from the start date "April 19, 2022" to the end date "April 21, 2022".

[0023] The store data 104 includes store information including at least one of the store name, the store code assigned to each store, and the address or location of each store, information identifying at least one of the promotional plans implemented at each store, the plan type, the plan name, and the plan code, and information such as the start date and end date. The promotional plans implemented at each store can be linked to the promotional plan data 103 based on the plan type or plan code. The store data 104 is stored and managed for each store or on a management server. In this example, the first line of the store data 104 indicates that the store name "Yokohama store" with the store code "390" implemented a promotional plan with the plan name "Saturday and Sunday Miscellaneous Goods" with the plan type "Saturday and Sunday Menu" from "October 19, 2022" indicated as the start date to "October 20, 2022" indicated as the end date. The second line of store data 104 indicates that the store name "Shibuya Store" with store code "130" implemented a promotional plan with the plan name "Mon., Tues., Wed. Uniform Price" and plan type "Mon., Tues., Wed. Uniform Price" from the start date "October 12, 2022" to the end date "October 14, 2022."

[0024] The calendar data 106 includes information on at least one of non-business days of the store, national holidays, and days related to commonly recognized customs. Days related to commonly recognized customs are days that are not legally specified but are expected to change purchasing activities, such as Valentine's Day, Christmas, and Obon. In this example, non-business day data, Golden Week / year-end data, and holiday data are shown as examples of the calendar data 106. In this example, the first line of the non-business day data indicates that "January 1, 2021" shown in the date was a "Friday" day shown in the day of the week, and was a non-business day of the store. The second line of the non-business day data indicates that "January 2, 2021" shown in the date was a "Saturday" day shown in the day of the week, and was a non-business day of the store. In this example, the first line of the Golden Week / year-end data indicates that "May 1, 2022" shown in the date was the "1st" day shown in the "Golden Week" day shown in the holiday name. The second line of the Golden Week / Year-end data indicates that the date, "May 2, 2022," was the "2nd" day of the "Golden Week" holiday, as indicated by the holiday name. In this example, the first line of the holiday data indicates that the date, "July 23, 2020," was Marine Day, as indicated by the holiday name. The second line of the holiday data indicates that the date, "August 10, 2020," was Mountain Day, as indicated by the holiday name.

[0025] The weather data 107 includes at least one of information regarding the date of the store's business day, the weather in the store's location, the minimum temperature, the maximum temperature, the average temperature, the humidity, the air pressure, and the probability of precipitation. In this example, the first row of the weather data 107 indicates that the date "January 1, 2021" had an average temperature of "7"°C, a maximum temperature of "10"°C, a minimum temperature of "1"°C, a total probability of precipitation of "0", and a maximum probability of precipitation of "0", and the weather was "clear". The second row of the weather data 107 indicates that the date "January 2, 2021" had an average temperature of "6"°C, a maximum temperature of "7"°C, a minimum temperature of "4"°C, a total probability of precipitation of "10", and a maximum probability of precipitation of "5", and the weather was "clear".

[0026] The event data 108 includes information about social events. A social event is a change in social conditions that affects purchasing activities, such as the spread of an infectious disease, the issuance of a state of emergency declaration due to the spread of an infectious disease, and a change in the tax system. In this example, corona event data, corona infection data, and pre-tax increase period data are shown as examples of the event data 108. In this example, the first line of the corona event data indicates that the date "April 1, 2020" is the first day since the "State of Emergency" in the event name was issued as indicated by the flag "1001". The second line of the corona event data indicates that the date "April 2, 2020" is the second day since the "State of Emergency" in the event name was issued as indicated by the flag "1002". The first line of the corona infection number data indicates that the number of corona infections confirmed on the date "May 10, 2020" was "103" as indicated by the number of infections. The first line of the data on the number of coronavirus infections indicates that the number of coronavirus infections confirmed on the date "May 11, 2020" was "140" as indicated by the number of infections. The first line of the data on the period before the tax increase indicates that the date "September 1, 2019" is "30 days before the tax increase" as indicated by the explanation, and is the "1st" day as indicated by the day after the tax increase announcement. The second line of the data on the period before the tax increase indicates that the date "September 2, 2019" is "29 days before the tax increase" as indicated by the explanation, and is the "2nd" day as indicated by the day after the tax increase announcement. The event data 108 is not limited to the coronavirus event data, coronavirus infection data, and pre-tax increase period data listed here, and may record any social event that may affect purchasing activities. For example, when analyzing a magazine promotional plan, it may include information such as the most recent major incident and the date on which the follow-up report was reported.

[0027] The generating unit 22 calculates the sales volume characteristic for each of the multiple products, classifies the multiple products into at least one product group according to the similarity, and generates an evaluation model using the sales trend data 101 and the sales promotion plan data 103 for the products classified into each product group from the sales data 100. Here, the sales volume characteristic is a histogram of the number of unit periods of sales (also called reference period width) based on the sales volume of the product, and the histogram is determined by normalizing or standardizing for each product based on the sales volume of the product per unit period and the number of unit periods in which a specific sales volume was recorded. By using the histogram, it is possible to compare sales promotion plans with different reference period widths or different sales volume scales. The generation of the evaluation model by the generating unit 22 will be described later.

[0028] When generating the evaluation model, the generation unit 22 creates input data 109 by editing the sales data 100 so as to include information on factors that affect the sales volume, and generates the evaluation model based on the input data 109. The evaluation model generated by the generation unit 22 is output as a plurality of parameters (also called evaluation model parameters) 110 that indicate the degree of influence of each factor on the sales volume.

[0029] Input data 109 is generated by extracting and organizing items required for model generation from sales data 100. In this embodiment, in order to clarify the effect of the day of the week on sales, sales volume information for each day up to six days prior to the sales date is extracted and included in input data 109.

[0030] FIG. 2B shows an example of input data 109. In the input data 109, the sales as the objective variable for model generation are shown in the shaded columns, and the items of each factor are shown as explanatory variables in the dotted columns. In this example, the input data 109 includes, as explanatory variables, sales quantity (-1 day), sales quantity (-2 days), sales quantity (-3 days), sales quantity (-4 days), sales quantity (-5 days), sales quantity (-6 days), category indicating product classification, discount amount at the time of promotion planning, standard selling price (e.g., master selling price) which is the basic selling price, each group name of the promotion planning group, social event data (e.g., corona event, number of corona infected people, period before tax increase), and weather data including temperature (average temperature, maximum temperature, minimum temperature) and weather, and id. Here, id is an entry number for identifying data assigned by product and date. In this example, the first row of input data 109 shows that on the day when the sales quantity was "190" and the ID "1020" was assigned, the sales quantity (-1 day) was "292" units, the sales quantity (-2 day) was "123" units, the sales quantity (-3 day) was "134" units, the sales quantity (-4 day) was "194" units, the sales quantity (-5 day) was "133" units, and the sales quantity (-6 day) was "185" units, and the product classifications were "Western daily products" in the large category, "Milk" in the medium category, and "Processed milk" in the small category. A promotional plan of the planning group "Promotion_Mon-Tues-Wed Menu" was implemented, which deducted a discount amount of "20" yen from the master selling price of the product of "190" yen, the corona event "1021" occurred on the sales day, the number of corona infections was "120", it was a holiday that was not a consecutive holiday, it was not a period before the tax increase, the average temperature was "19" degrees Celsius, the maximum temperature was "25" degrees Celsius, the minimum temperature was "13" degrees Celsius, the total probability of precipitation was "10"%, and the maximum probability of precipitation was "5"%.The second line of input data 109 shows that on the day when the sales quantity was "204" units and the ID "1021" was assigned, the sales quantity (-1 day) was "324" units, the sales quantity (-2 day) was "180" units, the sales quantity (-3 day) was "193" units, the sales quantity (-4 day) was "173" units, the sales quantity (-5 day) was "221" units, and the sales quantity (-6 day) was "253" units, and the product classifications were "Large category "Alcohol", "Medium category "Sake", and "Small category "Sake". The product "Nationwide Sake" had a master selling price of "1500" yen and no promotional campaign was implemented, the corona event "1022" occurred on the sales date, the number of corona infections was "181", it was a holiday that was not a consecutive holiday, it was not a period before the tax increase, the average temperature was "24" degrees Celsius, the maximum temperature was "29" degrees Celsius, the minimum temperature was "22" degrees Celsius, the total probability of precipitation was "0"%, and the maximum probability of precipitation was "0"%.

[0031] The evaluation model parameters 110 include the influence (SHAP value) of each factor included in the input data 109 on the sales volume in the evaluation model generated by the generation unit 22 based on the input data 109. The evaluation model parameters 110 are calculated and output by the generation unit 22 for each product and date.

