Artificial intelligence-based personalized commodity recommendation system for improving regular delivery resubscrption rate

The AI-based personalized product recommendation system addresses the challenge of maintaining high re-subscription rates in regular delivery services by analyzing customer data to recommend tailored products, improving convenience and meeting changing needs, thereby enhancing customer satisfaction and business benefits.

WO2024248279A1PCT designated stage expired Publication Date: 2024-12-05GIO VENTURES CO LTD

Patent Information

Application Number
PCT/KR2024/002800
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-30
Filing Date
2024-03-05
Publication Date
2024-12-05

AI Technical Summary

Technical Problem

Existing regular delivery services face challenges in maintaining high re-subscription rates due to the difficulty in designing products that meet consumers' preferences and changing needs, as customers often experience inconvenience in purchasing daily necessities in large quantities and face gaps between ordering and receiving products.

Method used

A personalized product recommendation system utilizing artificial intelligence, which collects survey data and product purchase history to recommend customized products and change recommendations based on customers' purchasing tendencies, improving the re-subscription rate by analyzing big data and providing tailored product suggestions.

Benefits of technology

The system effectively enhances re-subscription rates by recommending products that better meet customers' needs, ensuring timely delivery and convenience, while also providing compensation for product sharing and promoting affiliated companies, thus benefiting both customers and sellers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure KR2024002800_05122024_PF_FP_ABST
    Figure KR2024002800_05122024_PF_FP_ABST
Patent Text Reader

Abstract

The present invention relates to an artificial intelligence-based personalized commodity recommendation system for improving a regular delivery resubscription rate and, more particularly, to a system for increasing a resubscription rate by recommending a personalized commodity suitable for needs of a customer on the basis of artificial intelligence. The present invention has an effect of improving a subscription rate by analyzing survey data received from a customer and data of existing customers on the basis of artificial intelligence and recommending a subscription product including a commodity that may be more preferred by the consumer. In addition, there is an effect of providing a product more suitable for needs of a customer by analyzing product purchase tendency data of the customer and recommending a changed product of a commodity that is required to be changed from among subscribed commodities as a purchase tendency of the customer is changed.
Need to check novelty before this filing date? Find Prior Art

Description

AI-based personalized product recommendation system to improve subscription rates for regular delivery services.

[0001] The present invention relates to a personalized product recommendation system for improving the re-subscription rate of a regular delivery service based on artificial intelligence, and more specifically, to provide a system for improving the re-subscription rate by recommending personalized products that meet the needs of customers based on artificial intelligence.

[0002] For everyday necessities like toilet paper, bottled water, and diapers, it's inconvenient for consumers to purchase them in bulk and store them in their homes or offices. However, purchasing them on an as-needed basis not only entails the hassle of selecting and ordering each item individually, but also often leads to situations where the items they need immediately are unavailable due to the lag between ordering and receiving.

[0003] Recently, to alleviate this inconvenience, online shopping malls have begun offering subscription delivery services, which deliver a set quantity of desired products at regular intervals (e.g., on a set date each month) with a single order. Specifically, customers select products and quantities online and enter subscription order information, such as the delivery date, delivery address, and payment information, on the order form to request subscription. The online shopping mall then automatically processes the order and payment on the designated date based on the subscription order information entered by the customer, ensuring regular delivery of the products to the customer.

[0004] However, in order to maintain the subscription rate for regular delivery, it is necessary to compose products that meet the preferences and needs of consumers, but this is currently difficult.

[0005] The purpose of the present invention is to improve subscription rates by recommending subscription products comprised of products more likely to be preferred by consumers by analyzing survey data received from customers and data of existing customers based on artificial intelligence.

[0006] In addition, the purpose is to provide products that better suit the needs of customers by analyzing customer product purchase tendency data and recommending products that need to be changed among the products subscribed to according to the change in customer purchase tendency.

