Method and system for integrated management of non-inventory cosmetics experience, order and delivery based on multi-industry offline stores

Abstract

According to one embodiment, when a customer terminal scans a QR code of an experience zone installed in another store to access it, the system loads a store profile based on a store identification code included in the QR code, receives skin data from the customer terminal to create a customer skin profile, performs a similarity analysis between the customer skin profile and product attribute information, calculates a suitability score for each product by reflecting the industry characteristics of the store, generates a customized recommended product list based on the suitability score and transmits it to the customer terminal, processes payment for the product selected by the customer, transmits a delivery instruction to a logistics center upon completion of payment, calculates a commission per store based on the store identification code to perform automatic settlement, and controls the system to integrate and accumulate data generated during the experience and purchase process.

Description

Technology Field

[0001] The present invention relates to the field of O4O (Online for Offline) and artificial intelligence-based customized data processing technology that integrates offline retail distribution and e-commerce.

[0002] More specifically, this invention relates to an offline store-based zero-inventory cosmetic experience, order, and delivery integrated management method and system, in which a cosmetic brand headquarters operates a zero-inventory experience zone within a store other than its own dedicated store, and streamlines the processes of customer physical experience, AI-based personalized recommendations, QR payment, and direct delivery from a logistics center to eliminate the inventory burden on offline stores and innovate the distribution structure. Background Technology

[0004] Recently, offline distribution methods such as road shops, health and beauty (H&B) stores, and department stores, where customers purchase products after applying them directly to their skin and experiencing them, still account for a significant portion of the cosmetics industry. However, this conventional offline distribution structure based on dedicated stores entails high initial investment costs and rent for opening stores. Furthermore, since each branch must directly possess physical products to be sold, the burden of managing obsolete inventory and the risk of disposal due to expired products have been a significant burden for store owners.

[0005] Meanwhile, offline simple payment systems utilizing QR codes or mobile devices are being introduced as an attempt to connect online shopping and offline stores, but this remains at the level of merely replacing existing POS systems. In other words, it has not yet been realized as a fully integrated three-way (e.g., customer-headquarters-store owner) zero-inventory distribution platform that encourages customers to try out cosmetics in stores of other industries—such as nail salons, beauty salons, and cafes—rather than dedicated cosmetics stores; where artificial intelligence (AI) recommends customized products that take into account the store environment based on skin data collected there; and where products are delivered immediately from the central headquarters without any sales inventory in the store.

[0006] Furthermore, when managing dozens to hundreds of dispersed base experience stores, existing franchise management methods suffered from the problem of consuming massive human resources and operational costs, such as through on-site inspections by headquarters sales personnel. Objectively evaluating the performance of each store and manually identifying optimal marketing channels tailored to regional characteristics and industry types to allocate advertising budgets presented the challenge of difficult real-time response. The problem to be solved

[0008] The present invention aims to reduce the inventory burden and distribution risks for store owners by utilizing an existing store network without opening a dedicated store for the company's brand, establishing an inventory-free experience zone that stocks only samples for trial without sales inventory, and processing direct orders and delivery from the headquarters based on QR codes.

[0009] In addition, the present invention aims to provide a customized recommended product list optimized for the customer and the store environment through a unique AI algorithm that converts skin data entered by the customer into a multidimensional vector and weights color difference-based color matching that goes beyond simple product matching and reflects the industry affinity of the store where the experience zone is located and the specificity of a particular industry (e.g., a flower shop).

[0010] The present invention aims to establish a transparent profit-sharing structure without complex payment collection procedures by automatically transmitting a delivery instruction to a logistics center system immediately upon completion of payment via a customer terminal, and simultaneously automatically calculating and paying a store-specific settlement fee for the corresponding order in real time based on the store identification code of the experience zone.

[0011] The present invention aims to efficiently operate a multi-store experience network with minimal management personnel by automatically determining multiple performance check items, such as inducing experiences at individual stores, Naver reviews and community postings, and customer response metrics, to conditionally approve the payment of advertising fees, and by automatically selecting the optimal external marketing channel based on regional characteristics and target customer base. means of solving the problem

[0013] According to one embodiment, when a customer terminal scans a QR code of an experience zone installed in a store of a different industry to access it, the system loads a store profile based on a store identification code included in the QR code, receives skin data from the customer terminal to create a customer skin profile, performs a similarity analysis between the customer skin profile and product attribute information, calculates a suitability score for each product by reflecting the industry characteristics of the store, generates a customized recommended product list based on the suitability score and transmits it to the customer terminal, processes payment for the product selected by the customer, transmits a delivery instruction to a logistics center upon completion of payment, calculates a commission per store based on the store identification code to perform automatic settlement, and controls the system to integrate and accumulate data generated during the experience and purchase process.

[0014] According to one embodiment, when a customer terminal scans a QR code of an experience zone installed in a store of a different industry to access it, the system (400) loads a store profile containing information on the industry and commission rate of the store based on a store identification code encoded in the QR code, and transmits a skin type diagnosis interface to the customer terminal to receive skin data including skin type, skin tone, and skin concerns from the customer. The system (400) converts the received skin data into a multidimensional vector to generate a customer skin profile vector, performs a similarity analysis between the customer skin profile vector and the product attribute vector, calculates a skin suitability score for each product by reflecting the industry affinity between the store industry and the product category corresponding to the store identification code, generates a customized recommended product list based on the skin suitability score and transmits it to the customer terminal, processes payment for the product selected by the customer through an external payment server, automatically transmits a delivery instruction to a logistics center system upon completion of payment, automatically calculates a commission for each store based on the store identification code to perform automatic settlement to the store owner's account, and controls the system to integrate and accumulate data generated during the experience and purchase process by customer, store, and product.

[0015] According to one embodiment, the system (400) loads an access URL containing a store identification code, a product identification code, and an experience zone type code encoded in a QR code into a web browser of a customer terminal, queries a store profile database based on the store identification code to load the industry classification, location information, commission rate, and target customer base information of the corresponding store, quantifies the skin type, skin tone, major concerns, and age group information received from the customer, calculates a seasonal correction value based on the current month, converts it into a multidimensional vector including a skin type value, a skin tone value, a weight for each concern item, an age group value, and a seasonal correction value, and can generate a customer skin profile vector. The system (400) calculates a cosine similarity between a customer skin profile vector and a product attribute vector pre-set for each product, calculates a product-specific skin suitability score by adding a value with a first weight applied to the cosine similarity, a value with a second weight applied to a product-specific seasonal correction value according to the current season, a value with a third weight applied to an industry affinity between a store industry and a product category, and a value with a fourth weight applied to a purchase history bonus based on the customer's past purchase history, and normalizes the sum of the first to fourth weights to be 1.0, and controls the system to generate a customized recommended product list along with the skin suitability scores of each product by selecting a predetermined number of top products in order of high skin suitability scores.

[0016] According to one embodiment, the system (400) can automatically determine whether a plurality of performance check items are achieved for each experience zone store, and can determine whether the number of experience photos sent by the store to the headquarters server within a predetermined period is above a preset standard, whether the number of new reviews of the store collected through the Naver Place API is above a preset standard, whether the number of community platform postings within a predetermined period is above a preset standard, whether the headquarters inspection score and the negative review ratio each meet the preset standards, whether the number of non-responses to reservation customers is 0, and whether the customer SNS authentication postings and goods distribution performance are above a preset standard. The system (400) can automatically approve the payment of advertising fees for the current month for the store if the sum of the number of passed items among multiple performance check items is greater than or equal to a preset passing standard, and automatically send a warning notification to the store owner including items that are not met if the sum is less than the passing standard, and at the end of a predetermined pilot period, automatically classify the store into multiple grades based on the number of monthly QR sales per store and control the determination of differential measures including contract extension, conversion to store self-funding, or termination of the pilot based on the classified grade.

[0017] According to one embodiment, the system (400) can determine the business classification of a store corresponding to a store identification code, and if the business of the store is a flower shop, extract a representative HSV color value of a flower variety handled by the store from a flower variety database, extract a HSV color value of each product from a product database belonging to the lip and cheek category, and calculate the color difference between the representative HSV color value of the flower variety and the HSV color value of each product by squaring the difference values ​​of hue, saturation, and brightness, summing them, and taking the square root. The system (400) can select a product whose color difference value is below a preset threshold as a flower color matching recommendation product, determine whether the current time corresponds to the gift season, and if it corresponds to the gift season, increase the recommendation weight for the matching recommendation product, and control the creation of a business-specific cross-selling recommendation list by integrating the flower color matching recommendation product with a customized recommendation product list based on skin suitability scores and transmitting it to a customer terminal.