[0032] 2C shows an example of the evaluation model parameters 110. In this example, among the items of factors included in the evaluation model parameters 110, the item corresponding to the effect of the sales promotion plan is surrounded by a thick frame. In this example, the first row of the evaluation model parameters 110 shows that, for id "1020" corresponding to the input data 109, the SHAP value of the sales quantity (-1 day) is "0.21", the SHAP value of the sales quantity (-2 days) is "0.43", the SHAP value of the sales quantity (-3 days) is "-0.41", the SHAP value of the sales quantity (-4 days) is "-0.41", the SHAP value of the sales quantity (-5 days) is "-0.63", and the SHAP value of the sales quantity (-6 days) is "0.56", and the SHAP value of the large category is "1.24", the SHAP value of the medium category is "0.51", and the SHAP value of the small category is "0.01". "The SHAP value of the discount amount is "5.31", the SHAP value of the product's master selling price is "9.43", the SHAP value of the promotional plan classified as the Promotion_Mon-Tues-Wed menu is "10.32", the SHAP value of the COVID-19 event is "-0.45", the SHAP value of the number of COVID-19 infections is "-0.53", the SHAP value of the holiday is "9.14", the SHAP value of the average temperature is "0.02", the SHAP value of the maximum temperature is "0.04", the SHAP value of the minimum temperature is "0.00", the SHAP value of the total probability of precipitation is "0.00", and the SHAP value of the maximum probability of precipitation is "0.13". The second row of the evaluation model parameters 110 shows that for ID "1021" corresponding to the input data 109, the SHAP value for sales quantity (-1 day) is "0.53", the SHAP value for sales quantity (-2 days) is "0.24", the SHAP value for sales quantity (-3 days) is "0.24", the SHAP value for sales quantity (-4 days) is "-0.21", the SHAP value for sales quantity (-5 days) is "-0.13", and the SHAP value for sales quantity (-6 days) is "1.43", and the SHAP value for the large category is "2.53", the SHAP value for the medium category is "1. The SHAP value for the average temperature was "0.12", the SHAP value for the maximum temperature was "0.42", the SHAP value for the minimum temperature was "0.03", the SHAP value for the total probability of precipitation was "0.00", and the SHAP value for the maximum probability of precipitation was "0.00".

[0033] The storage unit 23 stores various information such as the evaluation model created by the generation unit 22 and the sales data 100. The storage unit 23 communicates with the evaluation device 30 via a network 60 or via an interface such as SCSI or SATA, and can provide the evaluation model created by the generation unit 22 to the evaluation device 30.

[0034] The evaluation device 30 is a computer device that inputs data on the sales promotion plan and the promotion target product to be evaluated from the sales data 100 into the evaluation model generated by the model generation device 20, and evaluates the promotion effect in terms of the sales volume of each product. The evaluation device 30 includes a central processing unit (CPU), a communication device, and a storage device (all not shown). The CPU executes a dedicated program to realize the promotion effect evaluation function. The dedicated program is stored, for example, in a ROM (not shown), and read by the CPU, or stored in a storage medium such as a CD-ROM, and read by the CPU using a reading device (not shown), or stored in a cloud (multiple distributed servers or multiple subsystems, etc.), and is read by the CPU via the communication device and deployed in RAM to be started. The communication device is a means for communicating with the store terminal 10, the model generation device 20, the external database 40, and the external storage device 50 via the network 60, and can communicate using a protocol such as TCP / IP, for example. The output device is a means for presenting the results of the model evaluation, and may be, for example, a monitor or a printer, and may provide a visual representation, such as a screen display or printout on paper.

[0035] 1C shows a functional configuration of the evaluation device 30. The evaluation device 30 includes a second acquisition unit 31, an evaluation unit 32, and an output unit 33. The evaluation device 30 and the model generating device 20 may be incorporated in the same device.

[0036] The second acquisition unit 31 acquires, from the store terminal 10 installed in each store, sales data 100 relating to products sold in each store, sales quantities of each product, and sales promotion plans, and evaluation model parameters 110 created from the model generation device 20. Furthermore, the second acquisition unit 31 acquires plan cost data 111 including at least the implementation costs of the sales promotion plans from each store terminal 10 or the external storage device 50.

[0037] An example of the project cost data 111 is shown in Fig. 3A. The project cost data 111 includes information including a breakdown of costs involved in implementing a promotional project. For example, in the case of a flyer advertisement, the costs involved in implementing a promotional project include the cost of printing the flyer and the cost of inserting the flyer. The project cost data 111 includes project information specifying the project including at least one of the project code, project name, and project type, the cost of printing the flyer, the cost of inserting the flyer, the total cost of printing and inserting, the number of inserts which is the number of flyers inserted, the printing cost per copy of the flyer, the insert cost per copy of the flyer, etc. In this example, the first line of the project cost data 111 shows that the project named "Hokkaido Everyday Goods" of project type "Producer's Fair" had a printing cost of "1,493,032" yen and an insert cost of "2,194,039" yen, for a total cost of "3,687,071" yen, with 394,294 insert copies distributed, with a printing cost of "3.79" yen and an insert cost of "5.56" yen. The second line of the project cost data 111 indicates that a project with the project name "Weekdays Uniform" of project type "Weekdays Uniform" had a total cost of "3,532,748" yen, consisting of printing cost of "2,103,439" yen and insert cost of "1,429,309" yen, and distributed 452,357 insert copies, with a printing cost of "4.65" yen and an insert cost of "3.16" yen. Sales promotion other than flyers, for example, in the case of commercials, may include items such as "commercial filming cost," "commercial editing cost," "commercial broadcast cost," "total cost," "commercial seconds," "commercial broadcast destination," "commercial broadcast count," "average viewer rating during commercial broadcast / commercial skip rate," etc.

[0038] The evaluation unit 32 inputs the sales data 100 into the evaluation model generated by the model generation device 20, and evaluates the promotion effect in terms of the sales volume of each product. In this embodiment, the evaluation unit 32 calculates the sales volume increase due to each promotion plan from the impact of each promotion plan on the sales volume in the evaluation model generated by the model generation device 20 (evaluation model parameters 110) and the sales volume data included in the sales data 100, calculates the investment return from the sales amount increase obtained by multiplying the sales volume increase by the selling price and the expenses related to the implementation of the plan, and determines the improvement priority for each promotion plan. The evaluation unit 32 can also predict sales for each product and the sales amount increase for a social event.

[0039] The output unit 33 outputs the evaluation results of the promotional effectiveness by the evaluation unit 32 in the form of numerical information or a chart or the like created based on numerical information as a data file or printout. In this embodiment, the output unit 33 presents the investment return and improvement priority for each promotional plan created by the evaluation unit 32, but may also output the increase in sales amount for the promotional plan. The output unit 33 can also output a sales forecast for a product and a sales increase forecast for a social event. When the evaluation results of the promotional effectiveness are numerical information, they are output as evaluation data 112 for each product and promotional plan.

[0040] 3B shows an example of the evaluation data 112. The evaluation data 112 is an evaluation of the effectiveness of each promotional plan, and includes information such as the sales forecast for each promotional plan and the priority order of the promotional plans. The evaluation data 112 includes evaluations of promotional plan information including plan name and plan code, target product data such as product name and POS code, total sales volume, breakdown of promotional plan effectiveness such as announced discount increment and announced increment, target product selling price, investment costs such as plan implementation costs, promotional effectiveness including expected profit amount, ROI which is return on investment, and tier which is improvement priority for each promotional plan. In this example, the first row of the evaluation data 112 indicates that the promotional plan with the plan name "Monday, Tuesday, Wednesday, uniform price" had an announced discount increment of "34" units and an announced increment of "23" units added to the total quantity of "123" units, which is the sales quantity of the product name "xxx Milk 500ml", resulting in an announced discount / discount of "1.48", a selling price of "209" yen, and an investment cost of "29,341" yen, resulting in a profit of "2,945" yen, a quantity increment percentage of "21.4", an ROI of "10.3", and an improvement priority for each promotional plan of "4" as shown in Tier. The second row of the evaluation data 112 shows that, for the plan named "Monday, Tuesday, Wednesday, daily necessities," the announced discount increment of "19" units and the announced increment of "4" units were increased from the total quantity "93" units, which is the number of units sold of the product name "xxx pot puller," resulting in an announced discount / discount of "4.75," a selling price of "301" yen, and an investment cost of "30,921" yen, resulting in a profit of "1,920" yen, a quantity increase rate of "13.5," an ROI of "3.4," and an improvement priority for each sales promotion plan of "7," as shown in Tier.

[0041] The external database 40 is a database that provides various data to the sales promotion effect evaluation system 1. The various data is, for example, external data 105 including the above-mentioned calendar data 106, weather data 107, and event data 108. The external database does not need to be prepared separately, and for example, data from an external information provider may be used via the Internet. The data from the external information provider may be, for example, weather data from the Japan Meteorological Agency and infectious disease outbreak trend surveys from the National Institute of Infectious Diseases. When an external information provider is not used, the external database 40 may be a storage device such as a hard disk drive (HDD) communicably connected to the network 60, and is managed from the store terminal 10 or the model generation device 20 via the network 60. When an external information provider is not used, an external storage device 50 may be used as the external database 40.

[0042] The external storage device 50 is a storage device such as a hard disk drive (HDD) communicatively connected to the network 60, and can record communication history, a model generated by the model generating device 20, evaluation data of the sales promotion effect generated by the evaluation device 30, and the like. The external storage device 50 includes a central processing unit (CPU) and a communication device, and the CPU may periodically and automatically acquire and store external data 105 from the external database 40 using a dedicated program. The external storage device 50 may share functions with the storage device of the store terminal 10, the storage device of the model generating device 20, a storage device added to the evaluation device 30, the external database 40, and the like. The external storage device 50 may be managed in a cloud (multiple distributed servers or multiple subsystems, etc.) format.