[0007] In order to solve the above problem, the present invention provides a personalized product recommendation system for improving the re-subscription rate of regular delivery based on artificial intelligence, including a product subscription server (100) for collecting written questionnaire information created based on a pre-registered questionnaire from a customer terminal (200), collecting product purchase history information and forming a database into big data, and recommending customized products to customers based on artificial intelligence based on the collected questionnaire information and the big data-ized product purchase history information, and a customer terminal (200) for accessing the product subscription server (100), wherein the product purchase history information of the product subscription server (100) includes purchase time information, product history information, sales trends within a set period, product re-purchase rate, and purchase rate compared to product page clicks.

[0008] The present invention has the effect of improving subscription rates by recommending subscription products composed of products that consumers are more likely to prefer by analyzing survey data received from customers and data of existing customers based on artificial intelligence.

[0009] In addition, by analyzing the customer's product purchase tendency data, it is possible to recommend products that need to be changed among the products subscribed to according to the customer's purchase tendency change, thereby providing products that better meet the customer's needs.

[0010] Figure 1 is a general configuration diagram of the present invention.

[0011] Figure 2 is a detailed configuration diagram of the product subscription server (100) of the present invention.

[0012] Figure 3 is a drawing showing an embodiment of the present invention.

[0013] Figure 4 is a flowchart of the present invention.

[0014] Terms or words used in this specification and claims should not be interpreted as limited to their usual or dictionary meanings, but should be interpreted as meanings and concepts that conform to the technical idea of ​​the present invention, based on the principle that the inventor can appropriately define the concept of the term to explain his or her invention in the best possible way.

[0015] Hereinafter, a preferred embodiment of the present invention will be described in detail with reference to the attached drawings.

[0016]

[0017] The "user terminal" mentioned below may be implemented as a computer or portable terminal that can access a server or other terminal via a network. Here, the computer includes, for example, a notebook, desktop, or laptop equipped with a web browser, and the portable terminal may include, for example, all types of handheld-based wireless communication devices that ensure portability and mobility, such as IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), and LTE (Long Term Evolution) terminals. In addition, the "network" may be implemented as a wired network such as a local area network (LAN), a wide area network (WAN), or a value-added network (VAN), or all types of wireless networks such as a mobile radio communication network or a satellite communication network.

[0018]

[0019] The present invention relates to a personalized product recommendation system for improving the re-subscription rate of a regular delivery service based on artificial intelligence, and more specifically, to provide a system for improving the re-subscription rate by recommending personalized products that meet the needs of customers based on artificial intelligence.

[0020] Figure 1 is a diagram showing the overall configuration of a personalized product recommendation system for improving the re-subscription rate of a regular delivery service based on artificial intelligence according to the present invention, which is composed of a product subscription server (100) and a customer terminal (200).

[0021]

[0022] Figure 2 is a detailed configuration diagram of the product subscription server (100) of the present invention.

[0023]

[0024] The product subscription server (100) collects written survey information based on a pre-registered survey from a customer terminal (200), collects product purchase history information and forms a database to create big data, and recommends customized products to customers based on artificial intelligence based on the collected survey information and the big data-ized product purchase history information.

[0025]

[0026] The above product subscription server (100) is configured to include a survey information collection unit (110), a customer data collection unit (120), a product recommendation unit (130), a change product recommendation unit (140), a product sharing compensation unit (150), and an affiliate recommendation unit (160).

[0027]

[0028] The survey information collection unit (110) is intended to provide a pre-registered survey to a customer terminal (200) and collect the completed survey information.

[0029]

[0030] The customer data collection unit (120) collects product purchase history information including purchase time information, product details, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks, and creates a database to form big data.

[0031]

[0032] In other words, product purchase history information including purchase time information, product details, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks is collected and formed into big data, but the data can be divided and stored by time, purchase type, etc.

[0033]

[0034] The product recommendation section (130) is intended to recommend customized products to customers based on artificial intelligence based on the survey information collected through the survey information collection section and the product purchase history information converted into big data through the customer data collection section.

