Product recommendation device and product recommendation program

Abstract

We aim to further improve personalized recommendation technology, ultimately leading to increased customer convenience and higher average customer spending. [Solution] A product recommendation device for a regular home delivery service, comprising: a product recommendation means that recommends recommended products to the user, including new products that the user has not purchased before and experienced products that the user has purchased before, based on the user's product purchase history; and a recommendation control means that, among the recommended products, increases the number of experienced products compared to new products the longer the user has been purchasing the product.

Description

Technical Field

[0001] The present invention relates to a product recommendation device and a product recommendation program.

Background Art

[0002] As one of the recommendation techniques in online shopping and EC (electronic commerce) sites, personalized recommendations that analyze a user's past purchase history and browsing history using an algorithm and recommend products that the user is likely to buy for the purpose of improving customer convenience and customer unit price are known (see, for example, Patent Documents 1 and 2).

[0003] Non-Patent Document 1 also describes a recommendation engine having a plurality of respective features. For example, in the case of an EC site for a regular home delivery service of organic vegetables, new products are recommended by a novelty engine according to the customer's preferences, and then re-purchase is recommended by a repeatability engine. If the customer purchases at this timing, the purchase interval is grasped, and the periodicity engine recommends at a timing considered optimal for each customer, leading to regular product purchases.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Non-Patent Documents

[0005]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] Personalized recommendations are a technology that predicts the demand of each individual customer, asking "who is likely to buy what and when," and recommending products at the most effective time for that customer, thereby aiming to increase the average transaction value per customer.

[0007] However, on e-commerce sites, the area where recommended products are displayed is not unlimited and is limited by display space. Therefore, it is difficult to arbitrarily increase the number of recommended products displayed simply because they are recommended.

[0008] This invention was proposed in view of the above points, and in one aspect aims to further improve personalized recommendation technology and increase the average customer spending. [Means for solving the problem]

[0009] To solve the above problems, the product recommendation device according to the present invention is a product recommendation device for a regular home delivery service, and includes a product recommendation means that recommends to the user recommended products, including new products that the user has not purchased before and experienced products that the user has purchased before, based on the user's product purchase history, and a recommendation control means that increases the number of experienced products compared to new products among the recommended products as the user's continuous purchase period for products increases. [Effects of the Invention]

[0010] According to embodiments of the present invention, in one aspect, it is possible to further improve personalized recommendation technology and increase the average customer spending. [Brief explanation of the drawing]

[0011] [Figure 1] This figure shows an example configuration of a regular home delivery product recommendation system according to this embodiment. [Figure 2] This figure shows an example of the functional configuration of the recommendation machine according to this embodiment. [Figure 3] It is a diagram (part 1) for explaining the recommendation engine according to this embodiment. [Figure 4] It is a diagram (part 2) for explaining the recommendation engine according to this embodiment. [Figure 5] It is a diagram (part 3) for explaining the recommendation engine according to this embodiment. [Figure 6] It is a diagram (part 1) for explaining the relationship between the purchase continuation period of home-delivery products by users according to this embodiment and the number of purchased new products and experience products. [Figure 7] It is a diagram (part 2) for explaining the relationship between the purchase continuation period of home-delivery products by users according to this embodiment and the number of purchased new products and experience products. [Figure 8A] It is a diagram for explaining EC site screen 1 according to this embodiment. [Figure 8B] It is a diagram for explaining EC site screen 2 according to this embodiment. [Figure 9A] It is a diagram for explaining EC site screen 3 according to this embodiment. [Figure 9B] It is a diagram for explaining EC site screen 4 according to this embodiment. [Figure 10] It is a flowchart showing the recommendation product display control process according to this embodiment. [Figure 11] It is a diagram showing a configuration example of the regular home-delivery product recommendation system according to this modification example. [Figure 12] It is a diagram showing a functional configuration example of the recommendation machine according to this modification example. [Figure 13] It is a diagram showing an example of a personal flier according to this modification example.

Mode for Carrying Out the Invention

[0012] Embodiments of the present invention will be described in detail with reference to the drawings. <System Configuration> FIG. 1 is a diagram showing a configuration example of a regular home delivery product recommendation system according to the present embodiment. The regular home delivery product recommendation system 100 in FIG. 1 includes a home delivery online shopping server 10, a recommendation machine 20, various DBs (databases) 30, and a user terminal 40, and is connected via a network 50.

