Information processing device, information processing method, and information processing program

Abstract

This information processing device is provided with: an acquisition unit that acquires multiple pieces of movement path data representing the in-store movement path of a customer vising a store for each visit to the store; a calculation unit that calculates a variation in route of the customer in the store on the basis of the multiple pieces of movement path data acquired by the acquisition unit; and a determination unit that determines whether the customer exhibits planned purchase behavior or unplanned purchase behavior on the basis of the variation in route calculated by the calculation unit.

Description

Information processing device, information processing method, and information processing program

[0001] The disclosed technology relates to an information processing device, an information processing method, and an information processing program.

[0002] Conventionally, there are known techniques for estimating the purchasing intentions of customers who visit a store (for example, see Non-Patent Document 1). The technique disclosed in Non-Patent Document 1 is a technique for estimating a customer's purchasing intentions based on the customer's time spent in front of shelves where products are displayed, their line of sight, and their movements. Non-Patent Document 1 estimates whether a customer's purchasing behavior is planned or unplanned. As also disclosed in Non-Patent Document 1, planned purchases refer to purchasing behavior in which the customer has already decided which products to purchase before entering the store. On the other hand, unplanned purchases refer to purchasing behavior in which the customer decides which products to purchase after entering the store.

[0003] Sugiyama, Study on Behavioral Characteristics of Non-Planned Purchasers in Front of Shelves, Vol.10, pp.1-8 (UBI 2019.5)

[0004] The technology disclosed in Non-Patent Document 1 analyzes customer characteristics on a store visit basis or on a product basis. Specifically, the technology disclosed in Non-Patent Document 1 estimates whether a single purchase behavior of a customer is a planned purchase or an unplanned purchase, but does not estimate the purchasing characteristics that each customer consistently possesses. For example, it is thought that there are individual tendencies among customers, such as those that make them more likely to make planned purchases or unplanned purchases.

[0005] Therefore, the technique of Non-Patent Document 1 has a problem in that it is not possible to estimate the purchasing characteristics of each customer.

[0006] The disclosed technology has been made in consideration of the above points, and aims to provide an information processing device, an information processing method, and an information processing program that can estimate whether a customer tends to engage in planned purchasing behavior or unplanned purchasing behavior.

[0007] A first aspect of the present disclosure is an information processing device that includes an acquisition unit that acquires multiple pieces of flow line data that represent in-store flow lines of customers when they visit a store and represent the in-store flow lines of customers for each visit to the store; a calculation unit that calculates variability in the paths of the customers within the store based on the multiple pieces of flow line data acquired by the acquisition unit; and a determination unit that determines whether the customers have a tendency to make planned purchases or a tendency to make unplanned purchases based on the variability in the paths calculated by the calculation unit.

[0008] A second aspect of the present disclosure is an information processing method in which a computer executes a process of acquiring a plurality of pieces of flow line data representing in-store flow lines of customers when they visit a store and representing the in-store flow lines for each visit to the store, calculating variability in the paths of the customers within the store based on the acquired plurality of pieces of flow line data, and determining whether the customers have a tendency to make planned purchases or a tendency to make unplanned purchases based on the calculated variability in the paths.

[0009] According to the disclosed technology, it is possible to obtain an effect of being able to estimate whether a customer has a tendency to take planned purchasing behavior or whether a customer has a tendency to take unplanned purchasing behavior.

[0010] 1 is a block diagram showing an example of a hardware configuration of an information processing device according to an embodiment. FIG. 2 is a block diagram showing an example of a functional configuration of an information processing device according to an embodiment. FIG. 3 is a diagram for explaining planned purchasing and unplanned purchasing. FIG. 4 is a diagram for explaining variety seeking. FIG. 5 is a diagram showing an example of data stored in a data storage unit 100. FIG. 6 is a diagram showing an example of data stored in a data storage unit 100. FIG. 7 is a diagram for explaining flow line data for each visit of a certain customer to a store. FIG. 8 is a diagram for explaining in-store area setting. FIG. 9 is a diagram for explaining an example of transition data. FIG. 10 is a diagram for explaining an example in which the flow line data to be calculated is limited to daytime hours. FIG. 11 is a scatter diagram showing the relationship between route variance and customer unit price. FIG. 12 is a diagram for explaining a process of sorting customer IDs in order of route variance value. FIG. 13 is a diagram for explaining areas where customers with large route variance are likely to stop by. FIG. 14 is a diagram for explaining the correlation between route variance and customer unit price. FIG. 15 is a diagram for explaining a comparison between a customer group with small route variance and a customer group with large route variance. FIG. 16 is a scatter diagram showing the relationship between route variance and various values ​​according to the present embodiment.

[0011] An example of an embodiment of the disclosed technology will be described below with reference to the drawings. Note that in each drawing, the same or equivalent components and parts are given the same reference numerals. Also, the dimensional proportions in the drawings are exaggerated for the convenience of explanation and may differ from the actual proportions.

[0012] First, the hardware configuration of an information processing device 10 according to this embodiment will be described with reference to FIG.

[0013] FIG. 1 is a block diagram showing an example of the hardware configuration of an information processing apparatus 10 according to this embodiment.