[0035]

[0036] Artificial Intelligence (AI) refers to a technology that mimics human learning, reasoning, and perception abilities and implements them on a computer. It can include concepts such as machine learning and symbolic logic. Machine Learning (ML) is an algorithm technology that classifies or learns the characteristics of input data on its own. AI technology can analyze input data using machine learning algorithms, learn the results of that analysis, and make judgments or predictions based on the results of that learning. In addition, technologies that utilize machine learning algorithms to mimic human brain functions such as cognition and judgment can also be understood as part of the category of AI. For example, this can include technical fields such as linguistic understanding, visual understanding, inference / prediction, knowledge representation, and motion control.

[0037]

[0038] The above product recommendation unit (130) calculates product page interest and reflects it in customized product recommendations, which can be calculated using [Mathematical Formula 1]. In other words, the degree of customer interest can be objectively analyzed by quantifying the degree of focus on the product page, thereby enabling better recommendations of products in which customers have a high level of interest.

[0039]

[0040]

[0041] Here, N is the product page interest, d is the average page retention time, p is the total time to maintain a specific focus, and U is the average page retention time. his a specific concentration duration that has been sustained at least twice, l h is the number of duplicate detection areas of the hth duplicate detected, B i is the duration of the ith specific concentration, q is the frequency of generation of the specific concentration, and z is the sum of the frequencies of duplicate detection of the specific concentration in at least two areas.

[0042]

[0043] Here, both the average page retention time and the specific focus retention time are based on time, and preferably in seconds.

[0044] In addition, the specific concentration is analyzed through the position of the eyes extracted from the screen viewed by the customer when accessing the product subscription server (100) through the customer terminal (200). If the eyes are located at a specific position, it can be determined that the customer is concentrating and thus can be determined to have a specific concentration.

[0045]

[0046] For example, if the average page retention time of customers entering a specific product page is 180 seconds, and the total retention time of a specific focus when a customer enters the product page and views the product is 100 seconds, and the specific focus lasted a total of 3 times, and the first measured retention time of a specific focus was 10 seconds, the first measured retention time of a specific focus was 8 seconds, the first measured retention time of a specific focus was 15 seconds, and the specific focus was detected repeatedly in at least two places once, and the retention time of a specific focus detected repeatedly is 15 seconds, and the number of duplicate detection areas is 2, then the product page interest is calculated as follows.

[0047]

[0048]

[0049] In other words, the product page interest N is calculated as 7.51, and the product page interest calculated in this way can be objectively analyzed by calculating the degree to which the customer focused on the product page as a numerical value, so that the customer's preference can be analyzed more clearly by determining the degree of interest in the product page.

[0050]

[0051] The change product recommendation section (140) is intended to recommend change products based on artificial intelligence so that products that require change can be changed among the product compositions based on the analyzed product purchasing tendency.

[0052]

[0053] The above-mentioned change product recommendation section (140) further includes a product purchase tendency analysis section (141).

[0054]

[0055] The above product purchase tendency analysis unit (141) is intended to analyze the product purchase tendency by receiving the customer's Internet search pattern and product purchase history from the customer terminal (100).

[0056]

[0057] Specifically, based on customer purchasing tendencies, similar products are clustered and patterned, and the purchase frequency of products with specific patterns is analyzed. Furthermore, customer purchase cycles are analyzed to identify changes in customer purchasing tendencies, ultimately leading to a comprehensive analysis of product purchasing tendencies.

[0058]

[0059] In addition, the above-mentioned change product recommendation unit (140) can receive answers to a predetermined purchase satisfaction survey question for product K from the customer terminal (200), receive the expected satisfaction score of product K expected by the seller, and provide an actual satisfaction index value, which means a numerical representation of a scale of how accurately the seller understands the customer's needs and selects a product to sell. In other words, if the actual satisfaction index value calculated in [Mathematical Formula 2] below is low, it can be determined that the seller has a high level of understanding of the customer's needs, and if the satisfaction index value is high, it allows for the seller to quickly prepare countermeasures.

[0060]

[0061] The above actual satisfaction index value can be calculated using [Mathematical Formula 2].

[0062]

[0063] Here M i is the actual satisfaction index value, and σ W is the variance of the expected satisfaction score, σ R is the variance of the purchase satisfaction survey response scores, m is the total number of survey questions, and c is the number assigned to each survey question. In addition, c=1,2,...,m, and m>1.