[0018] According to one embodiment, the system (400) inputs the industry classification of the store corresponding to the store identification code, the regional characteristics of the store location, and the target customer base data, retrieves a basic suitability score by industry from an industry-channel suitability matrix for each of the multiple advertising channels, and calculates a regional correction coefficient based on the regional characteristics of the store location, wherein if the store is located in a dense residential area, the correction coefficient for a region-based community channel is increased; if it is located in a dense commercial area, the correction coefficient for a region-based search channel is increased; and if it is located in an area with a large floating population of young people, the correction coefficient for a social media channel is increased. The system (400) calculates a final score for each channel by reflecting the regional correction coefficient in the basic suitability score by industry, automatically selects the channel with the highest final score as the channel eligible for headquarters advertising cost support for the store, and controls the execution of advertising for the automatically selected channel by linking it with whether the store's performance check items have been achieved, so that advertising costs are automatically executed for the selected channel only for stores that meet the performance criteria. Effects of the invention

[0020] The offline store-based integrated management method and system for zero-inventory cosmetic experience, ordering, and delivery according to the present invention can reduce the massive initial investment costs and the risk of obsolete inventory associated with opening offline stores by utilizing idle space in existing stores of other industries to establish an experience-oriented sales channel with no inventory burden.

[0021] In addition, the present invention introduces an advanced artificial intelligence algorithm that comprehensively analyzes not only simple skin data of customers but also the industry characteristics (e.g., industry affinity) and seasonal factors of the store where the experience zone is located, thereby providing customers with a hyper-personalized beauty experience that is perfectly harmonized with the surrounding environment and maximizing the purchase conversion rate for the store.

[0022] The present invention enables immediate direct delivery by linking customer QR order data with the headquarters' logistics system in real time, and ensures transaction transparency between the headquarters and numerous offline branches and reduces administrative costs associated with payment settlement by automatically calculating and settling the commission for the relevant store upon the occurrence of an order.

[0023] The present invention automatically and quantitatively evaluates the performance indicators of multiple stores to determine whether to execute advertising expenses, and intelligently selects and supports marketing channels optimized for each store's regional location and target customer base, thereby enabling the efficient management of a vast nationwide offline experience network with minimal management personnel and increasing the return on investment (ROI). Brief explanation of the drawing

[0025] FIG. 1 is a diagram illustrating the overall structure of an artificial intelligence-based system according to one embodiment. FIG. 2 is a diagram illustrating the learning of a neural network according to one embodiment. FIG. 3 is a diagram illustrating the configuration of an artificial intelligence model according to one embodiment. FIG. 4 is a block diagram showing the configuration of an offline store-based inventory-free cosmetic trial order delivery integrated management system according to one embodiment. FIG. 5 illustrates an integrated management method for ordering and delivering cosmetics for trial use without inventory, according to one embodiment. FIG. 6 illustrates a method for automatic evaluation of store performance and conditional execution of advertising costs according to one embodiment. Figure 7 shows the information processing flow of an industry-specific cross-selling color matching algorithm according to one embodiment. FIG. 8 illustrates the information processing process of an automatic guidance method linked with a carrier AI assistant according to one embodiment. Specific details for implementing the invention

[0026] Hereinafter, embodiments are described in detail with reference to the attached drawings. However, various modifications may be made to the embodiments, and thus the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments are included within the scope of the rights.

[0027] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, the embodiments are not limited to the specific disclosed forms, and the scope of this specification includes modifications, equivalents, or substitutions that fall within the technical concept.

[0028] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0029] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or coupled with that other component, or that there may be other components in between.

[0030] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0031] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0032] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. In describing the embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiments, such detailed description is omitted.

[0033] The embodiments can be implemented in various forms of products such as personal computers, laptop computers, tablet computers, smartphones, televisions, smart home appliances, intelligent automobiles, kiosks, and wearable devices.

[0034] Artificial Intelligence (AI) systems are computer systems that implement human-level intelligence; unlike existing rule-based smart systems, they are systems in which machines learn and make decisions autonomously. As AI systems improve in recognition accuracy and gain a more accurate understanding of user preferences with continued use, existing rule-based smart systems are gradually being replaced by deep learning-based AI systems.

[0035] Artificial intelligence technology consists of machine learning and component technologies utilizing machine learning. Machine learning is an algorithmic technology that autonomously classifies and learns the characteristics of input data, while component technologies are technologies that mimic the cognitive and judgmental functions of the human brain by utilizing machine learning algorithms such as deep learning, and are comprised of technological fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control.

[0036] The various fields where artificial intelligence technology is applied are as follows. Linguistic understanding refers to technologies that recognize, apply, and process human language and text, including natural language processing, machine translation, dialogue systems, question answering, and speech recognition / synthesis. Visual understanding refers to technologies that perceive and process objects like human vision, including object recognition, object tracking, image search, people recognition, scene understanding, spatial understanding, and image enhancement. Inference and prediction refers to technologies that logically reason and predict by judging information, including knowledge / probability-based inference, optimization prediction, preference-based planning, and recommendation. Knowledge representation refers to technologies that automatically process human experiential information into knowledge data, including knowledge construction (data generation / classification) and knowledge management (data utilization). Motion control refers to technologies that control the autonomous driving of vehicles and the movement of robots, including motion control (navigation, collision, driving) and manipulation control (behavior control).

[0037] Generally, to apply machine learning algorithms to real-world situations, training is performed using a trial-and-error method due to the inherent characteristics of the fundamental methodologies. In particular, deep learning requires hundreds of thousands of iterations. Since it is impossible to execute this in a real physical external environment, training is instead performed through simulations that virtually recreate the actual physical environment on a computer.

[0038] In the present invention, Artificial Intelligence (AI) refers to a technology that imitates human learning ability, reasoning ability, and perceptual ability, and implements them on a computer, and may include concepts such as machine learning and symbolic logic. Machine Learning (ML) is an algorithmic technology that classifies or learns the characteristics of input data on its own. AI technology can analyze input data as a machine learning algorithm, learn from the results of the analysis, and make judgments or predictions based on the results of the learning. Furthermore, technologies that mimic the functions of the human brain, such as cognition and judgment, by utilizing machine learning algorithms can also be understood as falling within the category of AI. For example, technological fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control may be included.

[0039] Machine learning can refer to the process of training neural network models using experience in processing data. Through machine learning, computer software can improve its own data processing capabilities. A neural network model is constructed by modeling the correlations between data, and these correlations can be expressed by multiple parameters. A neural network model extracts and analyzes features from given data to derive correlations between them; machine learning can be defined as the process of optimizing the model's parameters by repeating this process. For example, a neural network model can learn the mapping (correlation) between inputs and outputs for data given as input-output pairs. Alternatively, even when only input data is provided, a neural network model can derive regularities between the given data and learn those relationships.

[0040] An artificial intelligence learning model or neural network model can be designed to implement the structure of the human brain on a computer and may include multiple network nodes that have weights and simulate neurons of a human neural network. The multiple network nodes may have interconnected relationships by simulating the synaptic activity of neurons, where neurons exchange signals through synapses. In an artificial intelligence learning model, multiple network nodes may be located in layers of different depths and exchange data according to convolutional connections. The artificial intelligence learning model may be, for example, an Artificial Neural Network (ANN) or a Convolutional Neural Network (CNN). As an embodiment, the artificial intelligence learning model may be machine learned according to methods such as supervised learning, unsupervised learning, and reinforcement learning. Machine learning algorithms for performing machine learning may include Decision Tree, Bayesian Network, Support Vector Machine, Artificial Neural Network, Ada-boost, Perceptron, Genetic Programming, and Clustering.

[0041] Among these, CNNs are a type of multilayer perceptron designed to use minimal preprocessing. CNNs consist of one or more convolutional layers and standard artificial neural network layers stacked on top, additionally utilizing weights and pooling layers. Thanks to this structure, CNNs can fully utilize two-dimensional input data. Compared to other deep learning architectures, CNNs demonstrate good performance in both image and audio fields. CNNs can also be trained using standard backpropagation. CNNs have the advantage of being easier to train than other feedforward artificial neural network techniques and using a small number of parameters.

[0042] Convolutional networks are neural networks comprising sets of nodes with bounded parameters. Many computer vision tasks have been significantly improved, driven by the increased size of available training data and the availability of computational power, combined with algorithmic advancements such as discriminative linear units and dropout training. In the case of massive datasets, such as those available for many tasks today, outfitting is not critical, and increasing the network size improves test accuracy. Optimal utilization of computing resources becomes a limiting factor. To address this, distributed, scalable implementations of deep neural networks can be employed.

[0044] FIG. 1 is a diagram illustrating the overall structure of an artificial intelligence-based system according to one embodiment.