[0014] 1, the information processing device 10 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a storage 14, an input unit 15, a display unit 16, and a communication interface (I / F) 17. Each component is connected to each other via a bus 18 so as to be able to communicate with each other.

[0015] The CPU 11 is a central processing unit that executes various programs and controls each part. That is, the CPU 11 reads programs from the ROM 12 or the storage 14 and executes the programs using the RAM 13 as a work area. The CPU 11 controls the above components and performs various arithmetic processing in accordance with the programs stored in the ROM 12 or the storage 14. In this embodiment, the ROM 12 or the storage 14 stores a learning program and an estimation program.

[0016] The ROM 12 stores various programs and various data. The RAM 13 temporarily stores programs or data as a working area. The storage 14 is configured with an HDD (Hard Disk Drive) or an SSD (Solid State Drive) and stores various programs including the operating system and various data.

[0017] The input unit 15 includes a pointing device such as a mouse and a keyboard, and is used to input various information to the device itself.

[0018] The display unit 16 is, for example, a liquid crystal display, and displays various information. The display unit 16 may be a touch panel type and function as the input unit 15.

[0019] The communication interface 17 is an interface for the device itself to communicate with other external devices. For this communication, for example, a wired communication standard such as Ethernet (registered trademark) or FDDI (Fiber Distributed Data Interface) or a wireless communication standard such as 4G, 5G, or Wi-Fi (registered trademark) is used.

[0020] The information processing apparatus 10 according to this embodiment is implemented as a general-purpose computer device such as a server computer or a personal computer (PC).

[0021] Next, the functional configuration of the information processing device 10 will be described with reference to FIG.

[0022] FIG. 2 is a block diagram showing an example of the functional configuration of the information processing device 10 according to this embodiment.

[0023] 2, the information processing device 10 includes, as functional components, a data storage unit 100, an acquisition unit 102, a calculation unit 104, a determination unit 106, and an output unit 108. Each functional component is realized by the CPU 11 reading out an information processing program stored in the ROM 12 or the storage 14, expanding the program in the RAM 13, and executing the program.

[0024] FIG. 3 is a diagram for explaining planned purchases and unplanned purchases. As shown in FIG. 3, planned purchases are purchase behaviors in which a customer decides which product to purchase before entering the store. On the other hand, as shown in FIG. 3, unplanned purchases are purchase behaviors in which a customer decides which product to purchase after entering the store. As shown in FIG. 3, in planned purchases, a customer decides to purchase a specific product before entering the store. In contrast, in unplanned purchases, a customer considers which product to purchase while in the store and ends up purchasing a certain product. In this case, promoting a certain product to a customer who tends to make planned purchases is unlikely to be effective. On the other hand, promoting a certain product to a customer who tends to make unplanned purchases is likely to be highly effective. For this reason, in sales promotion planning, it is important to classify customers as either in a planned purchase state or an unplanned purchase state.

[0025] In particular, customers who have a tendency to make unplanned purchases do not have specific products to purchase and select products in the store, so it is thought that sales promotion measures are more likely to induce them to purchase products. Therefore, it is thought to be useful to use data obtained in the store to estimate whether customers have a tendency to make unplanned purchases or a tendency to make planned purchases.

[0026] As described above, the technology disclosed in Non-Patent Document 1 is a technology for estimating whether a single purchase behavior of a customer is a planned purchase or an unplanned purchase, and is not a technology for estimating the consistent purchasing characteristics of a customer. Therefore, when adopting the conventional technology disclosed in Non-Patent Document 1, the only measures that can be proposed to a store are those based on trends in the store, product, and location (for example, measures for the entire store rather than for each customer) or measures that require a system for estimating customer behavior in real time (for example, a measure that observes customer behavior in front of shelves in the store and displays signage advertisements on the spot).

[0027] On the other hand, if it were possible to estimate customer characteristics such as whether they have a strong tendency to make planned purchases or a strong tendency to make unplanned purchases, the range of possible measures would be wider, such as formulating and appealing to those customers themselves. For example, measures such as offering coupons to customers with a strong tendency to make unplanned purchases could be considered.

[0028] Variety-seeking is a commonly known purchasing tendency associated with unplanned purchasing. Variety-seeking is also known as the pursuit of diversity through consumption. Customers with variety-seeking tendencies (hereinafter simply referred to as "variety-seeking customers") often do not limit themselves to a specific product or product brand, but prefer to purchase and consume a variety of brands. Because variety-seeking customers tend to purchase products that are different from their usual ones, they are thought to be more likely to make unplanned purchases. For this reason, customers with variety-seeking characteristics can also be said to have unplanned purchasing tendencies.

[0029] Therefore, if variety-seeking customers can be extracted, it is believed that it will be possible to extract customers who are consistently more likely to make unplanned purchases.

[0030] Figure 4 is a diagram for explaining variety-seeking. As shown in Figure 4, variety-seeking customers are often in a psychological state in a store, thinking, "I've never bought that product before, so I'll try it," or "I bought that product last time, so I'll try a different product this time." For this reason, variety-seeking customers tend to change their in-store movement paths each time they visit a store.