[0064]

[0065] For example, when the remaining values ​​are the same and the expected satisfaction score is set differently, the actual satisfaction index value M i If you calculate it, it is as follows.

[0066]

[0067] (Example 1)

[0068] m is 4, σ W This 32.8, σ R When these are 7.2, 6.1, 8.1, and 7 respectively, the actual satisfaction index value M is calculated using mathematical formula 1. i If you calculate it, it is as follows.

[0069]

[0070]

[0071]

[0072] (Example 2)

[0073] m is 4, σ W This 28, σ R When these are 7.2, 6.1, 8.1, and 7 respectively, the actual satisfaction index value M is calculated using mathematical formula 1. i If you calculate it, it is as follows.

[0074]

[0075]

[0076]

[0077] The actual satisfaction index value M is calculated by the above formula. i We can see that it is 1.08 in Example 1 and 0.99 in Example 2.

[0078]

[0079] Through this, it is possible to determine whether the actual satisfaction index value calculated through this is higher or lower than the standard value, and the calculated values ​​can be compared to determine whether they are higher or lower.

[0080]

[0081] Through the actual satisfaction index value calculated in this way, it is possible to determine how high or low it is compared to a preset standard value, and if it is lower than the standard value or in relative comparison, it is determined that the product meets the customer's needs, and the frequency of recommendation of K product can be increased in the future or a specific product can be set to be displayed at the top. In addition, if it is higher than the set standard value or in relative comparison, another product can be sold to replace K product.

[0082]

[0083] The product sharing compensation department (150) is intended to provide expected compensation information by analyzing the sales conversion rate of shared products compared to product sharing information shared on an external website through a customer terminal (200).

[0084] The above product sharing compensation unit (150) can calculate the percentage of views within the scope of exposure of shared content when product sharing is used as content. In other words, the compensation unit can provide compensation to customers by calculating the click-through rate, which measures how many people access the page where the content is posted, and the number and rate of purchases of clicked products.

[0085]

[0086] There may be a way to provide customers with expected compensation information about how much compensation they will receive by assigning compensation weights to the method of providing compensation.

[0087]

[0088] That is, for example, the product sharing compensation unit (150) assigns a compensation weight based on the number of product sharing posts, product sales price, number of product sales, number of visits to product sales pages, etc., and provides compensation prediction information, which is information that predicts the product compensation amount based on the assigned compensation weight.

[0089]

[0090] The reward weight is calculated by analyzing the number of product sales compared to the number of product sharing posts and assigning a reward weight based on a preset standard. However, by assigning a higher reward weight to products with a higher purchase conversion rate compared to the number of shared posts, it has the effect of allowing users to focus on sharing product sharing posts for products with a high purchase conversion rate.

[0091]

[0092] Additionally, the reward weight is awarded based on a pre-set standard by analyzing the number of visits to the product sales page compared to the number of product sharing posts, but it has the effect of allowing focus on creating product sharing posts.

[0093]

[0094] Additionally, if a reward weight is continuously set higher than the standard, a corresponding reward can be paid.

[0095] By actively promoting products that users use, if users see the promoted content and make a purchase, they are rewarded, and by giving a reward weight, they are motivated to promote, so both sellers and customers benefit.

[0096]

[0097] The affiliate recommendation section (160) is intended to recommend product information of affiliates located close to the customer's location based on customer information.

[0098]

[0099] The above affiliate recommendation section (160) is provided by big data formed by database of customer reviews, expert analysis, etc., or is provided by filtering based on artificial intelligence. It is desirable to recommend only affiliates located within a set distance from the customer's location so that the customer can easily access them.

[0100]

[0101] This can provide a structure that can benefit from cooperation with franchisees, and also has the effect of increasing customer trust by allowing customers to receive recommendations for companies and products that better meet their needs.

[0102]

[0103] Figure 3 is a drawing showing an embodiment of the present invention.