[0045] As illustrated in FIG. 1, an artificial intelligence-based system (100) may include a plurality of user terminals (110-1 to 110-n), a server (120), and a database (130). This system adopts a distributed architecture and operates based on a client-server model, with each component playing a unique role and contributing to increasing the efficiency and scalability of the entire system. According to one embodiment, the database (130) is depicted as being configured separately from the server (120), but is not limited thereto; depending on system design and operational efficiency, the database (130) may be integrated within the server (120). This integrated configuration has the advantage of improving data access speed and reducing system complexity. For example, the server (120) may include a plurality of artificial intelligence models and processing units for performing machine learning algorithms, and these can implement various types of deep learning and machine learning technologies (e.g., CNN, RNN, Transformer, reinforcement learning, etc.) to respond to user requests and provide intelligent services. According to one embodiment, a plurality of user terminals (110-1 to 110-n), a server (120), and a database (130) can be connected to communicate with each other through a network (N), which enables real-time data exchange and smooth interaction between system components.

[0046] According to one embodiment, the network (N) may perform wireless or wired communication between a plurality of user terminals (110-1 to 110-n), a server (120), a database (130), etc., and is located in the center of FIG. 1 to serve as a hub connecting all components. For example, the network may perform wireless communication according to methods such as 5G, LTE (Long-Term Evolution), LTE-A (LTE Advanced), CDMA (Code Division Multiple Access), WCDMA (Wideband CDMA), WiBro (Wireless BroadBand), WiFi (Wireless Fidelity), Bluetooth, NFC (Near Field Communication), GPS (Global Positioning System), or GNSS (Global Navigation Satellite System). The 5G network supports high-speed data transmission (up to 20Gbps), ultra-low latency (1ms or less), and large-scale connectivity, making it suitable for real-time processing of large-capacity AI models. For example, the network (N) may be configured to perform wired communication using methods such as USB (Universal Serial Bus), HDMI (High Definition Multimedia Interface), RS-232 (Recommended Standard 232), Ethernet, fiber optic cable, or POTS (Plain Old Telephone Service). In particular, in data center environments, high-performance network technologies such as 400Gbps Ethernet or InfiniBand can be utilized for communication between large-scale AI computing clusters.

[0047] According to one embodiment, user terminals (110-1 to 110-n) are various client devices that access the system and can be implemented in various forms, such as smartphones, tablets, desktop computers, wearable devices, and IoT devices. These terminals are responsible for transmitting service requests through a user interface and receiving and displaying results processed by a server (120). Each terminal may also run a lightweight AI model locally, which can reduce network latency and enhance privacy protection. Direct connection or peer-to-peer (P2P) communication between terminals may be possible, as indicated by the dotted line, which is particularly useful in distributed learning or edge computing scenarios.

[0048] According to one embodiment, the server (120) is the central processing unit of the system, responsible for processing user requests, executing artificial intelligence models, and managing data. The server is equipped with a high-performance processor (CPU, GPU, TPU, NPU, etc.) for large-scale computation, large-capacity memory, and a stable operating system, and may be operated in a virtualized manner on a cloud infrastructure. The server performs training and inference functions for complex deep learning models and transmits the results to user terminals or stores them in a database. In addition, the stability and reliability of the system can be ensured through functions such as load balancing, error recovery, and security management.

[0049] According to one embodiment, a database (130) is a storage facility for storing and managing various data, and is depicted in a cylindrical shape in FIG. 1. Data stored in the database (130) is data acquired, processed, or used by at least one component of a plurality of user terminals (110-1 to 110-n) or a server (120), and may include software (e.g., programs), user profiles, training datasets, model weights, log records, etc. Structured relational data may be stored in a SQL-based database (MySQL, PostgreSQL, etc.), and unstructured large-scale data may be stored in a NoSQL database (MongoDB, Cassandra, etc.) or a distributed file system (Hadoop HDFS, etc.). The database (130) may include volatile and / or non-volatile memory, where volatile memory (RAM) is used for caching or temporary data processing requiring fast data access, and non-volatile memory (SSD, HDD, tape drive, etc.) is used for permanent data storage. Databases also ensure the security and integrity of data through data encryption, access control, and backup and recovery mechanisms.

[0050] This artificial intelligence-based system (100) can be utilized in various application fields and can provide services such as user behavior analysis, recommendation systems, natural language processing, image and video recognition, and predictive analysis. In particular, the distributed architecture and scalable design of the system enable stable performance to be maintained even as the number of users and data volume increase.

[0052] FIG. 2 is a diagram illustrating the learning of a neural network according to one embodiment.

[0053] As illustrated in FIG. 2, the learning device can train a neural network (123) to classify review responses received from a plurality of user terminals (110-1 to 110-n) by item. Additionally, the learning device can train a neural network (123) to extract user stay history from user movement path information. The neural network (123) is also called an artificial neural network and is a computational model designed inspired by the neural structure of the human brain, specialized in recognizing and learning complex patterns within data. According to one embodiment, the learning device may be a separate entity from the server (120), but it may also be implemented integrated into the same system, so it is not limited thereto.

[0054] According to one embodiment, the neural network (123) includes an input layer (121) into which training samples are input and an output layer (125) that outputs training outputs, and can be learned based on the difference between the training outputs and labels (i.e., actual correct data). Here, the labels are defined based on items corresponding to review responses (e.g., service quality, price satisfaction, cleanliness, etc.) and can be defined based on user dwell history (place visited, time spent, frequency of visit, etc.) corresponding to movement path information. The neural network (123) is connected as a group of multiple nodes and is defined by weights between the connected nodes and an activation function that activates the nodes. Various forms of the activation function may be used, such as sigmoid, hyperbolic tangent (tanh), and ReLU, which enables the network to learn non-linear patterns.

[0055] According to one embodiment, the learning device can train a neural network (123) using a Gradient Descent (GD) technique or a Stochastic Gradient Descent (SGD) technique. The GD technique is a method of updating weights all at once by calculating gradients based on the entire dataset, while the SGD technique is a method of increasing computational efficiency and the possibility of escaping a local optimum by updating weights more frequently using only a randomly selected portion of data (mini-batch). The learning device can use a loss function designed by the outputs and labels of the neural network. The loss function is an important element that quantifies the difference between the model's predicted value and the actual correct answer to suggest the direction of learning.

[0056] The learning device can calculate the training error using a predefined loss function. The loss function can be predefined with labels, outputs, and parameters as input variables, where the parameters can be set by weights within the neural network (123). For example, the loss function can be designed in the form of Mean Square Error (MSE), entropy, etc. MSE is calculated as the squared mean of the difference between the predicted value and the actual value and is mainly used for regression problems, while cross-entropy is a function that measures the difference between the predicted probability distribution and the actual distribution and is suitable for classification problems. In addition, loss functions that are less sensitive to outliers, such as Huber Loss, can be utilized, and various techniques or methods can be employed in the embodiments where the loss function is designed.

[0057] According to one embodiment, the learning device can identify weights that influence the training error using a backpropagation technique. Backpropagation is a process of calculating the degree to which each weight contributes to the final error while propagating the error calculated in the output layer toward the input layer, and is performed through differential calculations using the chain rule. Here, the weights are relationships between nodes within the neural network (123). The learning device may use an SGD technique using labels and outputs to optimize the weights found through the backpropagation technique. For example, the learning device may update the weights of a loss function defined based on labels, outputs, and weights using an SGD technique. This process operates by calculating the gradient (the derivative of the loss function with respect to the weights) and adjusting the weights in the direction of this gradient to gradually decrease the value of the loss function.

[0058] According to one embodiment, the learning device extracts first objects from a review response, obtains first labels which are items corresponding to the first objects, applies the first objects to a first neural network to generate first training outputs corresponding to the first objects, and can train the first neural network based on the first training outputs and the first labels. At this time, the first objects may include key keywords, sentence structures, sentiment expressions, etc. extracted from the review text, and thereby learn the ability to identify the core content of the review and classify it into the corresponding items.

[0059] According to one embodiment, the learning device extracts second objects from movement path information, obtains second labels which are user dwell history corresponding to the second objects, applies the second objects to a second neural network to generate second training outputs corresponding to the second objects, and can train the second neural network based on the second training outputs and the second labels. The second objects may consist of data such as the user's GPS coordinate sequence, movement speed, stopping point, and movement pattern, thereby learning the ability to determine where and for how long the user stayed.

[0060] According to one embodiment, a learning device can generate first training feature vectors based on constituent features (e.g., sentence structure, word frequency, part-of-speech distribution), positional features (e.g., location of important words, location of key sentences), and pattern features (e.g., repetitive expressions, usage patterns of specific phrases) of a review response. These feature vectors are generated through a process of converting raw text data into a numerical form that can be processed by a neural network, and various methods such as TF-IDF, Word2Vec, and BERT may be employed to extract features.