[0031] Reference numerals 200 and 202 in FIG. 4 represent customer traffic routes when the store is viewed from above. The squares in FIG. 4 represent fixtures (e.g., product shelves) installed in the store, and the lines in FIG. 4 represent customer traffic routes within the store. Note that different line types represent different customer transitions between areas. As shown in FIG. 4, the in-store traffic routes 200 of variety-seeking customers are significantly different between the in-store traffic routes 200X when they visit the store on one day and the in-store traffic routes 200Y when they visit the store on another day. On the other hand, as shown in FIG. 4, the in-store traffic routes 202 of customers who do not have a variety-seeking tendency tend to be approximately the same as the in-store traffic routes 202X when they visit the store on one day and the in-store traffic routes 202Y when they visit the store on another day.

[0032] As such, variety-seeking customers tend to explore products in the store, and their in-store movement patterns vary greatly from visit to visit. However, no previous research on variety-seeking has been conducted that relates to in-store movement patterns in physical stores.

[0033] Therefore, the information processing device 10 of this embodiment collects multiple pieces of flow-line data representing the in-store flow lines of customers when they visit a store, and calculates the variability of the customer's in-store path based on the multiple pieces of flow-line data collected for each customer. The information processing device 10 then determines whether the customer is a variety-seeking customer based on the calculated path variability. If the customer is a variety-seeking customer, the information processing device 10 determines that the customer has a tendency to unplanned purchases. On the other hand, if the customer is not a variety-seeking customer, the information processing device 10 determines that the customer has a tendency to planned purchases. This allows the present embodiment to estimate, for each customer, whether the customer has a tendency to engage in planned purchasing behavior or unplanned purchasing behavior. Hereinafter, path variability will also be referred to simply as "path variance." Route variance is an example of in-store path variance.

[0034] If we can link in-store traffic flow with variety-seeking, we can show the relationship between customers and unplanned purchases from in-store traffic flow, which may lead to a better understanding of customers and the development of store improvements and measures.

[0035] The data storage unit 100 stores, for each of a plurality of customers, flow line data that represents the flow line of the customer within the store when the customer visits the store and the flow line of the customer within the store for each visit. FIGS. 5 and 6 show examples of data stored in the data storage unit 100. As shown in FIG. 5, for example, a store visit date and time table T1 and a flow line data table T2 are stored in the data storage unit 100. When a customer visits the store, a store visit ID is assigned, and the store visit date and time is stored in the store visit date and time table T1. Furthermore, the customer's flow line data at the time of the visit is stored in the flow line data table T2.

[0036] The data storage unit 100 may store, for example, a table T3 as shown in Fig. 6. In this case, when a customer visits a store, a store visit ID is assigned, and the store visit date and time and flow line data are stored in the table T3.

[0037] Fig. 7 is a diagram for explaining the flow line data of a certain customer each time they visit the store. Like Fig. 4, Fig. 7 is a diagram showing the interior of the store as seen from above. As shown in Fig. 7, each line of a different type represents the flow line data of a certain customer each time they visit the store. Such multiple pieces of flow line data, collected in advance for each customer, are stored in the data storage unit 100.

[0038] In this way, the data storage unit 100 stores a plurality of flow line data for each customer and each visit to the store.

[0039] The flow line data represents the path that a customer takes in the store. As will be described later, the flow line data is converted into transition data.

[0040] The acquiring unit 102 acquires a plurality of pieces of flow line data for each customer from the data storage unit 100 .

[0041] The calculation unit 104 calculates, for each of a plurality of customers, the route dispersion of the customer within the store based on the plurality of flow line data acquired by the acquisition unit 102 .

[0042] Specifically, first, the calculation unit 104 generates transition data representing the customer's transition between areas within the store for each of the multiple traffic flow data acquired by the acquisition unit 102, and calculates the route variance of the customer based on the multiple generated transition data.

[0043] There are several possible methods for converting flow line data into transition data, each of which is a combination of (1) an area setting method, (2) a transition condition, and (3) a transition extraction method. These methods will be described in detail below.

[0044] (1) Area setting methods There are roughly two ways to set up areas within a store: However, there is a high degree of freedom in how areas are divided, and other patterns are also possible, such as treating multiple independent areas as the same.

[0045] (A) Method of Setting a Specific Location as an Area In this method, a specific location within a store is set as an area to be transitioned. This method allows for the existence of locations within the store where no area is set. FIG. 8 is a diagram for explaining the setting of areas within a store. Like FIG. 4, FIG. 8 is an aerial view of the store. The hatched areas in FIG. 8 represent the locations of fixtures and installations within the store. For example, as shown by 204X in FIG. 8, a specific location within the store is set as an area to be transitioned. In this case, for example, an area is set in front of shelves within the store, and no area is set for aisles or the like where shelves are not located. In this case, transition data for locations where no area is set may be treated as other areas, or may be not recorded or deleted.

[0046] (B) A method of dividing the entire store into areas: This method divides all locations in the store at equal intervals, or sets areas for all locations in the store while taking into account the fixtures and installations in the store. In this case, each divided location is treated as an area to be transitioned. For example, as shown in 204Y in FIG. 8, when the store is divided into predetermined regions, each area is set as an area to be transitioned.