[0043] The network 60 is a communication network that connects the store terminal 10, the model generating device 20, the evaluation device 30, the external database 40, and the external storage device 50 so that they can communicate with each other. The network 60 is, for example, the Internet, but is not limited to this. As long as they can communicate with each other, any communication network such as a local area network or a telephone line may be used, or a network in which multiple communication networks are mixed may be used.

[0044] 4A shows a flow of model generation. The flow starts when the user inputs a start command on the model generation device 20.

[0045] In step S100, the model generation device 20 (first acquisition unit 21) acquires sales data 100 related to a plurality of products sold in each of the plurality of stores from the store terminals 10 installed in the plurality of stores. The sales data 100 is as described above with reference to FIG. 2A.

[0046] In step S200, the model generation device 20 (generation unit 22) calculates, for each of the multiple products, sales volume characteristics regarding the number of unit periods in which the sales volume was achieved relative to the sales volume per unit period, classifies the multiple products into at least one product group according to the similarity of the sales volume characteristics, and generates an evaluation model using sales trend data 101 and sales promotion planning data 103 for the products classified into each product group from the sales data 100.

[0047] FIG. 4B shows a sub-flow of step S200.

[0048] In step S210, the model generating device 20 (generation unit 22) organizes the acquired sales data 100 by combining and cleansing it as data combining and cleansing. In data combining, multiple tables included in the sales data 100 are combined and necessary data is consolidated into one, thereby reducing the number of items and unifying the data into a format suitable for processing. In cleansing, data not used in processing is deleted to reduce the data volume. Step S210 may include steps for supplementing missing data, for example, step S220 for identifying a promotional period and step S230 for identifying a standard selling price.

[0049] 5 shows a sub-flow of step S220. In step S220, the model generation device 20 (generation unit 22) determines the implementation period of the sales promotion plan from the trends in the selling price and / or sales volume of each product based on the sales data 100. Step S220 consists of the procedures of aggregating the selling price and sales volume by product and store, smoothing the daily selling price, determining the discount period, smoothing the daily sales volume, and determining the sales promotion period. Step S220 is performed when the store data 104 does not include promotion plan implementation information for each store, or when some of the promotion plan implementation information is missing.

[0050] In step S221, the model generation device 20 (generation unit 22) judges whether promotional plan implementation information for each store is not included or is partially missing in the store data 104. If it is judged that the promotional plan implementation information is not included or is partially missing, the process proceeds to step S222, and if it is judged that the promotional plan implementation information is included and is complete, the subflow ends.

[0051] In step S222, the model generation device 20 (generation unit 22) tabulates the selling price and sales volume for each product by store. The tabulation of the selling price and sales volume for each product by store is performed by extracting the selling price and sales volume for each product by day for each store based on the sales trend data 101, and creating data including the date, selling price, and sales volume information for the number of business days.

[0052] In step S223, the model generating device 20 (generating unit 22) smoothes the daily selling price. The daily selling price is smoothed by using a moving average method in which the time-series data of the selling price obtained by aggregation is moved in a certain direction over a range period to obtain an average value over the range period. A method other than the moving average method may be used for smoothing.

[0053] In step S224, the model generating device 20 (generation unit 22) determines whether a period in which the selling price falls below a predetermined threshold can be confirmed in the smoothed selling price data. A period in which the selling price falls below the threshold can be regarded as a discount period. The threshold may be set appropriately in advance. If a period in which the selling price falls below the threshold cannot be confirmed, step S225 is omitted.

[0054] In step S225, the model generating device 20 (the generating unit 22) determines the discount period. The discount period is determined from the period during which the selling price in the smoothed selling price data falls below the threshold value.

[0055] In step S226, the model generating device 20 (generating unit 22) smoothes the daily sales volume. The daily sales volume is smoothed by using a moving average method in which the time-series data of the sales volume obtained by aggregation is moved in a certain direction over a range period to obtain an average value over the range period. A method other than the moving average method may be used for smoothing.

[0056] In step S227, the model generating device 20 (generation unit 22) determines whether a period in which the sales volume increases beyond a predetermined threshold can be confirmed in the smoothed sales volume data. A period in which the sales volume exceeds the threshold can be regarded as a sales promotion period. The threshold may be set appropriately in advance. If a period in which the sales volume exceeds the threshold cannot be confirmed, step S228 is omitted.

[0057] In step S228, the model generating device 20 (the generating unit 22) determines a promotion period. The promotion period is determined from a period during which the sales volume in the smoothed sales volume data increased beyond a threshold. Promotion plans carried out during the promotion period are identified from the promotion plan data 103.

[0058] By omitting or completing step S228, the identification of the promotion period in the generation of the evaluation model is completed.

[0059] Fig. 6A shows an outline of a method for determining a discount period. Here, the vertical axis indicates the selling price of the product being promoted, and the horizontal axis indicates time. The dashed line in the figure indicates the change in selling price before smoothing, and the thin solid line indicates the change after smoothing. In this embodiment, the reference value for selling price is set to the most frequent selling price over the entire period, indicated by the thin line, and the period in which the smoothed selling price falls below the threshold value, indicated by the thick line and set at 95% of the most frequent selling price, is determined to be the discount period.

[0060] Fig. 6B shows an outline of a method for determining the specific sales promotion period. Here, the vertical axis represents the sales volume of the promoted product, and the horizontal axis represents time. The dashed line in the figure represents the change in sales volume before smoothing, and the thin solid line represents the change after smoothing. In this embodiment, the reference value for sales volume is set to the most frequent sales volume for the entire period, shown by the thin line, and the period in which the sales volume exceeds the threshold value, shown by the thick line and set at 110% of the most frequent sales volume, is determined to be the discount period.

[0061] FIG. 7 shows a subflow of step S230. In step S230, a standard selling price is determined from the most frequent selling price of each product based on the sales data 100. Step S230 consists of the procedures of tabulating the product prices by store and calculating the standard selling price. Step S230 is executed when the sales trend data 101 is missing information on the standard selling price, which is the basis for varying the selling price due to discounts, etc., for at least one product or at least one store. In this embodiment, the standard selling price is calculated by store, but it may also be determined collectively for multiple stores by region, for example.

[0062] In step S231, the model generating device 20 (generating unit 22) judges whether standard selling price information is not included or is partially missing in the sales transition data 101. If it is judged that standard selling price information is not included or is partially missing, the process proceeds to step S232, and if it is judged that standard selling price information is included and is not missing, the subflow is terminated.

[0063] In step S232, the model generating device 20 (generation unit 22) tabulates the product sales prices by store. The product sales prices by store are tabulated by extracting the sales price of each product at each store for each day based on the sales trend data 101, and organizing the daily sales price information from the implementation period of the sales promotion plan into data including the number of business days. The tabulated product sales prices by store are information that records only the daily sales prices for each product at each store.

[0064] In step S233, the model generating device 20 (the generating unit 22) calculates a standard selling price. The selling price that is the most frequent value among the selling prices collected by store and by product can be determined as the standard selling price.

[0065] By omitting or completing step S233, the determination of the standard selling price is completed.

[0066] Fig. 8 shows an outline of the method for determining the standard selling price. Here, the vertical axis indicates the number of days that a product was sold at a specific price in each store, and the horizontal axis indicates the selling price of the product in each store. In this embodiment, the standard selling price is determined as the selling price for which the number of days that the product was set at a specific price in each store during the period was the greatest.

[0067] 9 shows a subflow of step S240. In step S240, promotion plan grouping is performed based on the promotion plan data 103. The promotion plan grouping calculates the sales volume characteristics of products handled in each promotion plan implemented in each store, and classifies the promotion plans into at least one promotion plan group according to the similarity of the sales volume characteristics. In this embodiment, step S240 includes the steps of tallying up daily sales volumes by plan, extracting sales volume characteristics, calculating similarity, clustering, and determining promotion plan groups.

[0068] In step S241, the model generation device 20 (generation unit 22) judges whether the number of types of promotional plans exceeds an appropriately set threshold. If the number of types of promotional plans exceeds the threshold, the process proceeds to step S242, and if the number is below the threshold, the subflow ends. In this embodiment, the reference number for executing promotional plan grouping is n, and a number at which the number of types of promotional plans increases to the point where the processing burden becomes unacceptable is input. If it is difficult to determine n, n may be fixed at 1. Also, regardless of the processing burden, if a sufficient number of implementations for each promotional plan cannot be secured, the process proceeds to step S242 and the subsequent processes may be executed.

[0069] In step S242, the model generation device 20 (generation unit 22) tabulates the daily sales quantities by plan. By tabulating the daily sales quantities by plan, the sales quantities of each product at each store are extracted for each day based on the sales trend data 101 and the promotion plan data 103, and data is organized that includes the daily sales quantities for each promotion plan for the number of business days from the implementation period of the promotion plan. In this embodiment, the sales trends of all stores that implemented the promotion plan are the subject of the tabulation, but it is also possible to target only some of the stores that implemented the promotion plan. In this embodiment, all products that are the subject of the plan are the subject of the tabulation, but it is also possible to target only some of the products that are the subject of the plan.