[0061] According to one embodiment, the learning device can generate second training feature vectors based on constituent features of movement path information (e.g., shape of the path, estimation of the means of movement), length features (e.g., total travel distance, travel time, distance between each point), and pattern features (e.g., repeated visit patterns, movement patterns by day of the week / time of the day). Since location data has continuous values ​​over time, time-series data processing techniques or spatial data analysis methodologies may be applied, and various methods may be employed to extract features.

[0062] According to one embodiment, the learning device can obtain training outputs by applying first training feature vectors to the neural network (123). This process is also called feedforward and is a process in which input data passes through each layer of the network to generate a final prediction value. The learning device can train the review item extraction algorithm of the neural network (123) based on the training outputs and first labels. The learning device can train the review item extraction algorithm of the neural network (123) by calculating training errors corresponding to the training outputs and optimizing the connection relationships of nodes within the neural network (123) to minimize the training errors. This process is repeated over several epochs, and in each epoch, the entire training dataset is processed and weights are updated. The server (120) can automatically extract and classify items from new review responses using the first neural network that has been trained.

[0063] According to one embodiment, the learning device can obtain training outputs by applying second training feature vectors to the neural network (123). The learning device can train the user stay history acquisition algorithm of the neural network (123) based on the training outputs and second labels. The learning device can train the user stay history acquisition algorithm of the neural network (123) by calculating training errors corresponding to the training outputs and optimizing the connection relationships of nodes within the neural network (123) to minimize the training errors. Since movement path data has temporal continuity, a recurrent neural network (RNN) structure such as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit) can be applied, which enables effective learning of the patterns of the sequence data. The server (120) can obtain user stay history from movement path information using the second neural network after training is complete, and the information thus obtained can be utilized in various application fields such as improving location-based services, building customized recommendation systems, and analyzing traffic patterns.

[0065] FIG. 3 is a diagram illustrating the configuration of an artificial intelligence model according to one embodiment.

[0066] An artificial intelligence model according to one embodiment may include an input layer, a hidden layer, and an output layer. This multi-layer structure is effective for learning complex patterns and solving non-linear problems, and is utilized as a core structure, particularly in deep learning models. The input layer is a layer related to input values ​​input into the artificial intelligence model and serves to receive data coming from the outside. As illustrated in FIG. 3, the input layer consists of three nodes, and each node may represent a specific feature or attribute of the data. The number of nodes in the input layer may vary depending on the dimension or characteristics of the data to be processed, and may be determined by the number of pixels in the case of image data, the dimension of the word vector in the case of text data, etc.

[0067] According to one embodiment, a feature map can be output by performing a MAC (multiply-accumulate) operation and an activation operation on the input values ​​in the hidden layer. In FIG. 3, the hidden layer consists of four nodes, and these nodes are connected to all nodes of the input layer to form a fully connected structure. The MAC operation may be an operation that multiplies the input values ​​by their corresponding weights and sums the multiplied values, and this acts as the basic unit of information processing in a neural network.

[0068] According to one embodiment, the activation operation may be an operation that inputs the result of a MAC operation into an activation function and outputs a result value. The activation function is a key element that enables a model to learn complex patterns by introducing non-linearity into the result of a linear operation. The activation function may be of various types. For example, the activation function may include a sigmoid function (outputting a value between 0 and 1 and suitable for expressing probability), a tangent function (outputting a value between -1 and 1 and suitable for balanced data), a ReLU function (a function that outputs 0 for negative inputs and outputs as is for positive inputs, mitigating the vanishing gradient problem), a Leaky ReLU function (applying a linear function with a small gradient to negative inputs), a MaxOut function (selecting the maximum value among several linear functions), and / or an ELU function (an exponential linear unit function that provides a smooth saturation curve for negative inputs), but there are no limitations on the types.

[0069] According to one embodiment, the hidden layer may be composed of at least one layer, and the more layers there are, the more complex patterns can be learned, resulting in a deep learning structure. For example, when the hidden layer is composed of a first hidden layer and a second hidden layer, the first hidden layer performs MAC operations and activation operations based on the input value of the input layer to output a feature map, and the feature map, which is the result value from the first hidden layer, can become the input value for the second hidden layer. The feature map is an intermediate representation representing features extracted from input data; in image processing, it can represent visual features such as edges, textures, and shapes, and in text processing, it can represent semantic patterns or syntactic structures. The second hidden layer can perform MAC operations and activation operations based on the feature map, which is the result value of the first hidden layer, and through this hierarchical processing, increasingly abstract and high-dimensional features can be extracted.

[0070] According to one embodiment, the output layer may be a layer associated with the result of an operation performed in the hidden layer, and in FIG. 3, it consists of three nodes. The number of nodes in the output layer varies depending on the type of problem to be solved; for binary classification problems, it may be one node, for multi-class classification problems, it may be equal to the number of classes, for regression problems, it may be determined by the dimension of the value to be predicted, and for sequence generation problems, it may be determined by the length of the output sequence. In the output layer, a special activation function such as a softmax function or a sigmoid function is generally used to generate a final predicted value or a probability distribution.

[0071] In one embodiment, the learning model learns syllable (character) patterns that are frequently combined and used in a given corpus to automatically learn the boundaries of compound words and named entities, integrates object information from a first UI source with object information rendered in a browser to create a learning object information file, and uses the learning object information file to generate training data for training a deep learning network. Natural language processing (NLP) technology is utilized in this process, and architectures such as Word Embedding, Recurrent Neural Network (RNN), and Transformer may be applied. Additionally, the learning model receives data from various domains of a support system, standardizes the data from the various domains into an integrated format based on at least one standardization method corresponding to each of the various domains, learns and infers data from a specific domain, determines information to be transmitted for standardization from the data of the specific domain, and performs post-processing on the data from the various domains.

[0072] According to one embodiment, the first UI source includes an XML file, and the training object information file includes an input JSON file for feature learning and an output JSON file that serves as label data during training. This structured data format allows the model to process it easily and is particularly useful for analyzing UI elements of web-based applications. The output JSON file includes a file containing HTML DOM Tree information implemented in compliance with web standards, and the various domains include at least one of a RAN (radio access network), a transport, or a core. A RAN refers to a wireless access network where a mobile device connects to a cellular network; a transport refers to a network layer responsible for data transmission; and a core refers to the central part of a network responsible for central processing and routing. Post-processing may include a correlation function, which plays an important role in analyzing relationships between data collected from various domains and discovering meaningful patterns. Through correlation analysis, the model can identify the causes of anomalies or performance degradation occurring in complex systems and improve prediction accuracy.

[0074] FIG. 4 is a block diagram showing the configuration of an offline store-based inventory-free cosmetic trial order delivery integrated management system according to one embodiment.

[0075] Referring to FIG. 4, the system (400) may include a communication interface (410), a database (412), an input / output interface (414), a processor (420), and a memory (430). The system (400) may be a server device that integrally performs customer experience, ordering, payment, delivery, commission settlement, and store performance management through a zero-inventory cosmetics experience zone installed in a store of another industry. The system (400) may be implemented as a single physical server or as a distributed system in which multiple servers are connected via a network.

[0076] The communication interface (410) may be a means of communication for the system (400) to transmit and receive data with an external device and an external server. The communication interface (410) may include a wired communication module or a wireless communication module and may be connected to an external device via an information and communication network such as the Internet, 5G, or LTE. The communication interface (410) may transmit and receive data with a customer terminal (e.g., a smartphone). For example, the communication interface (410) may receive a connection request transmitted by a customer terminal scanning the QR code of the experience zone and transmit a skin type diagnosis interface, a personalized recommended product list, a delivery tracking URL, etc., to the customer terminal. The communication interface (410) may receive connection URL data including a store identification code (Store_ID), a product identification code (Product_ID), and an experience zone type code (Zone_Type) from the experience zone QR code device. The communication interface (410) may transmit payment request data and receive payment approval results by linking with an external payment server (e.g., a PG server). The communication interface (410) can transmit delivery instruction data to a logistics center system (e.g., WMS) and receive tracking numbers and delivery status information from the logistics center system. The communication interface (410) can transmit automatically generated greeting messages for each store type by linking with a carrier AI assistant (e.g., KT AI Call Assistant, LG U+ AI Phone). The communication interface (410) can receive store review data, reservation response data, posting data, etc. by linking with external platform APIs such as Naver Place API, Naver Reservation API, and community platform API (e.g., Carrot Market). The types of the above external devices and external servers are merely examples and are not limited thereto, and may vary depending on the operating environment of the system (400).