[0047] Next, the transition conditions between areas will be described.

[0048] (2) Transition Conditions There are roughly two possible transition conditions, but any transition condition may be set as long as it is possible to determine that a customer has transitioned to an area.

[0049] (A) Transition by Passage This method determines that a customer has transitioned to a predetermined area when the coordinates corresponding to the customer's flow line data indicate that the customer has entered that area. A problem that arises here is when multiple transitions are recorded in adjacent areas due to noise, etc., contained in the flow line data. To avoid this, as will be described later, entry into an area may be recorded when there is continuous stay in the same area, rather than a simple single entry into the area.

[0050] (B) Transition by Stay This method determines that a customer has transitioned to a pre-set area when the coordinates corresponding to the flow line data indicate that the customer has stayed in that area. There are several possible methods for determining stay. One method is to determine that a customer is staying in a specified area if the customer's stay time in that area exceeds a certain value. Another method is to determine that a customer is staying in a specified area if the customer's movement volume within that area is below a certain value for a certain period of time. When measuring the amount of customer movement, it is possible to reduce noise by using a moving average or smoothing of the movement volume.

[0051] Next, a method for extracting transitions between areas will be described.

[0052] (3) Method of Extracting Transitions Based on the above-described transition conditions, further necessary information is extracted from the information on the area determined to be a transition of the customer. There are roughly two possible methods for this:

[0053] (A) Extraction of the first transition area This method records only the area to which a customer first transitions. For example, suppose a customer transitions through areas A, B, and C in the order "A → B → C → B → A." In this case, one possible method is to exclude the second transitions, B and A, and extract only the first transition area, "A → B → C." This extraction makes it possible to perform processing that focuses on the customer's initial movements after entering the store.

[0054] (B) Extraction of cumulative transition areas This method records the areas to which a customer has transitioned. In this case, if a customer passes through areas A, B, and C in the order of "A → B → C → B → A," the area is recorded as "A → B → C → B → A."

[0055] Next, the area transition information converted by the above conversion is converted into a character string. For example, the area transition information "A → B → C → B → A" is converted into the character string "ABCBA." Here, one area needs to be represented by a unique single character.

[0056] For example, if the area to be transitioned is expressed in the form of a food area, a fresh fish area, and a confectionery area, "food → fresh fish → confectionery → fresh fish → confectionery" cannot be converted to "food, fresh fish, confectionery, fresh fish, confectionery."

[0057] Also, two areas that are treated independently should not be represented by the same letter. For example, two different transition target areas should not both be represented by the same letter C. However, there is no problem if different areas are treated as the same area. In such cases, they must be converted into a single unique letter. For example, food area = K, fresh fish area = L, confectionery area = M, and so on, and converted to "KLMLK."

[0058] In this way, the calculation unit 104 converts the customer flow line data into transition data that represents transitions between areas.

[0059] FIG. 9 is a diagram illustrating an example of transition data according to this embodiment. As shown in 206X in FIG. 9 , the flow line data is expressed in the form of coordinate data (x1, y1) (x2, y2). The calculation unit 104 converts the flow line data shown in 206X in FIG. 9 into transition data shown in 206Y in FIG. 9 . As shown in 206Y in FIG. 9 , the transition data represents transitions between areas within a store. In this case, as described above, each area within the store is represented by a single character, and the transition data is represented by a character string. In the example of FIG. 9 , the entrance is represented by "entrance," and the other areas within the store are represented by a single character, and the transition data is expressed as [entrance, I, E, I, H, G, B, G, H, H, entrance]. For example, when extracting the first transition area as described above, the calculation unit 104 converts the transition data [entrance, I, E, I, H, G, B, G, H, H, entrance] into [I, E, H, G, B]. Then, the calculation unit 104 converts the transition data [I, E, H, G, B] into transition data [IEHGB], which is a character string.

[0060] Next, the calculation unit 104 calculates the edit distance between the character strings represented by the plurality of transition data, and calculates the route dispersion of the customer based on the edit distance.

[0061] Specifically, the calculation unit 104 reads all or part of the flow line data obtained for each customer during past store visits from the data storage unit 100. Next, the calculation unit 104 generates transition data from the flow line data for each customer. The calculation unit 104 then calculates the distance between the transition data for each customer. Note that the shorter the distance between the transition data, the higher the similarity between the transition data. The calculation unit 104 then calculates the route variance for that customer by aggregating the distances between the transition data.

[0062] (1) Calculating the distance between customer transition data (Selection of calculation target) All movement data of a certain customer may be used as the target, but it is generally believed that a customer's purchasing trends change depending on the time of day. For example, it is believed that the purchase contents will be significantly different even for the same customer between purchases made in the morning (e.g., single purchases of coffee, etc.) and purchases made in the afternoon (e.g., purchases of lunch, etc.). Naturally, if the purchase contents differ, the movement data within the store will also change. For this reason, it is better to divide the targets to be treated in a way that separates customer purchasing trends to some extent, such as by time of day. For example, if limited to the afternoon time period, it is possible to target only the movement lines that entered the store between 11:00 and 1:00 p.m.