[0070] FIG. 10A shows an example of the daily sales volume for each promotional plan organized in step S242 from the sales trend data 101 and the promotional plan data 103. In this example, the first line shows that when a promotional plan "Flyer_Nagano Specialty Fair" was implemented to distribute flyers related to the Nagano Specialty Fair for a certain product on "January 10, 2023," the sales volume of the aforementioned product was "30," and the second line shows that when a promotional plan "Flyer_Spring New Life Special" was implemented to distribute flyers related to the spring new life special for the aforementioned product on "February 24, 2023," the sales volume of the aforementioned product was "25."

[0071] In step S243, the model generating device 20 (generation unit 22) extracts sales volume characteristics from the collected sales volume. The sales volume characteristics are determined for each promotional plan from the sales volume and the number of days during which a specific sales volume was recorded within a period. The sales volume characteristics have normalized or standardized elements, and are, for example, a histogram of the number of days based on the sales volume. In this embodiment, the reference period width is days, but it may be arbitrarily changed to, for example, several days, weeks, months, or years.

[0072] FIG. 10B shows an example of the sales volume characteristics for each promotional plan extracted in step S243. The histogram on the left shows the sales volume characteristics for "Flyer_Nagano Specialty Product Fair." The characteristics exhibit one high peak and a long tail from the peak to the increase in sales volume. The histogram on the right shows the sales volume characteristics for "Flyer_Spring New Lifestyle Special." The characteristics exhibit two peaks.

[0073] In step S244, the model generation device 20 (generation unit 22) calculates the similarity between the promotional plans. The similarity is calculated by comparing the sales volume characteristics between the promotional plans. In this embodiment, the similarity of the sales volume characteristics is determined by formula (1) using Euclidean distance from the difference in the sales volume characteristics for each combination of promotional plans. When Euclidean distance is used, the closer the similarity of the sales volume characteristics is to 0, the more similar they are. The similarity may be determined by extracting and comparing the sales volume characteristics of different promotional plans using different reference time widths depending on the scale of the period in which the promotional plans are implemented. f = ((x1-y1) 2 +(x2-y2) 2 +···+(xn-yn) 2 ) 1 / 2 (1) Here, xn is the ratio of the number of days in sales volume interval n in the sales volume characteristics of promotional plan x to the total number of days during the promotional plan implementation period, and yn is the ratio of the number of days in sales volume interval n in the sales volume characteristics of promotional plan y to the total number of days during the promotional plan implementation period.

[0074] 10C shows an example of the similarity between the sales volume characteristics of the promotional plans extracted in step S244. In this example, the similarity between the sales volume characteristics of "Flyer_Nagano Specialty Products Fair" and "Flyer_Spring New Lifestyle Special" is "32.6", the similarity between the sales volume characteristics of "Flyer_Nagano Specialty Products Fair" and "Event_Food Education Festival" is "158.2", and the similarity between the sales volume characteristics of "Flyer_Spring New Lifestyle Special" and "Event_Food Education Festival" is "192.9".

[0075] In step S245, the model generation device 20 (generation unit 22) performs clustering of the promotional plans based on the similarity of the sales volume characteristics between the promotional plans. In this embodiment, the clustering of the promotional plans is performed by determining the position of each promotional plan on the Euclidean plane according to the mutual similarity between the promotional plans, and classifying the promotional plans into groups. The clustering of the promotional plans is performed based on the similarity, for example, using machine learning such as the Kmeans method.

[0076] In step S246, the model generation device 20 (generation unit 22) determines a promotion plan group for the promotion plan. The promotion plan group is determined by extracting groups that satisfy a reference grouping accuracy from among the groups created by clustering the promotion plans. In this embodiment, a promotion plan group is extracted when the coefficient of determination is 0.6 or more and the total root mean square error rate is 40% or less. The reference grouping accuracy may be determined as appropriate in advance, or the promotion plan group created in step S245 may be directly reflected without extracting the promotion plan group based on the grouping accuracy.

[0077] By omitting or completing step S246, the promotion plan grouping in the generation of the evaluation model is completed.

[0078] FIG. 10D shows an example of what groups are formed in the promotion plan grouping S240. The distribution plane of each promotion plan and promotion plan group is a Euclidean plane with the promotion plan similarity as the Euclidean distance, and expresses the mutual relationships between the promotion plans. In this example, the plans "Flyer_Spring New Life Special" and "Flyer_Nagano Specialty Product Fair" are classified into group α, the plans "Flyer_Halloween", "Event_Halloween", and "Event_Christmas" are classified into group β, and the plans "Event_Food Education Festival", "Event_Great Thanksgiving Festival", and "Flyer_Weekday Only Discount" are classified into group θ. The promotion plans assigned to each group can be considered to have a similar impact on sales of the product.

[0079] According to the above-mentioned promotion plan grouping, promotion plans with similar sales trends can be grouped together based on the similarity of the sales volume characteristics calculated for each promotion plan, and narrowing down the types of sales plans can improve the learning accuracy of model construction, etc. Promotion plan grouping is particularly effective when the number of promotion plan types increases and model construction becomes complicated, or when there are still few examples of promotion plans whose effects you want to confirm but many partially similar promotion plans have been implemented in the past.

[0080] FIG. 11 shows a subflow of step S250. Step S250 performs store grouping based on the sales promotion plan data 103 and the store data 104. Store grouping evaluates the similarity of the implementation status of the sales promotion plan for each of multiple stores, and classifies the multiple stores into at least one store group according to the similarity. Step S250 is executed when there is variation in the implementation status of the sales promotion plan between stores, and the amount of data is insufficient to generate a highly accurate model by simply aggregating sales trend data of stores with the same implementation status of the sales promotion plan. Step S250 may be executed for the purpose of reducing the number of items in model generation even when there is no variation in the implementation status of the sales promotion plan between stores. In this embodiment, step S250 consists of the procedures of aggregating the implementation status of the sales promotion plan for each store, calculating the correlation coefficient, clustering, and determining the store group.

[0081] In step S251, the model generation device 20 (generation unit 22) judges whether there is variation in the promotion plan implementation status between stores. The criterion for judging whether there is variation may be, for example, when the difference in the promotion plan implementation status exceeds 20% of the total number of combinations consisting of the type of promotion plan and the number of stores. If there is variation in the promotion plan implementation status between stores, the process proceeds to step S252, and if it is judged that there is no variation in the promotion plan implementation status between stores, the subflow ends.

[0082] In step S252, the model generation device 20 (generation unit 22) tally up the promotion plan implementation status for each store. By tallying up the promotion plan implementation status, data is created in the store data 104 in which the implementation status of each promotion plan included in the promotion plan data 103 is extracted and the implementation status of each promotion plan for each store is organized.

[0083] 12A shows an example of the data collected in step S252. In this example, in the data collected in step S252, a "1" is assigned to a promotional plan that has been carried out in each store, and a "0" is assigned to a promotional plan that has not been carried out. For example, the first line indicates that the store "Yokohama Store" has carried out "Flyer_Nagano Specialty Product Fair" and "Flyer_Spring New Life Special" but has never carried out "Event_Great Thanksgiving Festival," and the second line indicates that the store "Kawasaki Store" has carried out "Flyer_Spring New Life Special" but has never carried out "Flyer_Nagano Specialty Product Fair" or "Event_Great Thanksgiving Festival."

[0084] In step S253, the model generation device 20 (generation unit 22) calculates a similarity coefficient of the promotion plan implementation status for each store. The similarity coefficient is calculated for each combination of stores by comparing whether or not each promotion plan is implemented between stores. In this embodiment, formula (2) is used as an example. f=(|x1-y1|+|x2-y2|+···+|xn-yn|) / N (2) Here, xn is the implementation status (0 or 1) of the nth promotional plan in store x, yn is the implementation status (0 or 1) of the nth promotional plan in store y, and N is the total number of promotional plans.

[0085] 12B shows an example of similarity coefficient data for the promotional plan implementation status between stores calculated in step S253. In this example, the similarity coefficient between the "Yokohama store" and the "Kawasaki store" is "0.34", the similarity coefficient between the "Yokohama store" and the "Shinagawa store" is "0.45", and the similarity coefficient between the "Kawasaki store" and the "Shinagawa store" is "0.78".

[0086] In step S254, the model generation device 20 (generation unit 22) performs clustering of stores based on the similarity of the promotional plans implemented between the stores. In this embodiment, the clustering of stores is performed by determining the position of each store on the Euclidean plane according to the similarity of the promotional plan implementation status between the stores, and classifying the stores into groups. The clustering of stores is performed using machine learning such as the Kmeans method based on the correlation coefficient of the implementation status of the promotional plans.

[0087] In step S255, the model generation device 20 (generation unit 22) determines the store groups. The store groups are determined by the store groups created by store clustering. Store groups that can secure a certain amount of data when aggregating sales trend data 101 within the store groups are extracted, and data from store groups that cannot secure a sufficient amount of data for one sales promotion plan is excluded because it becomes noise.