[0013] The home delivery online shopping server 10 is a web server that provides an EC site of an operator who operates a regular home delivery service for, for example, foodstuffs and daily necessities. Users (members) of the regular home delivery service access and log in to the EC site using the user terminal 40, and select and order home delivery products. On the EC site, home delivery products are categorized and displayed in a predetermined product display column, and a recommended product display area for displaying recommended products (recommended items) determined by the recommendation machine 20 is provided.

[0014] The recommendation machine 20 has a plurality of recommendation engines described later, and is a product recommendation device that analyzes and calculates the past purchase history (purchase history) for each user, and outputs products that the user is likely to buy. The output recommended products are displayed in the recommended product display area for the user on the EC site.

[0015] The user terminal 40 is, for example, a PC (personal computer), a smartphone, a tablet terminal, etc., and is a terminal of a user who uses the regular home delivery service. A predetermined application program, a general-purpose web browser, etc. for accessing and logging in to the EC site of the home delivery online shopping server 10 are installed in advance in the user terminal 40. The user accesses and logs in to the EC site using the user terminal 40, and selects and orders (purchases) desired products. The ordered products are delivered to the delivery destination such as the user's home according to the delivery schedule. Generally, in the case of a regular home delivery service, an orderable period is determined, orders are placed within a fixed period every week, and the products are delivered on a fixed day of the week after the order deadline.

[0016] (Functional configuration) Figure 2 shows an example of the functional configuration of the recommendation machine according to this embodiment. The recommendation machine 20 according to this embodiment has, as its main functional units, a product recommendation unit 201, a recommendation control unit 202, and a storage unit 206.

[0017] The product recommendation unit 201 has the function of recommending products to the user, including new products that the user has not purchased before and products that the user has purchased before, based on the user's product purchase history.

[0018] The recommendation control unit 202 has a function that, among the recommended products recommended to the user, increases the number of experienced products rather than new products, as the user's continued purchase period for those products is longer.

[0019] The recommendation control unit 202 includes a purchase continuation period acquisition unit 203, a display ratio calculation unit 204, and a display control unit 205. The purchase continuation period acquisition unit 203 has the function of acquiring information on the user's purchase continuation period for products. The display ratio calculation unit 204 has the function of calculating the display ratio of new products and experienced products among the recommended products displayed to the user in the recommended product display area, according to the purchase continuation period. The display control unit 205 displays recommended products (new products and experienced products) in the recommended product display area on the e-commerce site according to the display ratio.

[0020] The memory unit 206 has the function of storing each recommendation priority list calculated for each recommendation engine in a memory device.

[0021] The recommendation machine 20 can be implemented using a general-purpose computer. Specifically, the recommendation machine 20 includes hardware such as an arithmetic processing unit (CPU, etc.), memory, input / output interfaces, and communication interfaces. The functions of the recommendation machine 20 are realized by the arithmetic processing unit executing processing according to computer programs stored in memory. That is, each functional unit is realized by a computer program executed on the hardware resources such as the arithmetic processing unit and memory of the computer that constitutes the recommendation machine 20. These functional units may also be read as "means," "modules," "units," or "circuits." Furthermore, some of the functional units may be located in the memory of the recommendation machine 20 or in external storage devices on the network. In addition, each functional unit of the recommendation machine 20 may be implemented not only by a single server device, but also as a system consisting of multiple devices with distributed functions. Furthermore, the above computer programs may be stored in a storage medium that can be read by the computer.

[0022] (Database) The DB30 according to this embodiment includes a user DB, a product DB, a product purchase history DB, and order delivery schedule information.

[0023] The user database is a database where user information of members who use the regular home delivery service is registered. For example, it includes user information such as the user's name, age, address, family, preferences, membership start date, and membership duration, as well as the login ID and login password for the e-commerce site on the home delivery online shopping server.

[0024] The product database is a database (product master) where products sold as home delivery items are registered. Product information includes, for example, product code, product name, JAN code, product category, price, quantity, stock quantity, and producer. Recommended products are selected from the products registered in the product database.

[0025] The product purchase history database is a database that records the purchase history of products that a user has purchased in the past. It includes at least product purchase history information for each user, such as the date and time of purchase, product code, product name, purchase price, quantity purchased, and number of purchases.