[0063] FIG. 10 is a diagram illustrating an example in which the flow-line data to be calculated is limited to daytime hours. In the example of FIG. 10, customer A is a customer who visits the store multiple times a day. In this case, as described above, it is assumed that customer A purchases different products during daytime hours than during other time periods. For this reason, for example, the flow-line data to be calculated for distance is limited to daytime hours. Note that when calculating distance, the distance may be calculated for all data from June 22 to July 10, for example. Alternatively, the distance may be calculated between adjacent data, for example. For example, the distance may be calculated between the data from June 22 and June 23, and the distance between the data from June 23 and June 24, and the average of these distances may be calculated.

[0064] (Calculating distance between flow paths) First, the edit distance between the transition data of the same customer selected by the above method is calculated. Specifically, the calculation unit 104 inputs two pieces of transition data and measures the edit distance between the two character strings. For example, the edit distance between the transition data "ABA" and the transition data "ABCBA" is equal to the cost of inserting area C into the route. Note that the cost in this case may be calculated using Euclidean distance or the inverse of the transition probability between areas.

[0065] (Normalization) Next, the calculation unit 104 normalizes the distance calculated by the above-described method by dividing it by the maximum length of the character string of the transition data. For example, when calculating the distance between the transition data "ABCDA" and the transition data "ABA", the distance is divided by the maximum length of 5. This makes the range of values ​​within the range of 0 to 1, making it easier to handle and reducing the impact of the flow line length. Note that the distance may be handled as is without being normalized.

[0066] (Combination of Traffic Lines) Combinations of traffic lines can be made by using all of the traffic line data of the same selected customer, or by combining only traffic line data that are consecutive in time series, but either can be used.

[0067] (2) Calculation of Aggregation and Route Variance Then, the calculation unit 104 aggregates the multiple distances obtained by calculating the distance between transition data of a certain customer. Basically, the average of these distances is used, but a weighted average may also be used depending on the importance of the visit. The average value of the distance for each customer calculated by the above method is treated as the route variance.

[0068] For example, as described above, the edit distance between the transition data [IEHGB] and the transition data [HGIEDBAJKML] generated from the flow line data of a certain customer is 9. In this case, by dividing the edit distance of 9 by the maximum length of these two data, 11, the distance between the transition data [IEHGB] and the transition data [HGIEDBAJKML] is calculated as 0.818.

[0069] The calculation unit 104 calculates the normalized distance for each pair of the flow line data for each customer, as described above, and calculates the sum of the normalized distances for each pair.The calculation unit 104 then calculates the route variance for each customer by dividing the sum of the normalized distances by the number of visits to the store.

[0070] The formula for calculating route variance is shown in the following formula (1). In the following, k represents the character string of the transition data converted from the flow line data, i and j are numbers for identifying the store visit or flow line data, and n represents the number of customer visits or the number of flow line data. distance() represents the edit distance, and len() represents the length of the character string.

[0071] (1)

[0072] The determination unit 106 determines whether the customer has a planned purchasing tendency or an unplanned purchasing tendency based on the path variance calculated by the calculation unit 104. For example, if the path variance calculated by the calculation unit 104 is equal to or greater than a preset threshold, the determination unit 106 determines that the customer has an unplanned purchasing tendency. On the other hand, if the path variance calculated by the calculation unit 104 is less than the preset threshold, the determination unit 106 determines that the customer has a planned purchasing tendency. In this case, it is also determined whether the customer is a variety-seeking customer.

[0073] The output unit 108 outputs the characteristics of each customer obtained by the determination unit 106 as a result.

[0074] 11 is a flowchart showing an example of the flow of processing by the information processing program according to this embodiment. The processing by the information processing program is realized by the CPU 11 of the information processing device 10 writing a learning program stored in the ROM 12 or the storage 14 to the RAM 13 and executing it.

[0075] When the information processing device 10 receives a predetermined instruction signal, it executes the information processing of FIG.

[0076] First, in step S100 of FIG. 11, the CPU 11 functions as the acquisition unit 102 and reads out a plurality of pieces of flow line data for each customer from the data storage unit 100.

[0077] In step S101, the CPU 11 sets one customer as the calculation unit 104.

[0078] In step S102, the CPU 11 functions as the calculation unit 104 to convert each of the plurality of customer flow line data set in step S101 into transition data.

[0079] In step S104, the CPU 11 functions as the calculation unit 104 to calculate the edit distance between the plurality of transition data obtained in step S102.

[0080] In step S106, the CPU 11 functions as the determination unit 106 and calculates the sum of the edit distances obtained in step S104 as the route variance.

[0081] In step S108, the CPU 11, functioning as the determination unit 106, determines whether the customer has an unplanned purchasing tendency or a planned purchasing tendency based on the route dispersion obtained in step S106.

[0082] In step S109, the CPU 11, functioning as the determination unit 106, determines whether or not the processing of steps S101 to S108 has been executed for all customers stored in the data storage unit 100. If the processing of steps S101 to S108 has been executed for all customers, the process proceeds to step S110. On the other hand, if there is a customer for whom the processing of steps S101 to S108 has not been executed, the process returns to step S101.