[0088] When step S255 is completed, store grouping in generating a rating model is completed.

[0089] FIG. 12C shows an example of the type of store group formed in step S250. The distribution plane for each store and store group is a plane created by treating the correlation coefficient of the implementation status of the promotional plan as Euclidean distance, and expresses only the mutual relationship between the stores. In this example, the stores "Yokohama store" and "Ikebukuro store" are classified into group I, the stores "Shinagawa store", "Chiba store", "Kawasaki store", and "Tokyo store" are classified into group II, and the store "Saitama store" is classified into group III. Stores assigned to each group can be considered to have similar implementation status of the promotional plan. Data from a store group that does not have enough data volume for one promotional plan is excluded because it becomes noise and deteriorates prediction accuracy.

[0090] According to the above-mentioned store grouping, each promotional plan has a similar concept, and because of this similarity, the sales trends of the products handled in the promotional plans are also similar, so learning accuracy can be improved by consolidating the sales data 100 based on the similarity of the implemented promotional plans. In particular, in the case of a large franchise chain, where individual promotional plans are implemented and there are a huge number of types of promotional plans, while there are few stores implementing the same promotional plans, great effects can be obtained by grouping stores in addition to grouping promotional plans.

[0091] 13 shows a subflow of step S260. In step S260, product grouping is performed based on sales trend data 101. Product grouping calculates, for each of a plurality of products, a sales volume characteristic for the number of unit periods in which the sales volume was achieved relative to the sales volume per unit period, and classifies the plurality of products into at least one product group according to the similarity of the sales volume characteristics. In this embodiment, step S260 includes the procedures of tabulating daily sales volumes by product, extracting sales volume characteristics, calculating similarity, clustering, extracting groups, and determining product groups.

[0092] In step S261, the model generation device 20 (generation unit 22) tabulates the daily sales volume for each product. The tabulation of the daily sales volume for each product is performed by extracting the sales volume for each product in each store or each store group for each day based on the sales trend data 101 and the product attribute data 102, and organizing the daily sales volume for each product into data including the number of business days.

[0093] 14A shows an example of the daily sales quantity for each product organized in step S261 from the sales trend data 101. In this example, the first line shows that the sales quantity of the product "Refill Body Soap C" with the POS code "49013013" was "40" units on "January 10, 2023", and the second line shows that the sales quantity of the product "Shampoo E" with the POS code "26135429" was "20" units on "January 10, 2023".

[0094] In step S262, the model generating device 20 (the generating unit 22) extracts the sales quantity characteristic for each product. The sales quantity characteristic is determined for each product from the sales quantity and the number of days during the period when a specific sales quantity was recorded. The sales quantity characteristic has normalized or standardized elements, and is, for example, a histogram of the number of days based on the sales quantity. In this embodiment, the reference period width is set to days, but it may be arbitrarily changed to, for example, several days, weeks, months, or years depending on the purchasing cycle. The similarity is calculated by comparing the sales quantity characteristics between products. By classifying using the sales quantity characteristic instead of the product attribute, it is possible to treat products that have different product attributes but have similar sales quantity characteristics due to the relevance of their use and that may have similar effects from sales promotion planning as a group. For example, sports drinks and sweat wipes have different product attributes, but can be expected to have similar sales quantity characteristics in situations such as high temperature periods, areas where activities are popular, and during exercise promotion fairs.

[0095] FIG. 14B shows an example of the sales volume characteristic for each product extracted in step S262. The histogram on the left shows the sales volume characteristic for "Refill Body Soap C." The characteristic exhibits one high peak and a long tail from the peak to the increase in sales volume. The histogram on the right shows the sales volume characteristic for "Shampoo E." The characteristic exhibits two peaks.

[0096] In step S263, the model generating device 20 (generation unit 22) calculates the degree of similarity of the sales volume characteristics between the products. In this embodiment, the similarity of the sales volume characteristics is determined by formula (3) using Euclidean distance from the difference in the sales volume characteristics for each combination of products. When Euclidean distance is used, the closer the similarity of the sales volume characteristics is to 0, the more similar they are. The sales volume characteristics of different products may be extracted for different reference period widths depending on the product purchasing cycle, and compared to determine the similarity. f = ((x1-y1) 2 +(x2-y2) 2 +···+(xn-yn) 2 ) 1 / 2 (3) Here, xn is the ratio of the number of days in sales volume interval n in the sales volume characteristic of product x to the total number of days in the comparison period, and yn is the ratio of the number of days in sales volume interval n in the sales volume characteristic of product y to the total number of days in the comparison period.

[0097] 14C shows an example of the similarity of sales quantity characteristics between products extracted in step S263. In this example, the similarity of sales quantity characteristics between "Refill Body Soap C" and "Shampoo E" is "28.6", the similarity of sales quantity characteristics between "Refill Body Soap C" and "Solid Laundry Detergent F" is "156.2", and the similarity of sales quantity characteristics between "Shampoo E" and "Solid Laundry Detergent F" is "241.9".

[0098] In step S264, the model generating device 20 (the generating unit 22) performs clustering of products based on the similarity of the sales volume characteristics between the products. In this embodiment, the clustering of products is performed by determining the position of each product on the Euclidean plane according to the mutual similarity of the products, and classifying the products into groups. The clustering of products is performed based on the similarity using machine learning such as the Kmeans method.

[0099] In step S265, the model generation device 20 (generation unit 22) determines the product groups. The product groups are determined by extracting groups that satisfy a standard accuracy from among the product groups created by product clustering. In this embodiment, product groups are extracted when the coefficient of determination is 0.6 or more and the total root mean square error rate is 40% or less. The remaining groups that satisfy the conditions are set as product groups. The standard grouping accuracy may be determined in advance as appropriate.

[0100] Upon completion of step S265, product grouping in the generation of the evaluation model is completed.

[0101] FIG. 14D shows an example of what kind of groups are formed in step S260. The distribution plane of each product and product group is a Euclidean plane in which the similarity of the products is the Euclidean distance, and expresses only the mutual relationships between the products. In this example, the products "Liquid Laundry Detergent B", "Solid Laundry Detergent F", and "Shampoo H" are classified into group A, the products "Toothbrush A" and "Dental Gum D" are classified into group B, and the products "Refill Body Soap C", "Shampoo E", and "Shampoo G" are classified into group C. Here, as an example, products with similar characteristics are assigned to the same group, but products with different properties can also be assigned to the same group.

[0102] According to the above-mentioned product grouping, a plurality of products are classified according to the similarity of sales volume characteristics, and an evaluation model is generated for each product group using the product sales trend data 101 and the sales promotion plan data 103. Therefore, even if the sales volume of each product is small, an evaluation model can be generated with high accuracy using a sufficient amount of sales data 100.

[0103] FIG. 15 shows a subflow of step S270. In step S270, the model generating device 20 (generation unit 22) performs new product classification based on the product groups and product attribute data 102 created in step S260. In new product classification, attributes common to the products classified into each product group are extracted, and products that have not been classified into any group among multiple products are classified into each product group based on the attributes, according to the similarity of sales volume characteristics. In step S270, products with no or little accumulated sales statement data, such as new products or renewed products, are assigned to the product groups created in the product grouping subflow S260. Step S270 consists of the steps of acquiring product attribute data of products belonging to a product group, setting product group classification rules, and assigning new products to product groups.

[0104] In step S271, the model generation device 20 (generation unit 22) judges whether or not a product with a low sales volume that was not included in the product grouping, such as a new product or a renewed product, is included in the target of the sales promotion plan. The judgment of whether to treat a product as having a low sales volume may be based on the sales volume or the percentage of days in which the sales volume included in the sales trend data is less than a specific number. If it is judged that a product with a low sales volume is included, the process proceeds to step S272, and if no product with a low sales volume can be confirmed, the subflow ends.

[0105] In step S272, the model generation device 20 (generation unit 22) acquires attributes of products included in each product group created in the product grouping subflow S260. For example, the product DNA, list price, category, brand, etc. included in the product attribute data 102 are acquired as attributes of each product.

[0106] In step S273, the model generation device 20 (generation unit 22) creates classification rules for the product groups created in S260 based on the features acquired in S272. The classification rules can be set appropriately by extracting product features that many products in a particular product group have in common as features of the product group in order of the number of products that have them in common, and dividing the product groups based on the top few features. In addition, by setting the objective variable to the product group and the explanatory variables to product features including the product DNA, list price, category, etc. of each product, it is possible to automatically create classification rules by machine learning using a method such as random forest.

[0107] In step S274, the model generation device 20 (generation unit 22) can assign products that were not included in the product grouping in step S260 due to low sales volume to product groups in accordance with the classification rules created in S273.

[0108] By omitting or completing step S274, new product classification in generating an evaluation model is completed.