[0026] Order and delivery schedule information refers to the period during which products can be ordered and the delivery date. Generally, in the case of regular delivery services, there is a set period during which orders can be placed, and products are delivered on a fixed day of the week. For example, in the case of a weekly one-term regular delivery service, users can place orders during the order period, for example, from 12:00 AM on Tuesday (order start time) to 12:00 PM on Sunday (order deadline time). Orders for the week are closed at the order deadline time, and those orders are delivered, for example, on Wednesday of the following week. Note that order and delivery schedule information may be set individually for each region based on the user's address. It may also be a two-term or more-term system per week.

[0027] (Recommendation engine) Figure 3 is a diagram (part 1) illustrating the recommendation engine according to this embodiment. As shown in Figure 3, the recommendation machine 20 according to this embodiment has a recommendation engine (recommendation engine program) that has multiple different algorithms, including, for example, a novelty engine, a repetition engine, a periodicity engine, and a preference engine.

[0028] The novelty engine is a recommendation engine designed to encourage first-time purchases. It has the function of recommending new products, such as new products or promotional items, that the user has not previously purchased, thereby contributing to the expansion of their purchase range.

[0029] The repetition engine is a recommendation engine designed to encourage repeat purchases. It recommends products that the user has previously purchased, contributing to increased purchase frequency.

[0030] The periodicity engine is a recommendation engine designed to encourage purchases three or more times. It recommends products that a user has previously purchased two or more times, contributing to increased purchase frequency. Based on the purchase history of products repeatedly bought in the past, the periodicity engine identifies the purchase cycle (purchase interval) of those products and recommends them when the purchase cycle (e.g., a cycle week) arrives, encouraging repeat purchases. Product recommendations by the periodicity engine can also be seen as a function to prevent users from forgetting to buy cyclical products.

[0031] The preference engine is a recommendation engine designed to encourage users to purchase products that match their preferences. In the case of the preference engine, recommended products can include both new products and products the user has previously purchased. For example, based on user information such as preferences, age, and family, and / or preference information based on the user's purchase history, it can recommend matching new products (situations where you want to encourage first-time purchases) and products the user has previously purchased (situations where you want to encourage repeat purchases), either alone or in combination with other novelty engines or periodicity engines.

[0032] Figure 4 is a diagram (part 2) illustrating the recommendation engine according to this embodiment. The products recommended in the novelty engine are new products, while the products recommended in the repetition engine and the periodicity engine are products that have been experienced. The recommendation machine 20 calculates a purchase likelihood score and a priority based on the score for each product from among the multiple or numerous recommended products for each engine, and recommends products with higher priority to be displayed more frequently based on the calculation results shown in Figure 4 (the recommendation priority list for each recommendation engine). The score is calculated comprehensively by combining various parameters, such as user preferences (preference engine), as well as the product's market sales (popularity), time elapsed since release, whether it is a staple product, seasonality, inventory level, and promotion level (a weight value indicating the degree to which the delivery service provider particularly wants to sell the product).

[0033] Figure 5 is a diagram (part 3) illustrating the recommendation engine according to this embodiment. The recommendation processing of the recommendation engine 20 will be explained using the following model case.

[0034] Step S1: The recommendation engine 20 recommends product A, a new product, based on the novelty engine (or preference engine). At this point, the user does not purchase the new product (week n).

[0035] Step S2: The recommendation engine 20 recommends product A again based on the novelty engine (or preference engine). At this time, the user is considered to have purchased the new product (week n+1). Information about the purchased product A, along with the purchase date and time, is recorded in the product purchase history DB as the first purchase history.

[0036] Step S3: The recommendation engine 20 recommends product A from the experienced products (purchased only once in the past) based on the iteration engine. At this time, the user does not purchase the new product (week n+2).

[0037] Step S4: Based on the repetition engine, the recommendation engine 20 recommends product A from the experienced products (purchased only once in the past). At this time, the user is considered to have purchased the new product (week n+3). Information about the purchased product A is recorded in the product purchase history DB as the second purchase, along with the purchase date and time. The recommendation engine 20 also determines that the purchase interval (purchase cycle) for product A for that user is 2 weeks based on the product purchase history DB for the first purchase date and time (week n+1) and the second purchase date and time (week n+3), and stores the information of the user, product A, and purchase cycle in association.

[0038] Step S5: Based on the cycle engine, the recommendation engine 20 recommends product A from the experienced products (purchased at least twice in the past) for which this week (this week) is the purchase cycle (week n+5). Note that in week n+4, the recommendation engine 20 does not recommend product A, which is not the purchase cycle. In week n+4, the recommendation engine 20 should recommend some other product, such as another product, that is the purchase cycle for that week.