[0083] In step S110, the CPU 11 outputs the determination result for each customer obtained in step S108 as the output unit 108. This determination result indicates whether the customer has a tendency to unplanned purchases or a tendency to planned purchases.

[0084] As described above, the information processing device of this embodiment acquires multiple pieces of flow line data that represent the in-store flow lines of customers when they visit a store, and that represent the in-store flow lines for each visit to the store.The information processing device calculates the variability of customers' routes within the store based on the multiple pieces of flow line data acquired.Based on the variability of routes, the information processing device determines whether the customer has a tendency to make planned purchases or unplanned purchases.This makes it possible to estimate whether the customer has a tendency to make planned purchases or unplanned purchases.

[0085] Furthermore, by converting the flow line data for each customer into transition data, it is possible to appropriately reduce the amount of data. Although it is possible to calculate route dispersion from the flow line data, the amount of calculation increases. In contrast, by converting the flow line data for each customer into transition data, which is character string data, it is possible to reduce the amount of calculation. Furthermore, by converting the flow line data for each customer into transition data, it is possible to appropriately express the meaning of the customer's movement within the store. For example, while simple flow line data is coordinate data, transition data represents movement between areas within the store. Therefore, by using the transition data, it is possible to interpret the meaning, such as which product area the customer moved from to which product area.

[0086] Furthermore, store managers can consider measures that take into account the purchasing characteristics of each customer. As will be described later, planned purchases, unplanned purchases, and unit purchase prices are thought to be related. In particular, the route distribution for each customer obtained by the information processing device of this embodiment is thought to be closely related to customers with unplanned purchasing tendencies and unit customer prices. Therefore, the following application examples are possible.

[0087] (1) Use as an Index for Discovering Appeal Targets By utilizing the route variance obtained by the information processing device of this embodiment, it is believed possible to discover customers for whom appeal is highly effective. FIG. 12 is a scatter diagram showing the relationship between route variance and customer unit price. The customer unit price is the total price of products purchased by a customer in one shopping trip. Note that the scatter diagram shown in FIG. 12 is a scatter diagram obtained in the example described below. As shown in FIG. 12, the greater the route variance, the higher the customer unit price tends to be. Customers with large route variance are likely to have unplanned purchasing tendencies, and product appeal is likely to be highly effective. Therefore, as shown in FIG. 12, by extracting a customer group with large route variance and providing coupon campaigns or product information notifications to this customer group, it is possible to directly appeal to this customer group with unplanned purchasing tendencies.

[0088] Furthermore, it is also possible to extract the characteristics of a customer group with high route dispersion and implement mass measures that match those characteristics (for example, measures that are implemented for all customers, such as signage displays or discounts on specific products). For example, consider a case where characteristics are extracted from a customer group with high route dispersion, such as that they tend to stay in a specific location in the store, purchase specific products, or have a certain attribute (for example, gender). In this case, it is possible to plan measures such as installing a display device that displays signage in a specific location to promote the customer, stocking new flavors of specific products, or preparing products for a specific attribute (for example, women). In this way, it is also possible to convert a customer group with high route dispersion into customer characteristics and then appeal to customers who have those customer characteristics.

[0089] Considering the above-described application, for example, a system can be implemented that extracts customers with large route dispersion and notifies the store manager of their presence or characteristics. To realize this application example, a system with the function of sorting all customers according to route dispersion based on the route dispersion output from the information processing device described above and extracting customers with the highest route dispersion values ​​can be implemented. FIG. 13 is a diagram illustrating the process of sorting customer IDs in order of route dispersion value. As shown in FIG. 13, all customer IDs are sorted so that, for example, the customer ID with the larger route dispersion value is ranked higher. Furthermore, the flow data of customers with large route dispersion is mapped on an in-store map, and the frequency of visits by each area of ​​the customer is calculated. It is also possible to implement a function that intensively distributes products that customers with large route dispersion tend to prefer (e.g., new products) to signage near areas where customers with large route dispersion tend to visit. FIG. 14 is a diagram illustrating areas where customers with large route dispersion tend to visit. The multiple curves shown in FIG. 14 represent the flow data of the top customers shown in FIG. 14. For example, the multiple curves shown in Figure 14 represent customer traffic data for customer IDs "c_00001," "c_00002," and "c_00003" in Figure 13. Customers with customer IDs "c_00001," "c_00002," and "c_00003" have a large degree of route dispersion. For this reason, for example, by concentrating the delivery of products likely to be preferred by customers with customer IDs "c_00001," "c_00002," and "c_00003" to signage near area H in Figure 14, where the pass-through rate of customers with IDs "c_00001," "c_00002," and "c_00003" is high, it is possible to enhance the appealing effect.