[0109] Fig. 16 shows an example of a method for determining a product group for an unclassified product in step S270. In this example, "product DNA", "list price", and "category" are extracted as attributes for classifying products. A product with product DNA "whitening", list price "1000" yen, and category "moisturizer" is classified into product group C (see Fig. 14D), a product with product DNA "damage care", list price "1800" yen, and category "conditioner" is classified into product group A (see Fig. 14D), and a product with product DNA "moisturizing", list price "700" yen, and category "shampoo" is classified into product group B (see Fig. 14D). In this example, the classification rules for the product groups are to classify in the order of product DNA ("moisturizing," "damage care," "whitening"), list price ("under 500 yen," "over 500 yen and under 1500 yen," "over 1500 yen," "under 1000 yen," "over 1000 yen"), and product category ("shampoo and body wash," "other than shampoo and body wash"), however the items and order of the classification conditions are not limited to this and may be determined appropriately depending on the design of the classification rules.

[0110] According to the above-described new product classification, it is possible to classify products that have not been classified into any group due to their small sales volume, such as new products, among a plurality of products, according to the similarity of sales volume characteristics.

[0111] 17 shows a subflow of step S280. In step S280, the model generation device 20 generates an evaluation model. The evaluation model generation generates an evaluation model of a sales promotion plan using sales trend data 101 and sales promotion plan data 103 tabulated for each product group. Step S280 includes data tabulation for each product group, addition of weekly sales trend data, model construction, and verification.

[0112] In step S281, the model generating device 20 (generation unit 22) extracts the sales quantity for each store group and each product group for each day based on the sales trend data 101 and the sales promotion plan data 103, and aggregates the data to include the sales quantity for each product by day and the promotion implementation status. The aggregated data includes each factor data that may affect the sales quantity obtained from the external data 105. In this example, as shown in FIG. 2B, the aggregated data includes the category indicating the product classification as each factor data that may affect sales, the discount amount at the time of the sales promotion plan, the standard selling price data (e.g., master selling price), the grouped sales promotion plan group name, social event data (e.g., corona event, number of corona infected people, period before tax increase), and weather data including temperature (average temperature, maximum temperature, minimum temperature) and weather.

[0113] In step S282, the model generation device 20 (generation unit 22) adds weekly sales trend data to the data collected in step S281. The weekly sales trend data consists of daily sales quantity data for the six days prior to the date of the data collected in step S281. The weekly sales trend data is used to determine the effect of the day of the week on the number of sales units when building a model.

[0114] In step S283, the model generating device 20 (the generating unit 22) constructs and verifies the evaluation model. The evaluation model may be constructed using, for example, machine learning of a regression model. In this embodiment, a model is constructed using gradient boosting, a machine learning technique, with the sales volume as the objective variable and the factors as the explanatory variables, and cross-validation, specifically Kfold, is executed five times as the verification method, and the coefficient of determination and the root mean square error rate are used as the evaluation method. Models that satisfy the standard accuracy during model construction are extracted, and one of the extracted models is determined as the evaluation model. In this embodiment, a model with a coefficient of determination of 0.6 or more and a total root mean square error rate of 40% or less is extracted as the evaluation model. The evaluation model can be determined by selecting, for example, the model with the highest accuracy standard.

[0115] Upon completion of step S283, the evaluation model parameters 110 shown as an example in FIG. 2C are obtained, and the generation of the evaluation model is completed.

[0116] In step S300, the model generation device 20 (storage unit 23) tabulates and records the evaluation model created in step S200. Since the evaluation model created in step S200 has a coefficient for obtaining a target variable for each factor set as an explanatory variable, the coefficient is recorded as an influence value (SHAP value) for each factor. The closer the SHAP value is to 0, the less the factor affects sales, and the higher the SHAP value, the greater the factor affects sales. For each product group and each promotional plan, the SHAP value of the factor items in parameter 110 is extracted and recorded as the effect of each promotional plan on the sales volume for each product group.

[0117] Upon completion of step S300, model generation is complete.

[0118] 18A shows a flow of evaluating the sales promotion effect. The flow is started when the user inputs a start command on the evaluation device.

[0119] In step S400, the evaluation device 30 (second acquisition unit 31) acquires data corresponding to the promotional plan to be evaluated and the promotional product from the sales data 100 from the store terminal 10, and the generated evaluation model (evaluation model parameters 110) from the storage unit 23 of the model generation device 20. Also, in step S400, the evaluation device 30 (second acquisition unit 31) acquires plan cost data 111 including a breakdown of costs for each promotional plan from each store terminal 10 or the external storage device 50. The plan cost data 111 is as described above with reference to FIG. 3A.

[0120] In step S500, the evaluation device 30 (evaluation unit 32) inputs the sales data 100 acquired in step S400 into the evaluation model to evaluate the promotion effect in terms of the sales volume of each product. In this embodiment, the evaluation unit 32 calculates the increase in sales amount due to each promotion plan from the evaluation model parameters 110 generated by the model generation device 20, and further calculates the return on investment from the increase in sales amount due to each promotion plan and the costs involved in implementing the promotion plan, thereby determining the improvement priority for each promotion plan.

[0121] FIG. 18B shows a sub-flow of step S500.

[0122] In step S510, the evaluation device 30 (evaluation unit 32) extracts the costs associated with implementing the promotional plan from the plan expense data 111, and calculates the allocated costs for each promotional plan by day and product by organizing the implementation costs for the promotional plan by day and product.

[0123] In step S520, the evaluation device 30 (evaluation unit 32) calculates the profit amount and the return on investment (ROI). The profit amount and ROI are calculated by multiplying the sales volume increase for each promotional plan group by the sales amount to calculate the amount increase, and determining the amount from the difference between the amount increase and the allocated cost for each day, product, and promotional plan.

[0124] In step S530, the evaluation device 30 (evaluation unit 32) determines improvement priorities for each product and promotion plan from the ROI and / or quantity increment. The improvement priorities for each product and promotion plan are priorities for which promotion plans should be actively improved for each product. The improvement priorities for each promotion plan can be determined by comparing a predefined ROI threshold and a quantity increment threshold. By using both a monetary-based ROI threshold and a quantity-based quantity increment threshold, the accuracy of the improvement priorities for each promotion plan can be further improved.

[0125] FIG. 19 shows an improvement priority determination table for each product and each promotion plan in this example. In this example, the improvement priority for each product and each promotion plan is determined from the ROI and the net increase (quantity increase ratio), and it is determined that the higher the value, the more immediate the improvement is required. In this example, when the ROI exceeds 100%, the improvement priority is the lowest regardless of the net increase, when the ROI is in the range of 50-100%, the improvement priority is the second highest regardless of the net increase, when the ROI is in the range of 25-50%, the improvement priority is the second highest regardless of the net increase, when the ROI is in the range of 10-25% and the net increase is 20% or more, when the ROI is less than 10% and the net increase is 20% or more, the improvement priority is the second highest, when the ROI is in the range of 10-25% and the net increase is less than 20%, the improvement priority is the second highest, when the ROI is in the range of 10-25% and the net increase is less than 20%, the improvement priority is the second highest, when the ROI is less than 10% and the net increase is less than 20%, the improvement priority is the highest.

[0126] In step S540, the evaluation device 30 (evaluation unit 32) executes sub-evaluation functions such as customer attraction of the promotion plan, optimal selling price in normal sales or discount sales, optimal flyer composition ratio, suggestion of alternative products for low-selling products, etc. Step S540 may be omitted and is executed as appropriate when checking a more detailed evaluation of the promotion plan or an evaluation other than the promotion plan.

[0127] By omitting or completing step S540, the evaluation of the promotion effect in the promotion plan evaluation is completed.

[0128] In step S600, the evaluation device 30 (output unit 33) outputs the promotion effect evaluation data created in step S500. The output method may be determined appropriately depending on the mode of use. The output method may be, for example, numerical output using a table, or illustration using a bar graph or line graph, and may be displayed on a monitor screen or printed on paper.

[0129] When step S600 is completed, the evaluation and output of the effectiveness of the sales promotion plan is completed.

[0130] As described above, according to the sales promotion effect evaluation system 1 of this embodiment, the model generation device 20 acquires and organizes sales data 100 including product sales trend data 101 based on sales records input to the store terminal 10 and store data 104 recording the implementation status of sales promotion plans at each store, generates an evaluation model for evaluating the effect of each sales plan, acquires the evaluation model using the evaluation device 30, determines the ROI and improvement priority of the sales plan based on the variables of the evaluation model, and outputs the requested evaluation result. By aggregating stores, sales plans, and products with similar sales characteristics and increasing the amount of each data that can be used for model generation, it becomes possible to predict the effect of a sales promotion plan with high accuracy even when the information input by the store terminal 10 is insufficient or when targeting a product with little accumulated data, such as a product that has just been on sale.

[0131] For simplicity, the promotional effect evaluation system 1 in this embodiment has been described as evaluating the promotional effect, but when generating the evaluation model, it is possible to calculate the increase in sales volume for each factor and predict the impact on sales during holidays and social events.