[0039] Step S6: Based on the cycle engine, the recommendation engine 20 recommends product A from among the experienced products (purchased at least twice in the past) for which this week (this week) is the purchase cycle (week n+7). In this way, the cycle engine recommends product A at periodic timings when the user is highly likely to purchase product A, thereby leading to regular purchases of product A.

[0040] <Relationship between user purchase retention period and the number of purchased items by new and experienced product categories> Figure 6 is a diagram (part 1) illustrating the relationship between the user's continued purchase period of home-delivered products and the number of new and experienced products purchased according to this embodiment. Figure 6 is a graph based on big data from the user DB and product purchase history DB accumulated in the past, with the user's continued purchase period of home-delivered products (membership period) on the horizontal axis and the number of new and experienced products purchased (average value) on the vertical axis.

[0041] The graph shows that the number of products purchased per term that a member has previously purchased and repurchased tends to increase with longer purchase history, while the number of new products purchased for the first time remains relatively constant (or shows a very slight increase) across all member segments, regardless of purchase history.

[0042] Figure 7 is a diagram (part 2) illustrating the relationship between the purchase duration of home-delivered products by users according to this embodiment and the number of new and experienced products purchased. For example, for user A with a purchase duration of 6 months, the number of new products purchased per term is 3, and the number of experienced products purchased is 2. For user A with a purchase duration of 12 months, the number of new products purchased per term is 3, and the number of experienced products purchased is 3. For user A with a purchase duration of 24 months, the number of new products purchased per term is 3, and the number of experienced products purchased is 6. In other words, while the number of purchased products increases with the purchase duration, looking at the breakdown of new and experienced products, the number of experienced products purchased increases as the purchase duration lengthens (2, 3, 6), while the number of new products purchased remains at 3 regardless of the purchase duration. Experienced products contribute significantly from the perspective of expanding the range of purchased items and improving the average customer spending.

[0043] Therefore, the recommendation machine 20 according to this embodiment controls the recommendation machine to reduce the proportion of new products recommended in the recommended product display area as the user's purchase continuation period (membership continuation period) lengthens, and to increase the proportion of experienced products recommended, as these products contribute more significantly from the perspective of expanding the range of purchased items and increasing the average customer spending. This can also be described as a control that changes the display ratio of new products to experienced products as the user's purchase continuation period lengthens. More specifically, as the user's purchase continuation period lengthens, the display proportion of recommended products (new products) by the novelty engine is reduced, and the display proportion of recommended products (experienced products) by the repetition engine and periodicity engine is increased.

[0044] This allows e-commerce sites offering regular home delivery services to proactively recommend experienced products to customers at more effective times as their purchasing history lengthens, and conversely, to passively recommend new products at less effective times, thereby increasing average customer spending and other benefits.

[0045] As can be seen from Figure 6, in cases where the purchase duration is short, such as less than 1 to 6 months, the proportion of new products purchased by the user (100% of first-time purchases are new products) is large, and the number of experienced products is still small. Therefore, this trend can be excluded from the target purchase duration.

[0046] <Online shopping site for home delivery> The EC site screen according to this embodiment is generated by the home delivery online shopping server 10 and displayed on the screen of the user terminal 40. A detailed explanation follows, with reference to an example of the EC site screen. (EC site screen during the first purchase continuation period) Figure 8A is a diagram illustrating the EC site screen 1 according to this embodiment. The top screen of the EC site for user A shown in Figure 8A includes, for example, the logged-in user 401 as of 10 / 1 / 2024, the current date and time 402, the order and delivery schedule for the current term 403, a content area 404, and a recommended product display area 405. As of 10 / 1 / 2024, user A's purchase continuation period (membership continuation period) is assumed to be 6 months.

[0047] Furthermore, the recommended product display area 405 shown in Figure 8A includes the first recommended product display area 405a, the second recommended product display area 405b, and the third recommended product display area 405c. For example, the first recommended product display area 405a displays new products recommended by the novelty engine, the second recommended product display area 405b also displays new products recommended by the novelty engine, and the third recommended product display area 405c displays new or experienced products recommended by the preference engine.