[0090] (2) Use as a Purchase Price Index Furthermore, measures aimed at increasing the average customer price may be implemented by utilizing the correlation between route variance and average customer price. As described below, route variance is considered to have a high correlation with average customer price. However, it is considered that route variance and average customer price do not necessarily have a linear correlation. Depending on the store layout or size and the route variance calculation method used in the information processing device, the relationship between route variance and average customer price may be strongly correlated with a linear function or a strong exponential function. For this reason, the scatter diagram showing the relationship between route variance and average customer price may be transformed to strongly express the correlation by performing a logarithmic transformation or the like on each plot point. Figure 15 is a diagram illustrating the correlation between route variance and average customer price. As shown in the upper diagram of Figure 15, the correlation between route variance and average customer price is not necessarily linear (solid line in Figure 15). For this reason, for example, the correlation between route dispersion and average customer price may be observed by logarithmically transforming each plot point in the scatter diagram of Fig. 15. When logarithmic transformation is performed, it is expected that a relationship such as that shown by the dashed line in Fig. 15 will be observed.

[0091] Furthermore, it is also possible to extract a customer group from the entire customer population that exhibits a strong correlation between route variance and average customer price. For example, as shown in the lower diagram of Figure 15, the customer group within the ellipse has a particularly strong correlation between route variance and average customer price. Therefore, by implementing measures to increase route variance for such a customer group, it is possible to increase average customer price.

[0092] In the above-described application scenario, a scatter diagram showing the relationship between route variance and average customer spending can be displayed to the store manager, and the strength of the correlation between route variance and average customer spending and the areas where the correlation is strong can be presented. In this case, for example, customers with small route variance and low average customer spending can be extracted, and customers with large route variance and high average customer spending can be extracted. A system or the like can then be implemented that presents the differences in the characteristics of both to the store manager. To realize this application example, an information processing device can be realized that has the function of plotting a scatter diagram plotting route variance and average customer spending for all customers based on the route variance output from the information processing device described above. Furthermore, a function can be implemented that extracts a customer group with small route variance and a customer group with large route variance and displays the difference in customer characteristics between the two groups. FIG. 16 is a diagram illustrating a comparison between a customer group with small route variance and a customer group with large route variance. As shown in Figure 16, by comparing indicators such as the category of products purchased, the time of day when customers enter the store, and the rate of stay in specific areas between customer groups with low route variance and customer groups with high route variance, it is possible to use this information to help plan measures within the store.

[0093] The information processing device of this embodiment may have the following modifications: Modifications 2 to 4 are modifications in which the above-described application examples are implemented as an information processing device.

[0094] [Variation 1] In the above embodiment, transition data is used to calculate route dispersion, but this is not limiting. For example, route dispersion may be calculated using the above-described flow line data. In this case, for example, the distance between the coordinates of each pair of flow lines among the multiple flow line data may be calculated using DTW (dynamic time warping) or the like instead of edit distance.

[0095] [Variation 2] In the above embodiment, the flow line data is continuous coordinate data, and the transition data is generated from the flow line data. However, the present invention is not limited to this. For example, if the data stored in the data storage unit 100 is originally transition data, the process of generating transition data from the flow line data is unnecessary. In this case, the transition data can be converted into a character string and then passed to the calculation unit 104.

[0096] [Variation 3] As in the above-described example, the calculation unit 104 of the information processing device 10 sorts all customer IDs so that customer IDs with larger route variances are ranked higher. Furthermore, the calculation unit 104 of the information processing device 10 maps the traffic data of customers with large route variances on a store map and calculates the frequency of visits by each area for these customers. The output unit 108 of the information processing device 10 then intensively distributes products likely to be preferred by these customers (e.g., new products) to signage near areas where customers with large route variances are likely to visit. More specifically, when the determination unit 106 determines that a customer has an unplanned purchasing tendency, the output unit 108 of the information processing device 10 may display product-related information (e.g., information introducing the product or coupon information for the product) available in the store on a display device (not shown) located in an area where the customer's frequency of visits during past visits to the store is equal to or greater than a predetermined threshold. In this case, when a customer visits the store, the information processing device 10 acquires the customer ID in some way (for example, by identifying the customer via a terminal carried by the customer or by capturing an image of the customer with a store camera), and outputs the above-described product-related information based on the route distribution linked to the customer ID (or the determination result indicating whether the customer has a tendency to make unplanned purchases or a tendency to make planned purchases).

[0097] [Variation 4] As in the above-described example of use, the output unit 108 of the information processing device 10 may output a scatter diagram in which, for each of a plurality of customers, a combination of a value representing the route dispersion of that customer and a value representing the unit purchase price of that customer is plotted as a point.

[0098] [Variant 5] Furthermore, the calculation unit 104 of the information processing device 10 may calculate the route variance for each customer for each specified time period, and the determination unit 106 may determine whether or not a customer has a tendency to make unplanned purchases when the customer visits the store based on the route variance corresponding to the time period in which the customer visits the store.

[0099] Next, the relationship between the route dispersion obtained by the information processing device of this embodiment and various values ​​will be described. Figure 17 is a scatter diagram showing the relationship between the route dispersion and various values ​​obtained by the experiment of this embodiment. Table 1 below is a table showing a list of correlation coefficients between each value.

[0100]

[0101] Referring to Figure 17 and Table 1, it can be seen that the correlation coefficient between route variance and average customer price is the largest. It is generally said that extending the length of a customer's flow line increases the average purchase price. However, in the experimental results of this example, the correlation coefficient between route variance and average customer price was greater than the correlation coefficient between flow line length and average customer price. Furthermore, with regard to the number of products purchased, flow line length and dwell time were dominant. This also suggests that customers with large route variance tend to purchase high-priced products.