[0077] The database (412) may be a storage medium for the system (400) to permanently store and retrieve data necessary for the operation and management of the experience zone. The database (412) may be implemented as a storage device provided within the system (400) or as a remote database built on an external cloud server. The database (412) may include a store profile database. A store profile database may refer to a database that stores the store identification code, industry classification, location information, commission rate, target customer profile, etc., of each experience zone store. The database (412) may include a product attribute database. A product attribute database may refer to a database that stores the product identification code, product name, category, price, product attribute vector, HSV color value, etc., of each cosmetic product. The database (412) may include a customer skin profile database. A customer skin profile database may refer to a database that stores each customer's skin type, skin tone, concerns, age group, skin profile vector, purchase history, etc. The database (412) may store an industry-channel fit matrix. The industry-channel suitability matrix may refer to a mapping table that pre-defines the suitability between a store industry classification and multiple advertising channels as a numerical value. The database (412) may store a flower variety database. The flower variety database may refer to a database that stores representative HSV color values ​​for each major flower variety handled in the flower shop experience zone. The database (412) may additionally store order data, settlement data, store performance checklist data, and integrated accumulated experience and purchase data.The database (412) is a permanent storage space distinct from the memory (430). Unlike the memory (430), which stores instructions and temporary data required for the operation of the processor (420), the database (412) can store data that is continuously referenced and updated throughout the operation of the system (400). The type of database is merely an example and is not limited thereto; additional databases may be included as the functionality of the system (400) expands.

[0078] The input / output interface (414) may be an interface for an administrator of the system (400) to access the system (400) to control operational settings and monitor operational status. The input / output interface (414) may include input devices such as a keyboard, mouse, and touchscreen, and output devices such as a display and monitor. The input / output interface (414) may be implemented as a web-based administrator console. Through the input / output interface (414), the administrator can register new stores in the store profile database or change the commission rate of existing stores. Through the input / output interface (414), the administrator can set or change the standard values ​​of performance check items (e.g., the standard number of monthly experience photos sent, the standard number of new reviews, etc.). Through the input / output interface (414), the administrator can set operational parameters such as advertising cost execution conditions, pilot grade classification standard values, and settlement cycles. Through the input / output interface (414), the administrator can monitor the performance status by store, sales statistics, advertising cost execution details, and settlement status in real time. The administrator can adjust the score of the industry-channel fit matrix or change the calculation criteria for the region correction coefficient through the input / output interface (414). The administrator can update the value of the product attribute vector or register a new product in the product attribute database through the input / output interface (414). The above administrator functions are merely examples and are not limited thereto, and additional management functions may be included according to the operating policy of the system (400).

[0079] The processor (420) may be a processing unit that controls the overall operation of the system (400). The processor (420) may include computing units such as a CPU (Central Processing Unit), GPU (Graphics Processing Unit), AP (Application Processor), and NPU (Neural Processing Unit). By loading and executing instructions stored in memory (430), the processor (420) can perform operations such as QR code-based store recognition, skin data collection and profile vector generation, cosine similarity-based customized recommendation product list generation, payment processing and automatic transmission of delivery instructions, automatic settlement of commissions per store, automatic determination of store performance check items and conditional execution of advertising costs, automatic selection of optimal marketing channels per store, and generation of industry-specific cross-selling recommendation lists. The processor (420) can transmit and receive data with external devices and external servers through a communication interface (410), store or query data in a database (412), receive setting inputs from an administrator through an input / output interface (414), and output operational status. The processor (420) may be implemented as a single processor or as a structure in which multiple processors operate in parallel.

[0080] Memory (430) may be a storage device that stores instructions to be executed by the processor (420) and temporary data required for operations. Memory (430) may include RAM (Random Access Memory), ROM (Read Only Memory), flash memory, etc. Instructions stored in memory (430) can control the system (400) to perform each operation described in this specification when executed by the processor (420). Memory (430) is a component distinct from the database (412), wherein memory (430) stores instructions and temporary data for the immediate execution of operations by the processor (420), and the database (412) can store permanent data continuously required for the operation of the system (400).

[0082] FIG. 5 illustrates an integrated management method for ordering and delivering cosmetics for trial use without inventory, according to one embodiment.

[0083] The operations described through FIG. 5 can be implemented based on instructions that can be stored in memory (e.g., memory (420) of FIG. 4). The order of each operation of FIG. 5 may be changed, some operations may be omitted, and some operations may be performed simultaneously.

[0084] According to one embodiment, when a customer terminal scans a QR code of an experience zone installed in a store of another industry to access it, the system loads a store profile based on a store identification code included in the QR code, receives skin data from the customer terminal to create a customer skin profile, performs a similarity analysis between the customer skin profile and product attribute information, calculates a suitability score for each product by reflecting the industry characteristics of the store, generates a customized recommended product list based on the suitability score and transmits it to the customer terminal, processes payment for the product selected by the customer, transmits a delivery instruction to a logistics center upon completion of payment, calculates a commission per store based on the store identification code to perform automatic settlement, and controls the system to integrate and accumulate data generated during the experience and purchase process.

[0085] According to one embodiment, when a customer terminal scans a QR code of an experience zone installed in a store of a different industry to access it, the system (400) loads a store profile containing information on the industry and commission rate of the store based on a store identification code encoded in the QR code, and transmits a skin type diagnosis interface to the customer terminal to receive skin data including skin type, skin tone, and skin concerns from the customer. The system (400) converts the received skin data into a multidimensional vector to generate a customer skin profile vector, performs a similarity analysis between the customer skin profile vector and the product attribute vector, calculates a skin suitability score for each product by reflecting the industry affinity between the store industry and the product category corresponding to the store identification code, generates a customized recommended product list based on the skin suitability score and transmits it to the customer terminal, processes payment for the product selected by the customer through an external payment server, automatically transmits a delivery instruction to a logistics center system upon completion of payment, automatically calculates a commission for each store based on the store identification code to perform automatic settlement to the store owner's account, and controls the system to integrate and accumulate data generated during the experience and purchase process by customer, store, and product.

[0086] In operation 510, the system (400) can load a store profile based on a store identification code through a QR code scan. Specifically, when a customer terminal scans a QR code of an experience zone installed in a store of a different industry to access it, the system (400) can receive a connection URL containing a store identification code (Store_ID), a product identification code (Product_ID), and an experience zone type code (Zone_Type) encoded in the QR code. Based on the received store identification code, the system (400) can query a store profile database to load a store profile containing the industry classification, location information, commission rate, and target customer base information of the store. Here, stores of a different industry may refer to existing stores that are not dedicated to cosmetics, such as cafes, beauty salons, flower shops, nail salons, and bookstores, and each store may only have samples for experience without inventory for sale. The system (400) can distinguish orders generated in each experience zone by store using the store identification code, and thereby secure basic data for commission settlement and store performance management in subsequent stages.

[0087] In operation 520, the system (400) can receive customer skin data to generate a skin profile vector and generate a customized recommended product list. Specifically, the system (400) can transmit a skin type diagnosis survey interface to a customer terminal and receive information on the skin type (dry, oily, combination, sensitive), skin tone (warm tone, cool tone, neutral), major concerns (pores, blemishes, wrinkles, dryness, skin trouble, whitening, etc.) and age group entered by the customer. The system (400) can generate a customer skin profile vector by quantifying each item of the received skin data, calculating a seasonal correction value reflecting dryness and UV index based on the current month, and converting it into a multidimensional vector including a skin type value, a skin tone value, a weight for each concern item, an age group value, and a seasonal correction value. The system (400) can calculate the cosine similarity between the generated customer skin profile vector and a product attribute vector pre-set for each product. The system (400) can calculate a skin suitability score for each product by summing the value obtained by applying a first weight to the cosine similarity, the value obtained by applying a second weight to the seasonal correction value for each product according to the current season, the value obtained by applying a third weight to the industry affinity between the store industry and the product category corresponding to the store identification code, and the value obtained by applying a fourth weight to the purchase history bonus based on the customer's past purchase history, and normalizing the sum of the first to fourth weights to be 1.0. Here, the industry affinity may be a numerical value representing the degree of association between the store's industry characteristics and the product category, and, for example, may be applied in such a way that the weight of lip and cheek category products is increased for customers experiencing at a flower shop, and the weight of skincare products is increased at a yoga or Pilates shop. The system (400) can select a predetermined number of top products in order of high skin suitability scores, generate a customized recommended product list along with the skin suitability scores of each product, and transmit it to the customer terminal.