[0104]

[0105] Survey information sent by consumers is collected by the survey information collection department (n1), and based on the collected survey information and product purchase history information converted into big data through the customer data collection department, customized products are recommended to customers based on artificial intelligence (n2), and if a consumer (customer) who confirms this applies for a subscription service, he or she can receive a regular purchase product (n3).

[0106] Additionally, consumers can share the products they are using with other consumers (n4) and receive compensation if the shared product leads to a purchase.

[0107]

[0108] Referring to Figure 4, the method for recommending personalized products to improve the re-subscription rate for AI-based regular delivery is as follows.

[0109]

[0110] The personalized product recommendation method of the present invention, based on the personalized product recommendation system for improving the re-subscription rate of the above artificial intelligence-based regular delivery service, proceeds by including a customer data collection step (S1), a survey information collection step (S2), an artificial intelligence analysis step (S3), a product recommendation step (S4), and a changed product recommendation step (S5).

[0111]

[0112] S1) Customer data collection stage

[0113] The customer data collection stage (S1) is the stage where product purchase history information, including hourly information, product details, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks, is collected and databased to form big data.

[0114]

[0115] That is, the customer data collection step (S1) can form big data through the customer data collection unit (120) of the product subscription server (100).

[0116]

[0117] S2) Survey information collection stage

[0118] The survey information collection stage (S2) is a stage in which a pre-registered survey is provided to the customer terminal (200) and the written survey information is collected.

[0119]

[0120] The above survey information collection step (S2) can collect survey information through the survey information collection unit (110).

[0121]

[0122] S3) Artificial Intelligence Analysis Stage

[0123] The artificial intelligence analysis stage (S3) is a stage that analyzes, based on artificial intelligence, the survey information collected through the survey information collection stage (S2) and the product purchase history information converted into big data through the customer data collection stage (S1).

[0124]

[0125] The above artificial intelligence analysis step (S3) can analyze the survey information and product purchase history information through the product recommendation unit (130) based on artificial intelligence, and can recommend the most optimal product to the customer through artificial intelligence-based analysis.

[0126]

[0127] S4) Product recommendation stage

[0128] The product recommendation stage (S4) is the stage where customized products are recommended to customers based on the analyzed information.

[0129]

[0130] The product recommendation step (S4) can recommend a customer-tailored product analyzed based on artificial intelligence through the artificial intelligence analysis step (S3) above, and can be performed in the product recommendation unit (130).

[0131]

[0132] In addition, the method for recommending personalized products to improve the re-subscription rate of the above-mentioned artificial intelligence-based regular delivery service may proceed with the change product recommendation stage (S5), product sharing reward stage (S6), and affiliate recommendation stage (S7).

[0133]

[0134] S5) Change product recommendation stage

[0135] The change product recommendation stage (S5) is a stage that recommends change products based on artificial intelligence so that products that need to be changed can be changed among the products in the composition based on the analyzed product purchasing tendency.

[0136]

[0137] That is, it is carried out through the change product recommendation section (140) to analyze the product purchase tendency by receiving the customer's Internet search pattern and product purchase history from the customer terminal (100).

[0138]

[0139] Specifically, based on customer purchasing tendencies, similar products are clustered and patterned, and the purchase frequency of products with specific patterns is analyzed. Furthermore, customer purchase cycles are analyzed to identify changes in customer purchasing tendencies, ultimately leading to a comprehensive analysis of product purchasing tendencies.

[0140]

[0141] S6) Product sharing compensation stage

[0142] The product sharing compensation stage (S6) is a stage that provides expected compensation information by analyzing the sales conversion rate of shared products compared to product sharing information shared on external websites through the customer terminal held by the customer.

[0143]

[0144] The above-mentioned product sharing compensation step (S6) is implemented via the product sharing compensation unit (150). When product sharing is used as content, the percentage of views within the scope of the shared content can be calculated. In other words, the click-through rate, which measures the number of visits to the page where the content is posted, and the number and rate of purchases of clicked products can be calculated to provide compensation to customers.

[0145]

[0146] There may be a way to provide customers with expected compensation information about how much compensation they will receive by assigning compensation weights to the method of providing compensation.