[0132] Furthermore, the model generation method according to this embodiment is a method for generating an evaluation model for evaluating a promotional effect on the sales volume of a product, and includes a step S100 of acquiring sales data 100 regarding a plurality of products sold at each store from a store terminal 10 installed in each store, the sales data 100 including sales trend data 101 regarding the trend in sales volume of each of the plurality of products, product attribute data 102 regarding the attributes of each of the plurality of products, and promotional plan data 103 regarding implemented promotional plans, and a step S200 of calculating, for each of the plurality of products, sales volume characteristics regarding the number of unit periods in which a sales volume was achieved relative to the sales volume per unit period, classifying the plurality of products into at least one product group according to the similarity of the sales volume characteristics, and generating an evaluation model using the sales trend data 101 and the promotional plan data 103 regarding the products classified into each product group from the sales data 100. According to this, by classifying a plurality of products according to the similarity of sales volume characteristics and generating an evaluation model for each product group using product sales trend data 101 and sales promotion plan data 103, even when the sales volume of each product is small, an evaluation model can be generated with high accuracy using a sufficient amount of sales data 100.

[0133] The evaluation method according to the present embodiment is a method for evaluating the promotion effect in terms of the sales volume of a product, and includes a step S400 of acquiring sales data 100 on at least one product sold at each store from a store terminal 10 installed in each store, the sales data 100 including sales trend data 101 on the trend of the sales volume of the at least one product, product attribute data 102 on the attributes of the at least one product, and promotion plan data 103 on the implemented promotion plan, and a step S500 of inputting the promotion plan data 103 into the evaluation model generated by the above-mentioned model generation method to evaluate the promotion effect in terms of the sales volume of the at least one product. According to this, the profit amount and return on investment of the promotion plan can be estimated by evaluating the promotion effect from the parameters of the evaluation model generated with high accuracy, and it is possible to determine the improvement priority, optimal price, optimal flyer composition ratio, target of product replacement, etc. for each promotion plan.

[0134] The model according to this embodiment is an evaluation model for evaluating the promotion effect in terms of the sales volume of a product, which is generated by the above-mentioned model generation method. According to this, by classifying a plurality of products according to the similarity of the sales volume characteristics and generating an evaluation model for each product group using the sales trend data 101 and the sales promotion plan data 103 of the products, a highly accurate evaluation model can be generated using a sufficient amount of sales data 100, even when the sales volume of each product is small.

[0135] In addition, the model generation device according to this embodiment is a model generation device 20 that generates an evaluation model for evaluating a promotional effect on the sales volume of a product, and includes a first acquisition unit 21 that acquires sales data 100 regarding a plurality of products sold at each store from a store terminal 10 installed in each store, the sales data 100 including sales trend data 101 regarding the trend in sales volume of each of the plurality of products, product attribute data 102 regarding the attributes of each of the plurality of products, and promotional plan data 103 regarding implemented promotional plans, and a generation unit 22 that calculates sales volume characteristics for each of the plurality of products regarding the number of unit periods in which the sales volume was achieved relative to the sales volume per unit period, classifies the plurality of products into at least one product group according to the similarity of the sales volume characteristics, and generates an evaluation model using the sales trend data 101 and the promotional plan data 103 for the products classified into each product group from the sales data 100. According to this, by classifying a plurality of products according to the similarity of sales volume characteristics and generating an evaluation model for each product group using product sales trend data 101 and sales promotion plan data 103, even when the sales volume of each product is small, an evaluation model can be generated with high accuracy using a sufficient amount of sales data 100.

[0136] The evaluation device according to the present embodiment is an evaluation device 30 for evaluating the promotion effect in the sales volume of a product, and includes a second acquisition unit 31 for acquiring sales data 100 on at least one product sold at each store from a store terminal installed in each store, the sales data 100 including sales trend data 101 on the trend of the sales volume of the at least one product, product attribute data 102 on the attributes of the at least one product, and promotion plan data 103 on the implemented promotion plan, and an evaluation unit 32 for inputting the sales data 100 into the evaluation model generated by the above-mentioned model generation device and evaluating the promotion effect in the sales volume of the at least one product. According to this, the profit amount and return on investment of the promotion plan can be estimated by evaluating the promotion effect from the parameters of the evaluation model generated with high accuracy, and it is possible to determine the improvement priority, optimal price, optimal flyer composition ratio, target of product replacement, etc. for each promotion plan.

[0137] Various embodiments of the present invention may be described with reference to flowcharts and block diagrams, where the blocks may represent (1) stages of a process in which operations are performed or (2) sections of an apparatus responsible for performing the operations. Particular stages and sections may be implemented by dedicated circuitry, programmable circuitry provided with computer readable instructions stored on a computer readable medium, and / or a processor provided with computer readable instructions stored on a computer readable medium. Dedicated circuitry may include digital and / or analog hardware circuitry and may include integrated circuits (ICs) and / or discrete circuits. Programmable circuitry may include reconfigurable hardware circuitry including logical AND, logical OR, logical XOR, logical NAND, logical NOR, and other logical operations, memory elements such as flip-flops, registers, field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), and the like.

[0138] A computer readable medium may include any tangible device capable of storing instructions that are executed by a suitable device, such that the computer readable medium having instructions stored thereon comprises an article of manufacture that includes instructions that can be executed to create means for performing the operations specified in the flowcharts or block diagrams. Examples of computer readable media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, etc. More specific examples of computer readable media may include floppy disks, diskettes, hard disks, random access memories (RAMs), read-only memories (ROMs), erasable programmable read-only memories (EPROMs or flash memories), electrically erasable programmable read-only memories (EEPROMs), static random access memories (SRAMs), compact disk read-only memories (CD-ROMs), digital versatile disks (DVDs), Blu-ray (RTM) disks, memory sticks, integrated circuit cards, etc.

[0139] The computer readable instructions may include either assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk®, JAVA®, C++, etc., and conventional procedural programming languages ​​such as the “C” programming language or similar programming languages.

[0140] Computer readable instructions may be provided to a processor or programmable circuitry of a programmable data processing apparatus, such as a general purpose computer, special purpose computer, or other computer, either locally or over a wide area network (WAN) such as a local area network (LAN), the Internet, etc., to execute the computer readable instructions to create means for performing the operations specified in the flowcharts or block diagrams. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, etc.

[0141] 20 illustrates an example of a computer 1200 in which aspects of the present invention may be embodied in whole or in part. Programs installed on the computer 1200 may cause the computer 1200 to function as or perform operations associated with an apparatus or one or more sections of the apparatus according to an embodiment of the present invention, and / or to perform a process or steps of the process according to an embodiment of the present invention. Such programs may be executed by the CPU 1212 to cause the computer 1200 to perform specific operations associated with some or all of the blocks of the flowcharts and block diagrams described herein.

[0142] The computer 1200 according to this embodiment includes a CPU 1212, a RAM 1214, a graphic controller 1216, and a display device 1218, which are connected to each other by a host controller 1210. The computer 1200 also includes input / output units such as a communication interface 1222, a hard disk drive 1224, a DVD-ROM drive 1226, and an IC card drive, which are connected to the host controller 1210 via an input / output controller 1220. The computer also includes legacy input / output units such as a ROM 1230 and a keyboard 1242, which are connected to the input / output controller 1220 via an input / output chip 1240.

[0143] The CPU 1212 operates according to a program stored in the ROM 1230 and the RAM 1214, thereby controlling each unit. The graphic controller 1216 retrieves image data generated by the CPU 1212 into a frame buffer or the like provided in the RAM 1214 or into the graphic controller 1216 itself, and causes the image data to be displayed on the display device 1218.

[0144] The communication interface 1222 communicates with other electronic devices via a network. The hard disk drive 1224 stores programs and data used by the CPU 1212 in the computer 1200. The DVD-ROM drive 1226 reads programs or data from a DVD-ROM 1227 and provides the programs or data to the hard disk drive 1224 via the RAM 1214. The IC card drive reads programs and data from an IC card and / or writes programs and data to an IC card.

[0145] The ROM 1230 stores therein a boot program or the like executed by the computer 1200 upon activation and / or a program that depends on the hardware of the computer 1200. The input / output chip 1240 may also connect various input / output units to the input / output controller 1220 via a parallel port, a serial port, a keyboard port, a mouse port, etc.

[0146] The programs are provided by a computer-readable medium such as a DVD-ROM 1227 or an IC card. The programs are read from the computer-readable medium, installed in the hard disk drive 1224, the RAM 1214, or the ROM 1230, which are also examples of computer-readable media, and executed by the CPU 1212. Information processing described in these programs is read by the computer 1200, and brings about cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be constructed by realizing information manipulation or processing according to the use of the computer 1200.

[0147] For example, when communication is performed between the computer 1200 and an external device, the CPU 1212 may execute a communication program loaded in the RAM 1214 and instruct the communication interface 1222 to perform communication processing based on the processing described in the communication program. Under the control of the CPU 1212, the communication interface 1222 reads transmission data stored in a transmission buffer processing area provided in the RAM 1214, the hard disk drive 1224, the DVD-ROM 1227, or a recording medium such as an IC card, and transmits the read transmission data to a network, or writes reception data received from the network to a reception buffer processing area or the like provided on the recording medium.

[0148] Furthermore, the CPU 1212 may cause all or a necessary portion of a file or database stored in an external recording medium such as a hard disk drive 1224, a DVD-ROM drive 1226 (DVD-ROM 1227), an IC card, etc. to be read into the RAM 1214, and perform various types of processing on the data on the RAM 1214. The CPU 1212 then writes back the processed data to the external recording medium.