[0088] In operation 530, the system (400) can process the payment and automatically transmit a delivery instruction to the logistics center upon completion of the payment. Specifically, when a customer selects a product they wish to purchase from the personalized recommended product list or the entire product list, the system (400) can create a shopping cart session containing price information, shipping fees, and store identification information for the selected product. When the customer enters delivery address information and requests payment, the system (400) can transmit payment request data to an external payment server (PG) and receive an approval result from the payment server regarding the completion of payment processing, thereby updating the order status to payment completed. Immediately upon completion of payment, the system (400) can automatically transmit delivery instruction data to the logistics center system, including an order number, product code, quantity, delivery address, and recipient information. When the system (400) receives a tracking number from the logistics center system after the shipment processing, it can send a notification containing a delivery tracking URL to the customer terminal. The system (400) can eliminate the inventory burden of the store and implement real-time order and delivery processing through a structure in which it does not hold sales inventory at the experience zone store and delivers directly to customers from the headquarters logistics center.

[0089] In operation 540, the system (400) can automatically settle commissions per store and accumulate integrated experience and purchase data. Specifically, the system (400) can look up a pre-set commission rate for the corresponding store based on the store identification code included in the order data, and automatically calculate the commission per store by multiplying the order amount by the said commission rate. The system (400) can deposit the calculated commission into a settlement waiting fund for each store and execute an automatic transfer to the account of the corresponding store owner when the pre-set settlement cycle arrives. The system (400) can set and operate the said settlement cycle on a weekly or monthly basis, and can eliminate separate manual settlement tasks by automating the profit distribution between the store owner and the headquarters. The system (400) can accumulate data generated during the entire process, including experience store information, experience product information, customer skin profile, recommendation results, purchase status, purchased products, and payment amounts, in an integrated database by customer, store, and product. The system (400) can use the accumulated data as basic data for future marketing optimization analysis, store performance evaluation, and product planning feedback.

[0090] According to one embodiment, when a customer terminal scans a QR code of an experience zone installed in a store of a different industry to access it, the system (400) loads a store profile containing information on the industry and commission rate of the store based on a store identification code encoded in the QR code, and transmits a skin type diagnosis interface to the customer terminal to receive skin data including skin type, skin tone, and skin concerns from the customer. The system (400) converts the received skin data into a multidimensional vector to generate a customer skin profile vector, performs a similarity analysis between the customer skin profile vector and the product attribute vector, calculates a skin suitability score for each product by reflecting the industry affinity between the store industry and the product category corresponding to the store identification code, generates a customized recommended product list based on the skin suitability score and transmits it to the customer terminal, processes payment for the product selected by the customer through an external payment server, automatically transmits a delivery instruction to a logistics center system upon completion of payment, automatically calculates a commission for each store based on the store identification code to perform automatic settlement to the store owner's account, and controls the system to integrate and accumulate data generated during the experience and purchase process by customer, store, and product.

[0091] Here, "stores in other business sectors" may refer to existing offline stores that do not primarily engage in the sale of cosmetics. Stores in other business sectors may include cafes, beauty salons, flower shops, nail salons, bookstores, yoga or Pilates studios, fashion shops, etc., and the above examples are merely illustrative and are not limited thereto. Each store in other business sectors may not stock inventory for sale and may only display samples for trial. An "experience zone" may refer to a designated area within the aforementioned store in other business sectors where cosmetic sample products and QR code devices are installed to allow customers to directly test and use the products.

[0092] A QR code may be encoded with a Store ID, a Product ID, and an Experience Zone Type. The Store ID is a code used to uniquely identify each experience zone store, and can be used as reference data for the system (400) to automatically branch the settlement of commissions per store, store performance management, and selection of marketing channels within the same order processing flow. The Product ID may be a code used to identify individual products or product groups placed in the experience zone. The Experience Zone Type may be a code indicating the classification of the experience zone according to the industry characteristics of the store where the experience zone is installed.

[0093] The system (400) can query a store profile database based on a store identification code when a customer terminal scans the QR code. A store profile may refer to a set of unique attribute information of a store, including the store's industry classification, location information, commission rate, and target customer profile. The system (400) can utilize the industry classification information included in the store profile to calculate industry affinity in a subsequent customized recommendation process, and can utilize the commission rate information in an automatic settlement process after payment is completed. The system (400) can manage the entire processing flow, from order reception to payment processing, delivery instructions, commission settlement, and data accumulation, by attributing it to the corresponding store based on a single store identification code.

[0094] The system (400) can collect skin data of a customer by transmitting a skin type diagnostic interface to a customer terminal. The skin type diagnostic interface may refer to a survey-type user interface configured to allow a customer to select and input their skin type, skin tone, and skin concerns. Skin type may include classifications such as dry, oily, combination, and sensitive; skin tone may include classifications such as warm tone, cool tone, and neutral; and skin concerns may include items such as pores, blemishes, wrinkles, dryness, skin troubles, and whitening. The above classification items are merely examples and are not limited thereto, and may be changed according to the product line configuration or operational policy.

[0095] The system (400) can generate a customer skin profile vector by converting the received skin data into a multidimensional vector. Conversion into a multidimensional vector means mapping qualitative data such as skin type, skin tone, concern items, and age group entered by the customer to numerical values ​​between 0 and 1, and adding a seasonal correction value calculated based on the current month to form a single numerical vector. The seasonal correction value is a correction coefficient to reflect changes in skin condition according to seasonal characteristics at the current time, and can be calculated based on seasonal environmental factors such as a dryness index or a UV index. The system (400) can quantitatively express the individual skin characteristics of the customer using the customer skin profile vector and can utilize this as input data for a product recommendation algorithm.

[0096] The system (400) can calculate a skin suitability score for each product by performing a similarity analysis between a customer skin profile vector and a product attribute vector. A product attribute vector may refer to attribute information set in advance for each product expressed as a numeric vector of the same dimension as the customer skin profile vector. The similarity analysis may be performed based on cosine similarity, which is a mathematical technique for measuring directional similarity between two vectors, and can represent the degree of matching between the customer's skin characteristics and the product attributes as a value between 0 and 1.

[0097] The system (400) may reflect the industry affinity between the store industry and the product category in addition to the similarity analysis results. Industry affinity may refer to a numerical value representing the degree of association between the industry characteristics of the store where the experience zone is installed and a specific product category. The system (400) may apply differential recommendation weights to specific product categories depending on the store's industry. For example, in the case of a flower shop, the weight of lip and cheek category products that can color-match with flowers may be increased, and in the case of a yoga or Pilates shop, the weight of skincare category products suitable for use before and after exercise may be increased. By reflecting the industry affinity in the product recommendation, the system (400) may prioritize recommending products that match the atmosphere and context of the store where the customer visited the experience zone. The association between the industry and the product category is merely an example and is not limited thereto, and may be updated as the store industry is added or the product lineup is changed.

[0098] The system (400) can generate a customized recommended product list based on skin suitability scores and transmit it to a customer terminal. The customized recommended product list may refer to a recommendation list composed of a predetermined number of products selected in order of highest skin suitability scores calculated above, along with the suitability scores of each product. When a customer selects a product they wish to purchase from the customized recommended product list or the entire product list, the system (400) can generate a shopping cart session including price information, shipping costs, and store identification information for the selected product.

[0099] The system (400) can process payment for a product selected by a customer through an external payment server. An external payment server may refer to a Payment Gateway (PG) server that handles credit card payments or simple payments. When a customer enters delivery address information and requests payment, the system (400) can transmit payment request data to the payment server and update the order status by receiving an approval result from the payment server. The system (400) can automatically transmit a delivery instruction to the logistics center system upon completion of payment. A delivery instruction may refer to transmitting data including an order number, product code, quantity, delivery address, and recipient information to the logistics center system to initiate the product dispatch and delivery procedure. Through a structure in which the system (400) does not hold sales inventory at the experience zone store and delivers directly to the customer from the headquarters logistics center, the system (400) can implement real-time order and delivery processing for customers while eliminating the store's inventory burden and expiration date management risks.

[0100] The system (400) can automatically calculate a commission per store for an order amount based on a store identification code. The system (400) can calculate a commission to be paid to the store owner by querying a commission rate pre-set in the store profile based on the store identification code included in the order data and multiplying the order amount by the said commission rate. Automatic settlement may mean accumulating the calculated commission into a settlement reserve fund per store and executing an automatic transfer to the store owner's account when a pre-set settlement cycle arrives. The settlement cycle may be set on a weekly or monthly basis and may vary depending on the setting. Through the above automatic settlement, the system (400) can automate the distribution of profits between the headquarters and the store owner and eliminate separate manual settlement tasks.

[0101] The system (400) can integrate and accumulate data generated during the experience and purchase process by customer, store, and product. Integrated accumulation may mean classifying data including experience store information, experience product information, customer skin profile, recommendation results, purchase status, purchased product, and payment amount into three standard axes—customer, store, and product—and storing it in an integrated database. The system (400) can utilize the accumulated data as basic data for marketing optimization analysis, store performance evaluation, and product planning feedback. By integrating and collecting data generated from all experience zones in real-time on the headquarters server, the system (400) can solve the problem of conventional technology where integrated analysis at the headquarters level was impossible because data was dispersed by store.