[0147]

[0148] That is, for example, the product sharing compensation unit (150) assigns a compensation weight based on the number of product sharing posts, product sales price, number of product sales, number of visits to product sales pages, etc., and provides compensation prediction information, which is information that predicts the product compensation amount based on the assigned compensation weight.

[0149]

[0150] The reward weight is calculated by analyzing the number of product sales compared to the number of product sharing posts and assigning a reward weight based on a preset standard. However, by assigning a higher reward weight to products with a higher purchase conversion rate compared to the number of shared posts, it has the effect of allowing users to focus on sharing product sharing posts for products with a high purchase conversion rate.

[0151]

[0152] Additionally, the reward weight is awarded based on a pre-set standard by analyzing the number of visits to the product sales page compared to the number of product sharing posts, but it has the effect of allowing focus on creating product sharing posts.

[0153]

[0154] Additionally, if a reward weight is continuously set higher than the standard, a corresponding reward can be paid.

[0155] By actively promoting products that users use, if users see the promoted content and make a purchase, they are rewarded, and by giving a reward weight, they are motivated to promote, so both sellers and customers benefit.

[0156]

[0157] S7) Affiliate recommendation stage

[0158] The affiliate recommendation stage (S7) is a stage that recommends product information from affiliates located close to the customer's location based on customer information.

[0159]

[0160] The above-mentioned franchise recommendation step (S7) can be carried out through the franchise recommendation unit (160), and is provided by big data formed by database of customer reviews, expert analysis, etc., or by filtering based on artificial intelligence. It is desirable to recommend only franchises located within a set distance from the customer's location so that the customer can easily access them.

[0161]

[0162] This can provide a structure that can benefit from cooperation with franchisees, and also has the effect of increasing customer trust by allowing customers to receive recommendations for companies and products that better meet their needs.

[0163]

[0164] The embodiments described in this specification and the configurations illustrated in the drawings are merely the most preferred embodiments of the present invention and do not represent all of the technical ideas of the present invention. Therefore, it should be understood that there may be various equivalents and modified examples that can replace them.

Claims

1. A product subscription server (100) for collecting written survey information based on a pre-registered survey from a customer terminal (200), collecting product purchase history information and forming a database to create big data, and recommending customized products to customers based on artificial intelligence based on the collected survey information and the big data-ized product purchase history information. Includes a customer terminal (200) for connecting to the above product subscription server (100), The product purchase history information of the above product subscription server (100) is Includes purchase time information, product details, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks. Product subscription server (100) A survey information collection unit (110) that provides a pre-registered survey to a customer terminal (200) and collects the completed survey information, A customer data collection unit (120) for collecting product purchase history information including purchase time information, product history information, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks, and creating a database to form big data; A product recommendation unit (130) for recommending customized products to customers based on artificial intelligence based on the survey information collected through the survey information collection unit and the product purchase history information converted into big data through the customer data collection unit, It includes a change product recommendation unit (140) for recommending change products so that products that need to be changed among the composition products can be changed based on the analyzed product purchase tendency, based on artificial intelligence. The change product recommendation department (140) It is characterized by further including a product purchase tendency analysis unit (141) for analyzing product purchase tendencies by receiving customer Internet search patterns and product purchase history from a customer terminal (100). The above product recommendation section (130) includes calculating product page interest using the following mathematical formula 1 and reflecting it in customized product recommendations. [Mathematical formula 1] is (Here, N is the product page interest, d is the average page retention duration, p is the total time to maintain a specific focus, U h is a specific concentration duration that has been sustained at least twice, l h is the number of duplicate detection areas of the hth duplicate detected, B i is the duration of the ith specific concentration, q is the frequency of generation of the specific concentration, and z is the sum of the frequencies of the specific concentration being detected repeatedly in at least two areas.) It is characterized by, The above change product recommendation section (140) Receiving answers to a purchase satisfaction survey question of a specific product from a customer terminal (200), and receiving an expected satisfaction score of a specific product expected by the seller, and providing an actual satisfaction index value, if the actual satisfaction index value calculated in [Mathematical Formula 2] is a low value, it is determined that the seller has a high level of understanding of the customer's needs. [Mathematical formula 2] is (Here M i is the actual satisfaction index value, and σ W is the variance of the expected satisfaction score, σ R A personalized product recommendation system for improving the re-subscription rate of artificial intelligence-based regular delivery, characterized in that (i) is the variance of the response scores to the purchase satisfaction survey, m is the total number of survey questions, and c is the number value assigned to each survey question. In addition, c=1,2,...,m, and m>1.