[0149] Various types of information, such as various types of programs, data, tables, and databases, may be stored in the recording medium and undergo information processing. The CPU 1212 may perform various types of processing on the data read from the RAM 1214, including various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, information search / replacement, etc., as described throughout this disclosure and specified by the instruction sequence of the program, and write back the results to the RAM 1214. The CPU 1212 may also search for information in a file, database, etc. in the recording medium. For example, when a plurality of entries each having an attribute value of a first attribute associated with an attribute value of a second attribute are stored in the recording medium, the CPU 1212 may search for an entry that matches a condition, in which the attribute value of the first attribute is specified, from among the plurality of entries, read the attribute value of the second attribute stored in the entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0150] The above-described programs or software modules may be stored in a computer-readable medium on the computer 1200 or in the vicinity of the computer 1200. In addition, a recording medium such as a hard disk or a RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer-readable medium, thereby providing the programs to the computer 1200 via the network.

[0151] Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It is clear to those skilled in the art that various modifications and improvements can be made to the above embodiments. It is clear from the description of the claims that such modifications and improvements can also be included in the technical scope of the present invention.

[0152] It should be noted that the order of execution of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not specifically stated as "before," "prior to," etc., and may be realized in any order unless the output of a previous process is used in a later process. Even if the operational flow in the claims, specifications, and drawings is explained using "first," "next," etc. for convenience, it does not mean that it is essential to perform the process in this order. [Explanation of symbols]

[0153] 1...sales promotion effect evaluation system, 10...store terminal, 20...model generation device, 21...first acquisition unit, 22...generation unit, 23...storage unit, 30...evaluation device, 31...second acquisition unit, 32...evaluation unit, 33...output unit, 40...external database, 50...external storage device, 60...network, 100...sales data, 101...sales trend data, 102...product attribute data, 103...sales promotion plan data, 104...store data, 105...external data, 106...calendar data, 107...weather data, 108...event data, 109... Input data, 110...evaluation model parameters, 111...projected cost data, 112...evaluation data, 1200...computer, 1210...host controller, 1212...CPU, 1214...RAM, 1216...graphics controller, 1218...display device, 1220...input / output controller, 1222...communications interface, 1224...hard disk drive, 1226...DVD-ROM drive, 1230...ROM, 1240...input / output chip, 1242...keyboard.

Claims

1. A model generation method for generating an evaluation model for evaluating a promotion effect on a sales volume of a product, comprising the steps of: acquiring sales data related to a plurality of products sold at at least one store from a store terminal installed in the at least one store, the sales data including sales trend data related to a trend in sales quantity of each of the plurality of products, product attribute data related to attributes of each of the plurality of products, and sales promotion plan data related to implemented sales promotion plans; a step of calculating a sales volume characteristic for each of the plurality of products with respect to the sales volume per unit period and the number of unit periods in which the sales volume was achieved, classifying the plurality of products into at least one product group according to the similarity of the sales volume characteristics, and generating the evaluation model using sales trend data and sales promotion plan data for the products classified into the at least one product group among the sales data; A model generation method comprising:

2. 2. The model generation method according to claim 1, wherein in the generating step, attributes common to the products classified into the at least one group are extracted, and based on the attributes, products among the plurality of products that have not been classified into any group according to the similarity of the sales volume characteristics are classified into the at least one group.

3. The model generation method described in claim 1, wherein, in the generating step, the sales volume characteristics of products handled in each of a plurality of promotional plans implemented in the at least one store are calculated based on the promotional plan data, the plurality of promotional plans are classified into at least one promotional plan group according to the similarity of the sales volume characteristics, and the evaluation model is generated using sales trend data and promotional plan data for products handled in the promotional plans classified into the at least one promotional plan group among the sales data.

4. In the acquiring step, the sales data including store data regarding the implementation status of a sales promotion plan in the store is acquired from the store terminals installed in a plurality of stores; The model generation method of claim 1, wherein in the generating step, a similarity of promotional plans implemented at each of the plurality of stores is evaluated based on the promotional plan data, the plurality of stores are classified into at least one store group according to the similarity, and the evaluation model is generated using sales data of the stores classified into the at least one store group among the sales data.

5. 2. The model generation method of claim 1, wherein the sales volume characteristic is a histogram of the number of unit periods of sales based on the sales volume of the product, and the histogram is determined by normalizing or standardizing for each product based on the sales volume of the product per unit period and the number of unit periods in which a specific sales volume was recorded.

6. The model generating method according to claim 1 , wherein the evaluation model determines a standard selling price from a mode of selling prices of the at least one product based on the sales data.

7. The model generating method according to claim 1 , wherein the evaluation model determines a period during which a sales promotion plan is to be implemented based on a change in selling price and / or sales volume of the at least one commodity on the basis of the sales data.

8. The model generating method according to claim 1 , wherein the sales trend data includes data on the number of sales items of an item for each of the unit periods.

9. The model generating method according to claim 1 , wherein the product attribute data includes information regarding at least one of a product category, a manufacturer, a brand name, a product name, a product DNA, a selling price, and a selling date.

10. The model generating method according to claim 1 , wherein the sales promotion plan data includes information on at least one of a plan type, a plan name, a start date, an end date, and a product to be handled.

11. The model generating method according to claim 4 , wherein the store data includes information on at least one of the store name, location, and implemented sales promotion plans.

12. The method of claim 1 , wherein the sales data further includes at least one of calendar data relating to a calendar of days the store is open, weather data relating to weather during days the store is open, and event data relating to social events during days the store is open.

13. A method for evaluating a promotion effect in terms of product sales volume, comprising the steps of: acquiring sales data from a store terminal installed in at least one store, the sales data including sales trend data relating to a trend in sales volume of the at least one product, product attribute data relating to attributes of the at least one product, and sales promotion plan data relating to implemented sales promotion plans; a step of inputting the sales data into an evaluation model generated by the model generation method according to claim 1 to evaluate a sales promotion effect in terms of sales volume of the at least one commodity; The evaluation method includes:

14. The evaluation method according to claim 13, further comprising a step of calculating a promotion effect in a sales amount of the at least one commodity based on the promotion effect in a sales quantity of the at least one commodity.

15. The evaluation method according to claim 13, further comprising a step of calculating a return on investment of a promotional plan for the at least one product based on the promotional effect in the sales volume of the at least one product, and calculating improvement priorities for each product and each promotional plan based on the promotional effect in the sales volume of the at least one product and / or the return on investment of the promotional plan for the at least one product.

16. An evaluation model generated by the model generation method according to claim 1.

17. A model generation device for generating an evaluation model for evaluating a sales promotion effect on a sales volume of a product, comprising: a first acquisition unit that acquires, from a store terminal installed in at least one store, sales data related to a plurality of products sold in the at least one store, the sales data including sales trend data related to a trend in sales quantity of each of the plurality of products, product attribute data related to attributes of each of the plurality of products, and sales promotion plan data related to an implemented sales promotion plan; a generation unit that calculates, for each of the plurality of products, a sales volume characteristic with respect to the number of unit periods in which the sales volume was achieved relative to the sales volume per unit period, classifies the plurality of products into at least one product group according to the similarity of the sales volume characteristics, and generates the evaluation model using sales trend data and sales promotion plan data for the products classified into the at least one product group from the sales data; A model generating device comprising:

18. An evaluation device for evaluating a promotion effect in terms of product sales volume, comprising: a second acquisition unit that acquires, from a store terminal installed in at least one store, sales data related to at least one product sold in the at least one store, the sales data including sales trend data related to a trend in sales quantity of the at least one product, product attribute data related to attributes of the at least one product, and sales promotion plan data related to implemented sales promotion plans; an evaluation unit that inputs the sales data into an evaluation model generated by the model generation device according to claim 17 and evaluates a sales promotion effect in terms of the sales volume of the at least one commodity; An evaluation device comprising:

19. A computer is provided to generate an evaluation model for evaluating the effect of a sales promotion on the sales volume of a product. acquiring sales data from a store terminal installed in at least one store, the sales data including sales trend data relating to a trend in sales quantity of each of the plurality of products, product attribute data relating to attributes of each of the plurality of products, and sales promotion plan data relating to implemented sales promotion plans; a step of calculating a sales volume characteristic for each of the plurality of products with respect to the sales volume per unit period and the number of unit periods in which the sales volume was achieved, classifying the plurality of products into at least one product group according to the similarity of the sales volume characteristics, and generating the evaluation model using sales trend data and sales promotion plan data for the products classified into the at least one product group from among the sales data; A program that executes the following.

20. In order to evaluate the effect of promotion on the sales volume of products, a step of acquiring sales data related to at least one product sold in at least one store from a store terminal installed in the at least one store, the sales data including sales trend data related to a trend in sales volume of the at least one product, product attribute data related to attributes of the at least one product, and sales promotion plan data related to implemented sales promotion plans; a step of inputting the sales data into an evaluation model generated by executing the program according to claim 19, and evaluating a sales promotion effect in the sales volume of the at least one product; A program that executes the following.

Images