[0102] According to one embodiment, the system (400) loads an access URL containing a store identification code, a product identification code, and an experience zone type code encoded in a QR code into a web browser of a customer terminal, and can load the store's industry classification, location information, commission rate, and target customer base information by querying a store profile database based on the store identification code. The system (400) can generate a customer skin profile vector by quantifying the skin type, skin tone, major concerns, and age group information received from the customer, calculating a seasonal correction value based on the current month, and converting it into a multidimensional vector including a skin type value, a skin tone value, a weight for each concern item, an age group value, and a seasonal correction value. The system (400) calculates a cosine similarity between a customer skin profile vector and a product attribute vector pre-set for each product, and calculates a product-specific skin suitability score by adding a value with a first weight applied to the cosine similarity, a value with a second weight applied to a product-specific seasonal correction value according to the current season, a value with a third weight applied to an industry affinity between a store industry and a product category, and a value with a fourth weight applied to a purchase history bonus based on the customer's past purchase history, and normalizes the sum of the first to fourth weights to be 1.0, and can generate a customized recommended product list along with the skin suitability scores of each product by selecting a predetermined number of top products in order of high skin suitability scores.

[0103] The access URL may refer to a web address that is automatically loaded into a web browser when a customer terminal scans a QR code. The access URL may include a store identification code, a product identification code, and an experience zone type code as parameters, and the system (400) can parse these parameters to provide a dedicated order page corresponding to the store to the customer terminal. By using the same order page structure while dynamically loading store profile information according to the store identification code, the system (400) can provide an order environment specialized for each store without building a separate order page for each store.

[0104] The system (400) can load the industry classification, location information, commission rate, and target customer base information of the store by querying the store profile database. Target customer base information may refer to data that summarizes the characteristics of customers who primarily use the store, such as age group, gender, and consumption patterns. The system (400) can use the above target customer base information as reference data for product recommendation and marketing channel selection in subsequent steps.

[0105] The system (400) can quantify each item of skin data received from a customer. Quantification may mean mapping qualitative classification data to a continuous numerical value between 0 and 1. For example, in the case of skin type, dryness can be mapped to 0.0, combination to 0.5, sensitivity to 0.7, oiliness to 1.0, etc., and in the case of skin tone, warm tone can be mapped to 0.0, neutral to 0.5, cool tone to 1.0, etc. In the case of concern items, a weight between 0 and 1 can be assigned to each item selected by the customer. In the case of age group, 20s can be mapped to 0.2, 30s to 0.4, 40s to 0.6, 50s and older to 0.8, etc. The above numerical mapping method is merely an example and is not limited thereto, and may vary depending on the product lineup configuration or operational policy.

[0106] The system (400) can calculate a seasonal correction value based on the current month. The seasonal correction value may refer to a value that quantifies the effect of the seasonal environment at the current time on the skin condition. For example, since the dryness index may increase during the winter, the suitability of moisturizing products may be corrected upward, and since the UV index may increase during the summer, the suitability of UV protection products may be corrected upward. By including the seasonal correction value as a component of the skin profile vector, the system (400) can ensure that different products are recommended depending on the season, even for customers of the same skin type.

[0107] The system (400) may sum multiple weighting factors when calculating the skin suitability score. The cosine similarity to which the first weight is applied may represent the basic degree of matching between the customer's skin characteristics and the product attributes. The seasonal correction value to which the second weight is applied may reflect changes in demand for a specific product category according to the current season. The industry affinity to which the third weight is applied may be a value for adjusting the recommendation weight for a specific product category according to the industry characteristics of the store where the experience zone is installed. The purchase history bonus to which the fourth weight is applied may be a value that grants additional points for products in the same or similar categories as the product categories purchased by the customer in the past. The system (400) can control the relative contribution of each factor within a certain range by normalizing the sum of the first to fourth weights to be 1.0. Normalization may mean adjusting each weight so that the sum of multiple weights becomes a constant value, and the system (400) can prevent a specific factor from having an excessive influence on the suitability score through normalization. The specific values ​​of each of the above weights may be adjusted according to operational policies or business strategies.

[0108] The system (400) can generate a customized recommended product list by selecting a predetermined number of top products in order of highest skin suitability scores. The predetermined number may be, for example, 3 to 5, but is not limited to this and may vary depending on the settings. The system (400) can support customers in intuitively recognizing the degree of suitability for their skin and selecting a product by displaying the skin suitability score for each recommended product.

[0110] FIG. 6 illustrates a method for automatic evaluation of store performance and conditional execution of advertising costs according to one embodiment.

[0111] The operations described through FIG. 6 can be implemented based on instructions that can be stored in memory (e.g., memory (420) of FIG. 4). The order of each operation of FIG. 6 may be changed, some operations may be omitted, and some operations may be performed simultaneously.

[0112] In operation 610, the system (400) can automatically determine whether a plurality of performance check items have been achieved for each experience zone store. Specifically, as a first check item, the system (400) can determine whether the number of experience photos sent by the store to the headquarters server via the store management app within a specified period is greater than or equal to a preset standard (e.g., 5 photos per month). As a second check item, the system (400) can automatically collect the number of new reviews registered for the store within a specified period by linking with the Naver Place API and determine whether the number of new reviews is greater than or equal to a preset standard. As a third check item, the system (400) can check for the existence of community posts for the store based on a URL submitted via the community platform API or the store app and determine whether the number of posts within a specified period is greater than or equal to a preset standard. As a fourth check item, the system (400) checks the inspection score based on the results of the visit inspection by the head office sales representative and the negative review ratio calculated through sentiment analysis of customer reviews, and can determine whether the inspection score is above a preset standard and the negative review ratio is below a preset standard. As a fifth check item, the system (400) automatically collects the number of unanswered reservations of the store through the Naver Reservation API and can determine whether the number of unanswered reservations is 0. As a sixth check item, the system (400) checks customer SNS authentication photos uploaded through the store app and records of goods distribution, and can determine whether the number of authentications within a predetermined period is above a preset standard.

[0113] In operation 620, the system (400) may sum the number of passed items and approve the payment of advertising fees if the criteria are met, and send a warning if they are not met. Specifically, the system (400) may calculate the total number of passed items by aggregating the pass status of each of the multiple performance check items determined in operation 610. If the total number of passed items is greater than or equal to a preset passing standard (e.g., 5 out of 6 items), the system (400) may automatically approve the payment of advertising fees for the current month for the store. If the total number of passed items is less than the preset passing standard, the system (400) may withhold the payment of advertising fees for the current month for the store and automatically send a warning notification to the store owner containing specific details of the items that are not met. By providing the store owner with an opportunity to recognize and improve the items that are not met through the warning notification, the system (400) may support the maintenance of store operation quality without the need for additional personnel input.

[0114] In operation 630, the system (400) can automatically classify stores into multiple grades based on the number of monthly QR sales at the end of the pilot period. Specifically, when a predetermined pilot period (e.g., 3 months) ends, the system (400) can aggregate the number of monthly QR sales that occurred at each store during that period. The system (400) can automatically classify stores with a monthly QR sales count greater than or equal to a first threshold value into a higher grade, stores with a monthly QR sales count less than the first threshold value but greater than or equal to a second threshold value into a middle grade, and stores with a monthly QR sales count less than the second threshold value into a lower grade. By performing the grade classification based on actual sales data collected through QR codes, the system (400) can automate store evaluation based on objective performance criteria.

[0115] In operation 640, the system (400) may determine differential measures, including contract extension, conversion to self-funded operation, or termination, based on the classified grade. Specifically, for stores classified as a higher grade in operation 630, the system (400) may designate them as subjects for negotiation regarding the extension of the experience zone operation contract and continue to provide advertising cost support from the headquarters to the stores. For stores classified as a middle grade, the system (400) may terminate the advertising cost support from the headquarters and designate them as subjects for conversion to self-funded operation. For stores classified as a lower grade, the system (400) may designate them as subjects for pilot termination and initiate a procedure to terminate the experience zone operation contract. The system (400) may automatically notify the store owners of the results of the decision on the differential measures, thereby enabling efficient management of the experience zone network based on performance.

[0116] According to one embodiment, the system (400) automatically determines whether each experience zone store has achieved a plurality of performance check items, and can determine whether the number of experience photos sent by the store to the headquarters server within a predetermined period is above a preset standard, whether the number of new reviews of the store collected through the Naver Place API is above a preset standard, whether the number of community platform postings within a predetermined period is above a preset standard, whether the headquarters inspection score and the negative review ratio each meet a preset standard, whether the number of unanswered cases for reservation customers is 0, and whether the performance of customer SNS authentication postings and goods distribution is above a preset standard. The system (400) can automatically approve the payment of advertising fees for the current month for the store if the sum of the number of passed items among the plurality of performance check items is above a preset passing standard, and can automatically send a warning notification including the items that are not met to the store owner if the sum is below the passing standard. The system (400) can automatically classify the store into multiple grades based on the number of monthly QR sales per store at the end of a predetermined pilot period, and determine differential measures including contract extension, conversion to store self-funding, or termination of the pilot based on the classified grade.