2. A product subscription server (100) for collecting written survey information based on a pre-registered survey from a customer terminal (200), collecting product purchase history information and forming a database to create big data, and recommending customized products to customers based on artificial intelligence based on the collected survey information and the big data-ized product purchase history information. Includes a customer terminal (200) for connecting to the above product subscription server (100), The product purchase history information of the above product subscription server (100) is Includes purchase time information, product details, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks. Product subscription server (100) A survey information collection unit (110) that provides a pre-registered survey to a customer terminal (200) and collects the completed survey information, A customer data collection unit (120) for collecting product purchase history information including purchase time information, product history information, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks, and creating a database to form big data; A product recommendation unit (130) for recommending customized products to customers based on artificial intelligence based on the survey information collected through the survey information collection unit and the product purchase history information converted into big data through the customer data collection unit, A change product recommendation unit (140) for recommending change products so that products that need to be changed among the products in the composition can be changed based on the analyzed product purchasing tendency, Includes a franchise recommendation unit (160) for recommending product information of franchises located close to the customer's location based on customer information. The change product recommendation department (140) It is characterized by further including a product purchase tendency analysis unit (141) for analyzing product purchase tendencies by receiving customer Internet search patterns and product purchase history from a customer terminal (100). The above product recommendation section (130) includes calculating product page interest using the following mathematical formula 1 and reflecting it in customized product recommendations. [Mathematical formula 1] is (Where N is the product page interest, d is the average page retention time, p is the total time to maintain a specific focus, is a specific concentration level that has been maintained for at least two consecutive times, is the number of duplicate detection areas of the hth specific concentration, is the duration of the ith specific concentration, q is the frequency of generation of the specific concentration, and z is the sum of the frequencies of the specific concentration being detected repeatedly in at least two areas.) It is characterized by, The above change product recommendation section (140) Receiving answers to a purchase satisfaction survey question for a specific product from a customer terminal (200), and receiving an expected satisfaction score for a specific product as expected by the seller, and providing an actual satisfaction index value. [Mathematical formula 2] is (Here M i is the actual satisfaction index value, and σ W is the variance of the expected satisfaction score, σ R An artificial intelligence-based personalized product recommendation and affiliate product recommendation system for improving subscription rate of regular delivery services, characterized in that (i) is the variance of the response scores to the purchase satisfaction survey, m is the total number of survey questions, and c is the number value assigned to each survey question. In addition, c=1,2,...,m, and m>1.

3. In paragraph 2, The above customer data collection unit (120) A personalized product recommendation and affiliate product recommendation system for improving the re-subscription rate of regular delivery services based on artificial intelligence, which collects product purchase history information including purchase time information, product details, sales trends within a set period, product repurchase rate, and purchase rate compared to product page clicks, and forms it into big data, but the data is divided and stored by time and purchase type.

Citation Information

Patent Citations

  • Apparatus and method for presenting personalized goods information based on artificial intelligence, and recording medium thereof

    KR100961782B1

  • Blockchain-based authentication system and method for preventing interception hacking attacks

    KR1020210129981A

  • Heat exchange apparatus for heat storage and cooling of greenhouse

    KR1020240043386A

  • Artificial intelligence-based customer-customized shopping mall product recommendation and automatic product page configuration system

    KR102225212B1

  • Personalized product recommendation system to improve the re-subscription rate of artificial intelligence-based regular delivery

    KR102585731B1

Cited By

  • Live broadcast advertisement video content optimization method and system based on artificial intelligence

    CN119515475A

  • Power business self-service intelligent recommendation method for enterprise users

    CN120763409A