[0048] Figure 8B is a diagram illustrating the EC site screen 2 according to this embodiment. The order cart screen of the EC site shown in Figure 8B is a screen that displays the items in the user's order cart (shopping basket). The items in the order cart are, for example, five items: "bananas," "1kg of shrimp," "pizza," "white bread," and "eggs." Of these, three items, "bananas," "1kg of shrimp," and "pizza," are new products, while two items, "white bread" and "eggs," are products the user has used before.

[0049] Furthermore, the shopping cart screen includes a fourth recommended product display area 405d, which displays previously recommended products by the repetitiveness engine and the periodicity engine. The periodicity engine, in particular, encourages repeat purchases by recommending products when a period (e.g., a periodic week) arrives. Specifically, on the shopping cart screen, which is the final confirmation screen (the screen just before purchase), it provides one last push to recommend cycle products that the user may have forgotten to buy but haven't yet added to their shopping cart, aiming to further increase purchase frequency and average transaction value.

[0050] Furthermore, from the perspective of improving purchase frequency and unit price, it is desirable that the recommended product display area 405d, which is recommended by the periodic engine, be placed on the shopping cart screen, but it is not necessarily limited to being displayed only on the shopping cart screen. It may also be displayed on the top screen or other screens, taking into consideration the overall balance of recommended product display areas from other engines.

[0051] (EC site screen during the second purchase continuation period) Figure 9A is a diagram illustrating the EC site screen 3 according to this embodiment. The top screen of the EC site for user A shown in Figure 9A includes, for example, the logged-in user 401 as of April 1, 2026, the current date and time 402, the order and delivery schedule for the current term 403, a content area 404, and a recommended product display area 405. As of April 1, 2026, user A's purchase continuation period (membership continuation period) is assumed to be 24 months.

[0052] Furthermore, the first recommended product display area 405a shown in Figure 9A displays new products recommended by the novelty engine, the second recommended product display area 405b displays experienced products recommended by the repetition engine and the periodicity engine, and the third recommended product display area 405c displays new or experienced products recommended by the preference engine.

[0053] Comparing Figure 8A and Figure 9A, Figure 8A shows the second recommended product display area 405b at the point when user A's purchase duration was 6 months, displaying new products recommended by the novelty engine, whereas Figure 9A shows the second recommended product display area 405b at the point when user A's purchase duration was 24 months, displaying experienced products recommended by the repetition engine and the periodicity engine. In other words, in the recommended product display area 405, as the user's purchase duration increases from 6 months to 24 months, the number of new products displayed as recommended products decreases, while the number of experienced products increases accordingly.

[0054] Figure 9B is a diagram illustrating the EC site screen 4 according to this embodiment. The order cart screen of the EC site shown in Figure 9B is a screen that displays the items in the user's order cart. The items in the order cart are, for example, nine items: "apples," "200g beef steak," "bone-in meat," "tomatoes," "eggs," "mayonnaise," "bananas," "1kg shrimp," and "bread." Of these, three items, "apples," "200g beef steak," and "bone-in meat," are new products, while six items, "tomatoes," "eggs," "mayonnaise," "bananas," "1kg shrimp," and "bread," are products that the user has used before.

[0055] As shown in Figure 9B 4th Recommended product display area 405d continues to display products recommended by the repetitive and periodic engines.

[0056] Comparing Figure 8B and Figure 9B, according to the shopping cart in Figure 8B at 6 months of continuous purchase by user A, 3 items are new products and 2 items are experienced products. In contrast, according to the shopping cart in Figure 9B at 24 months of continuous purchase by user A, 3 items are new products and 6 items are experienced products. In other words, the number of experienced products purchased increases from 2 to 6 as the continuous purchase period lengthens, while the number of new products purchased remains the same at 3 regardless of the continuous purchase period.

[0057] As mentioned above, the number of experienced products purchased per term tends to increase for members with longer purchase history, while the number of new products purchased for the first time remains almost constant (or shows a very slight increase) across all member segments, regardless of purchase history. Therefore, in this embodiment, as a user's purchase history lengthens, the proportion of new products displayed in the recommended product display area 405 is reduced, and instead, the proportion of experienced products, which contribute more to expanding the range of purchased items and increasing the average customer spending, is increased (the display ratio is changed so that the proportion of experienced products is greater than the proportion of new products). This makes it possible to actively recommend experienced products as recommended products as a user's purchase history lengthens, and conversely, to passively recommend new products, thereby further increasing the average customer spending.