[0102] In addition, the behavior of four customers was analyzed in the above experiment. Customer C1 had a route variance value of 0.717, Customer C2 had a route variance value of 0.685, Customer C3 had a route variance value of 0.399, and Customer C4 had a route variance value of 0.475. Customers C1 and C2 were a group of customers with high route variance, while Customers C3 and C4 were a group of customers with low route variance. Analysis of the behavior of these customers revealed that Customers C1 and C2 had variations in the products and sets of products they purchased. On the other hand, Customer C3 frequently purchased specific products individually. Furthermore, Customer C4 generally purchased sets of two loaves of bread. These analysis results also suggest that customers with high route variance tend to have higher average customer spending.

[0103] In the above embodiment, the processes that the CPU 11 reads and executes each program may be executed by various processors other than the CPU 11. Examples of processors in this case include dedicated electrical circuits, such as programmable logic devices (PLDs) (such as field-programmable gate arrays (FPGAs)) whose circuit configuration can be changed after manufacture, and application-specific integrated circuits (ASICs) that are processors having a circuit configuration specifically designed to execute specific processes. Each process may be executed by one of these various processors, or by a combination of two or more processors of the same or different types (e.g., multiple FPGAs, or a combination of a CPU and an FPGA). The hardware structure of these various processors is, more specifically, an electrical circuit that combines circuit elements such as semiconductor elements.

[0104] In the above embodiment, the program is pre-stored (also referred to as "installed") in the ROM 12 or the storage 14, but this is not limiting. The customer number prediction program may be provided in a form stored on a non-transitory storage medium such as a CD-ROM (Compact Disk Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), or a USB (Universal Serial Bus) memory. The program may also be downloaded from an external device via a network.

[0105] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[0106] The following additional notes are provided regarding the above-described embodiments.

[0107] (Supplementary Item 1) An information processing device comprising: a memory; and at least one processor connected to the memory, wherein the processor is configured to: acquire a plurality of pieces of flow line data representing the in-store flow lines of customers when they visit a store and the in-store flow lines for each visit to the store; calculate the variability of the customer's route within the store based on the acquired plurality of pieces of flow line data; and determine whether the customer has a tendency to make planned purchases or a tendency to make unplanned purchases based on the calculated variability of the route.

[0108] (Addendum 2) A non-transitory storage medium storing a program executable by a computer to perform information processing, wherein the information processing comprises: acquiring a plurality of pieces of flow line data representing the in-store flow lines of customers when they visit a store and representing the in-store flow lines for each visit to the store; calculating the variability of the customer's routes within the store based on the acquired plurality of pieces of flow line data; and determining whether the customer has a tendency to make planned purchases or a tendency to make unplanned purchases based on the calculated variability of the routes.

[0109] 10 Information processing device 100 Data storage unit 102 Acquisition unit 104 Calculation unit 106 Determination unit 108 Output unit

Claims

1. An information processing device comprising: an acquisition unit that acquires multiple pieces of flow line data representing in-store flow lines of customers when they visit a store, the in-store flow lines representing the in-store flow lines for each visit to the store; a calculation unit that calculates a variance in the customer's route within the store based on the multiple pieces of flow line data acquired by the acquisition unit; and a determination unit that determines whether the customer has a tendency to make planned purchases or a tendency to make unplanned purchases based on the variance in the route calculated by the calculation unit.

2. The information processing device according to claim 1, wherein the calculation unit generates transition data representing the transition of the customer between areas within the store for each of the multiple pieces of movement line data acquired by the acquisition unit, and calculates the variance of the route based on the multiple pieces of generated transition data.

3. The information processing device according to claim 2, wherein the area is represented by a single character, the transition data is represented by a character string, and the calculation unit calculates an edit distance between the character strings represented by the multiple transition data, and calculates a variance of the route based on the edit distance.

4. The information processing device of claim 1, further comprising an output unit that, when the judgment unit judges that the customer has an unplanned purchasing tendency, displays product-related information available in the store on a display device located in an area where the customer has visited the store with a frequency equal to or greater than a predetermined threshold value in the past.

5. The information processing device according to claim 1, further comprising an output unit that outputs a scatter diagram in which combinations of values ​​representing the variability of the routes and values ​​representing the unit purchase prices of the customers are plotted as points for each of the multiple customers.

6. The information processing device of claim 1, wherein the calculation unit calculates the variance of the route for each specified time period, and the determination unit determines whether or not the customer has a tendency to make unplanned purchases when the customer visits the store based on the variance of the route corresponding to the time period in which the customer visits the store.

7. An information processing method in which a computer executes the following processes: acquiring a plurality of pieces of flow line data representing in-store flow lines of customers when they visit a store and representing the in-store flow lines for each visit to the store; calculating the variance in the routes taken by the customers within the store based on the acquired plurality of pieces of flow line data; and determining whether the customers have a tendency to make planned purchases or a tendency to make unplanned purchases based on the calculated variance in the routes.

8. An information processing program for causing a computer to function as each part of the information processing device according to any one of claims 1 to 6.

Images