[0058] <Recommended Product Display Control Processing> Figure 10 is a flowchart illustrating the recommended product display control process according to this embodiment. This flowchart is executed when, for example, a user accesses and logs into an e-commerce site, and the delivery online shopping server 10 receives a request from the user terminal 40 to display the e-commerce site screen, and the recommendation machine 20 receives a request from the delivery online shopping server 10 to acquire recommended product information. Furthermore, the following steps (hereinafter referred to as "S") can be realized by having the arithmetic processing unit of the recommendation machine 20 read and execute a program capable of performing the processing.

[0059] S11: The recommendation machine 20 obtains information on the purchase duration of home delivery products for the logged-in user. The purchase duration can be obtained from the user database as the membership duration, with the membership registration date as the starting date. Alternatively, the purchase duration may be calculated from the current date and time, using the date and time of the first purchase in the product purchase history database as the starting date. If the user temporarily withdrew from the service, the period of temporary withdrawal may be excluded.

[0060] S12: The recommendation machine 20 determines whether the acquired user's purchase continuation period is longer than a predetermined period. If the purchase continuation period is short, the proportion of new products purchased by the user (100% new products in the first purchase) is large and the number of experienced products is small. The predetermined period is the period before the trend appears where the number of experienced products purchased increases and the number of new products purchased levels off as the user's purchase continuation period lengthens (Figure 6), and specifically, it may be set to less than 1 to 6 months.

[0061] S13: If the acquired user's purchase duration is longer than a predetermined period, the recommendation machine 20 calculates the display ratio (or display rate) of new products and experienced products to be displayed in the recommended product display area at that point in time, based on the user's purchase duration. In other words, if the user's purchase duration is long, the display ratio of recommended products is calculated so that there are more experienced products than new products compared to when the user's purchase duration is short.

[0062] (Calculation Example 1) For example, according to Figure 6, a method for calculating the display ratio of new products and experienced products for each predetermined category of purchase continuation period (number of months). For purchase duration (in months) between 6 and 12 months, the ratio of new products to experienced products is 3 / 5 and 2 / 5 (new products:experienced products = 3:2). For purchase duration (in months) between 12 and 18 months, the ratio of new products to experienced products is 3 / 6 (new products:experienced products = 3:3). For purchase duration (in months) between 18 and less than 24 months, the ratio of new products to experienced products is 3 / 8 (new products:experienced products = 3:5). For purchase duration (in months) between 24 and less than 36 months, the number of new products is 3 / 9 and the number of experienced products is 6 / 9 (new products:experienced products = 3:6).

[0063] (Calculation Example 2) For example, according to Figure 6, a method for calculating the display ratio of new products and experienced products for each purchase period (number of months) If the purchase duration (in months) is 6 months, the number of new products is 3 / 5 and the number of experienced products is 2 / 5 (number of new products:number of experienced products = 3:2). If the purchase duration (in months) is 7 months, the number of new products is 3 / 5.2, and the number of experienced products is 2.2 / 5.2 (number of new products:number of experienced products = 3:2.2). If the purchase duration (in months) is 8 months, the ratio of new products is 3 / 5.4, and the ratio of experienced products is 2.4 / 5.4 (new products:experienced products = 3:2.4). If the purchase duration (in months) is 9 months, the ratio of new products is 3 / 5.6, and the ratio of existing products is 2.6 / 5.6 (new products:existing products = 3:2.6). If the purchase duration (in months) is 10 months, the ratio of new products is 3 / 5.8, and the ratio of experienced products is 2.8 / 5.8 (new products:experienced products = 3:2.8). If the purchase duration (in months) is 11 months, the number of new products is 3 / 6 and the number of experienced products is 3 / 6 (number of new products:number of experienced products = 3:3). If the purchase duration (in months) is 12 months, the ratio of new products is 3 / 6.2, and the ratio of experienced products is 3.2 / 6.2 (new products:experienced products = 3:3.2). If the purchase duration (in months) is 13 months, the ratio of new products is 3 / 6.4, and the ratio of experienced products is 3.4 / 6.4 (new products:experienced products = 3:3.4). : Needless to say, the above-mentioned calculation examples 1 and 2 are illustrative examples for ease of understanding, and in practice, the calculation may be performed dynamically using a given formula.

[0064] S14: The recommendation machine 20 retrieves information on recommended products to be displayed in the recommended product display area according to the display ratio calculated in S13, from the recommended product calculation results (for example, each recommendation priority list in Figure 4), and sends it to the home delivery online shopping server 10.

[0065] The recommendation machine 20, taking into account the number of product slots available in the recommended product display area, for example, if the calculated display ratio is 3 / 5 for new products and 2 / 5 for experienced products, it will acquire information on recommended products in order of priority, with a ratio of 3 from the recommended products on the recommendation priority list by the novelty engine and 2 from the recommended products on the recommendation priority list by the periodicity engine. Also, for example, if the calculated display ratio is 3 / 8 for new products and 5 / 8 for experienced products, it will acquire information on recommended products in order of priority, with a ratio of 3 from the recommended products on the recommendation priority list by the novelty engine and 5 from the recommended products on the recommendation priority list by the periodicity engine (or iterative engine).

[0066] Recommended product information includes, for example, the product ID, product name, product image, product price, and display position (including display priority) that indicates which recommended product display area the product should be displayed in.

[0067] S15: If the user's purchase duration is not longer than a predetermined period, the recommendation machine 20 retrieves information on recommended products to display in the recommended product display area according to a predetermined display ratio or predetermined number of items, from the recommended product calculation results (for example, each recommendation priority list in Figure 4), and sends it to the home delivery online shopping server 10. The predetermined display ratio or predetermined number of items is, for example, an initial value when the purchase duration is not taken into consideration (Figure 8A).

[0068] <Summary> As described above, the recommendation machine 20 according to this embodiment, on an e-commerce site for home delivery shopping, reduces the proportion of new products displayed as recommended products in the recommended product display area 405 for users with a longer purchase history, and increases the proportion of experienced products displayed, which contribute more to expanding the range of purchased items and increasing the average customer spending (the display ratio is changed so that the number of experienced products is displayed in greater proportion than the number of new products). As a result, the recommendation machine 20 actively recommends experienced products as recommended products to users with a longer purchase history, and conversely, passively recommends new products, thereby making it possible to further increase the average customer spending and other benefits.

[0069] In e-commerce sites offering regular home delivery of groceries and daily necessities, as the length of a customer's subscription period increases, the number of products they have purchased daily (the number of products in the customer's past purchase history recorded in the product purchase history database) increases, resulting in a vast and diverse collection of accumulated products. On the other hand, while many attractive new products are developed daily, the number of purchases remains relatively constant (users don't immediately jump on new products just because they are recommended and catch their eye). Customers who make a significant contribution to a company's sales, such as those with high purchase frequency and average purchase value, are called valuable customers (loyal customers). In operating e-commerce sites offering regular home delivery of groceries and daily necessities, it goes without saying that it is important to propose attractive new products daily, but to increase customer loyalty, it is even more crucial to implement effective measures with the goal of increasing the number of products that customers continue to purchase over the long term.

[0070] The following points will also be mentioned. In fact, a comparison of A / B test results using an actual e-commerce site showed that for users with longer purchase retention periods (member retention periods), reducing the proportion of new products displayed as recommended products in the recommended product display area 405 and increasing the proportion of experienced products resulted in a noticeable increase in the average customer spending, particularly in the experienced products section, compared to when no such display control was implemented.

[0071] The recommendation machine 20 can perform the calculation process for recommended products (for example, the process of creating and updating the priority recommendation list in Figure 4) at times such as when a user accesses and logs into the e-commerce site, or when the delivery online shopping server 10 receives a request from the user terminal 40 to display the e-commerce site screen.

[0072] However, the timing of the recommended product calculation process is not limited to these times. For example, from the perspective of reducing the load cost associated with increasing the calculation frequency, the calculation may occur before the timing when the user accesses and logs into the EC site, or before the timing when the delivery online shopping server 10 receives a request from the user terminal 40 to display the EC site screen. Alternatively, the timing may be hourly, daily, weekly, or term-based. Even if the calculation is performed before the timing when the user accesses and logs into the EC site, or before the timing when the delivery online shopping server 10 receives a request from the user terminal 40 to display the EC site screen, if there are calculation results (for example, the recommendation priority list in Figure 4) available at that time, the recommendation machine 20 can obtain information on recommended products in order of priority from that recommendation priority list.

[0073] <Variation> In this modified example, the regular home delivery product recommendation system according to the above embodiment is applied to printing on paper flyers (referred to as personal flyers). Figure 11 shows an example of the configuration of a regular home delivery product recommendation system according to this modified example. The regular home delivery product recommendation system 100-2 in Figure 11 is composed of a recommendation machine 20-2, various DBs 30, and a printing device 60.

[0074] The recommendation machine 20-2 has the multiple recommendation engines described above and is a product recommendation device that analyzes and calculates each user's past purchase history to output products that the user is likely to buy. The outputted recommended products are displayed in the recommended product display area for that user on the personal flyer.

[0075] The printing device 60 is a printer for printing personal flyers on flyer paper. The personal flyers are delivered to the user's home or other designated delivery location according to a delivery schedule. The personal flyers include a product catalog and an order form (such as an OCR form), and the user can order their desired products by filling in the product numbers and other information listed in the product catalog and personal flyer and returning them. Alternatively, the user may order by providing the product numbers and other information to an operator over the phone.

[0076] Figure 12 shows an example of the functional configuration of a recommendation machine according to this modified example. The recommendation machine 20-2 according to this modified example has a product recommendation unit 201, a recommendation control unit 202, and a storage unit 206 as its main functional units.

[0077] The product recommendation unit 201 has the function of recommending products to the user, including new products that the user has not purchased before and products that the user has purchased before, based on the user's product purchase history.

[0078] The recommendation control unit 202 has a function that, among the recommended products recommended to the user, increases the number of experienced products rather than new products, as the user's continued purchase period for those products is longer.

[0079] The recommendation control unit 202 includes a purchase continuation period acquisition unit 203, a display ratio calculation unit 204, and a print control unit 205-2. The purchase continuation period acquisition unit 203 has the function of acquiring information on the user's purchase continuation period for products. The display ratio calculation unit 204 has the function of calculating the display ratio of new products and experienced products among the recommended products to be printed and displayed to the user in the recommended product display area on the personal flyer, according to the purchase continuation period. The print control unit 205-2 transmits print data to the printing device 60 and prints information on recommended products (new products and experienced products) in the recommended product display area on the personal flyer according to the display ratio.

[0080] Figure 13 shows an example of a personalized flyer related to this modification. The personalized flyer shown in Figure 12 is a flyer personalized for member user A. The personalized flyer has a recommended product display area, and new products recommended by the novelty engine, new or experienced products recommended by the preference engine, and experienced products recommended by the anti-replication engine / periodicity engine are printed and displayed according to user A's purchase continuation period (membership continuation period).

[0081] As described above, the recommendation machine 20-2 related to this modification, in a personal flyer for a regular home delivery service, reduces the proportion of new products displayed as recommended products in the recommended product display area for users with a longer purchase history, and increases the proportion of experienced products displayed, which contribute more to expanding the range of purchased items and increasing the average customer spending (the display ratio is changed so that the number of experienced products is greater than the number of new products), and is printed on the flyer paper. As a result, the more a user has a long purchase history, the more actively experienced products are recommended as recommended products on the personal flyer, and conversely, new products are recommended less actively, making it possible to further increase the average customer spending. Furthermore, since this modification performs product recommendations on a paper flyer, it is particularly effective for users who have difficulty accessing e-commerce sites.

[0082] While the present invention has been described with specific examples in the form of preferred embodiments, it is clear that various modifications and changes can be made to these examples without departing from the broad spirit and scope of the invention as defined in the claims. In other words, the details of the examples and the accompanying drawings should not be construed as limiting the present invention. [Explanation of Symbols]

[0083] 10 Home Delivery Online Shopping Server 20 Recommendation Machines 20-2 Recommendation Machine 30 DB 40 User terminals 50 Networks 100 Home Delivery Product Recommendation System 100-2 Home Delivery Product Recommendation System 201 Product Recommendation Department 202 Recommendation Control Unit 203 Purchase Continuation Period Acquisition Department 204 Display Ratio Calculation Unit 205 Display Control Unit 205-2 Printing Control Unit 206 Memory section

Claims

1. A product recommendation device for a regular home delivery service, A product recommendation means that recommends products to a user, including new products that the user has not previously purchased and products that the user has previously purchased, based on the user's product purchase history. A recommendation control means that, among the recommended products, increases the number of experienced products rather than new products the longer the user's purchase period for those products, A product recommendation device characterized by having [a certain feature].

2. Computers, A product recommendation means that recommends products to a user, including new products that the user has not previously purchased and products that the user has previously purchased, based on the user's product purchase history. A recommendation control means that, among the recommended products, increases the number of experienced products rather than new products the longer the user's purchase period for those products, A product recommendation program designed to function in this way.

Images