[0117] Performance check items may refer to multiple evaluation criteria defined in advance to quantitatively evaluate the operational activities of each experience zone store. The system (400) can automatically determine whether each of the above performance check items has been achieved through external API integration, data collection via a store management app, and input of headquarters inspection data. By combining multiple performance check items to evaluate the operational quality of the store in a multifaceted way, the system (400) can perform more balanced store management compared to an evaluation method that relies on a single indicator.

[0118] The system (400) can determine the number of experience photos transmitted by the store to the headquarters server within a specified period as a first check item. Experience photos may refer to photos taken by the store operator through a store management app to capture the operational status of the experience zone and uploaded to the headquarters server. The system (400) accumulates a reception count whenever a photo is uploaded from the store management app, and determines whether the accumulated count within a specified period is above a preset standard, processing it as either passing or failing. The system (400) can indirectly verify whether the store is actually operating the experience zone through the collection of the experience photos.

[0119] The system (400) can determine the number of new reviews for the store collected through the Naver Place API as a second check item. The system (400) can automatically collect the number of new reviews registered for each store within a specified period by linking with the Naver Place API. The system (400) can determine that the item passes if the number of new reviews collected is greater than or equal to a preset standard. Checking the number of new reviews may be intended to evaluate whether the store is encouraging customers to write reviews.

[0120] The system (400) can determine the number of community platform postings as a third check item. A community platform may refer to a local-based second-hand trading and community service, and may include Danggeun Market, etc. The system (400) can determine the number of postings by checking whether a post exists through a community platform API or by verifying the validity of a post URL submitted by a store operator through a store management app.

[0121] The system (400) can determine the cleanliness and service status of the experience zone as a fourth check item. The system (400) can check the inspection score entered as a result of the visit inspection by the head office sales representative and the negative review ratio calculated through sentiment analysis of customer reviews. Sentiment analysis may refer to a natural language processing technique that automatically classifies the positive or negative sentiment of review text written by customers. The system (400) can determine that the item is passed if the inspection score is above a preset standard and the negative review ratio is below a preset standard. By making a determination by combining the quantitative inspection score and the results of the sentiment analysis of customer reviews, the system (400) can evaluate the operational quality of the store from a complex perspective.

[0122] The system (400) can determine the number of unanswered cases regarding reservation customers as a fifth check item. The system (400) can automatically collect the number of unanswered cases among reservation customers for the store through the Naver Reservation API. The number of unanswered cases may refer to the number of cases where a customer has completed a reservation but the store has not confirmed or responded to them. The system (400) can determine that the item is passed if the number of unanswered cases is 0. Through the above determination, the system (400) can verify whether the store is faithfully fulfilling its obligation to respond to reservation customers.

[0123] The system (400) can determine customer SNS authentication postings and merchandise distribution performance as the sixth check item. Customer SNS authentication postings may refer to a customer who visited the experience zone writing a post related to the experience on their SNS, and the store operator uploading it as an authentication photo through the store management app. Merchandise distribution may refer to the performance of the store operator distributing promotional materials provided by the headquarters to customers. The system (400) can determine that the item is passed if the number of authentications and distribution records within a specified period exceed a preset standard.

[0124] The system (400) can calculate the total number of passed items by aggregating the pass status of each of the aforementioned multiple performance check items. If the total number of passed items is greater than or equal to a preset pass standard, the system (400) can automatically approve the payment of advertising fees for the current month for the store. The preset pass standard may be, for example, 5 out of 6 items, but is not limited to this and may be changed according to the operating policy. If the total number of passed items is less than the preset pass standard, the system (400) can withhold the payment of advertising fees for the current month for the store and automatically send a warning notification to the store owner containing specific details of the items that failed to meet the standard. By including both the current value and the standard value for each failed item in the warning notification, the system (400) can help the store owner recognize which specific items need to be improved. By linking the payment of advertising fees to the achievement level of the performance check items, the system (400) can induce the store operator to voluntarily maintain operational quality.

[0125] The system (400) can automatically classify the store into multiple grades based on the number of monthly QR sales per store at the end of a predetermined pilot period. The pilot period may refer to a pre-set trial operation period to verify the initial operational effectiveness of the experience zone, and may be, for example, three months. The number of monthly QR sales may refer to the number of monthly completed orders generated through the QR code of the experience zone installed in the store. The system (400) may classify stores with a monthly QR sales count greater than or equal to a first threshold value into a higher grade, stores with a count less than the first threshold value but greater than or equal to a second threshold value into a middle grade, and stores with a count less than the second threshold value into a lower grade. The specific numerical values ​​of the first threshold value and the second threshold value may be set according to the operation policy.

[0126] The system (400) can determine differential measures based on the classified grade. For stores classified as a high grade, the system (400) may designate them as subjects for negotiation on extending the experience zone operation contract and continue to provide advertising cost support from the headquarters. For stores classified as a middle grade, the system (400) may terminate the advertising cost support from the headquarters and designate them as subjects for transitioning to operation at the store's own expense. For stores classified as a low grade, the system (400) may designate them as subjects for pilot termination and initiate the procedure for terminating the experience zone operation contract. The system (400) may automatically notify the relevant store owner of the results of the determination of the above differential measures. By automating grade classification and differential measures based on actual sales data collected via QR code, the system (400) can efficiently manage the experience zone network according to objective performance criteria.

[0127] According to one embodiment, the system (400) can identify the business classification of a store corresponding to a store identification code, and if the business of the store is a flower shop, extract the representative HSV color value of the flower variety handled by the store from a flower variety database and extract the HSV color value of each product from a product database belonging to the lip and cheek category. The system (400) calculates the color difference between the representative HSV color value of the flower variety and the HSV color value of each product by squaring the difference values ​​of hue, saturation, and brightness, summing them, and taking the square root to calculate the color difference value, and can select products with a color difference value below a preset threshold as flower color matching recommended products. The system (400) can determine whether the current time corresponds to the gift season, and if it corresponds to the gift season, increase the recommendation weight for the matching recommended products, and integrate the flower color matching recommended products with a customized recommended product list based on skin suitability scores to create an industry-specific cross-selling recommendation list and transmit it to a customer terminal.

[0128] An HSV color value can refer to a value in a color space model that expresses color using three components: hue, saturation, and value. Hue can be an angle value indicating the type of color, saturation can be a value indicating the vividness of the color, and value can be a value indicating the brightness of the color. By using the HSV color space rather than other color spaces such as RGB, the system (400) can perform color matching based on a standard that is closer to how humans perceive color.

[0129] The system (400) can extract a pre-registered representative HSV color value for each flower variety from a flower variety database. The flower variety database may refer to a database that stores representative color information for each major flower variety handled by a flower shop where an experience zone is installed. A representative HSV color value may refer to a value that expresses the most common color of the flower variety as a number in the HSV color space. The system (400) can also extract a pre-registered HSV color value for each product from a product database belonging to the lip and cheek category. The lip and cheek category may refer to a cosmetics category where color is a major attribute of the product, such as lipstick, lip tint, and cheek blush.

[0130] The system (400) can calculate the color difference between the representative HSV color value of a flower variety and the HSV color value of each product. Color difference may refer to a value that quantifies the visual difference between two colors. The system (400) can calculate the color difference value by squaring the difference values ​​of hue, saturation, and brightness, summing them, and taking the square root. The system (400) can select products whose color difference value is below a preset threshold as recommended products for flower color matching. The threshold may refer to an upper limit standard value of color difference for determining that two colors are visually similar. The system (400) can control the strictness of the matching criteria by adjusting the threshold; if the threshold is set low, only products with more similar colors are recommended, and if it is set high, a wider range of products may be included in the recommendation target. The specific numerical value of the threshold may be set according to the operation policy.

[0131] The system (400) can determine whether the current time corresponds to a gift season. A gift season may refer to a period when the demand for gifts increases, such as Valentine's Day, Parents' Day, or Christmas. The system (400) can activate a gift season flag if the current month corresponds to a predefined gift season month. When the gift season flag is activated, the system (400) can increase the recommendation weight for flower color matching recommended products. Through the increase in the weight, the system (400) can induce the composition of gift sets combining bouquets and lip and cheek products, and can maximize cross-selling effects by utilizing the characteristics of the store's industry. The definition of the gift season month is merely an example and is not limited thereto, and may be changed or added according to the operating policy.

[0132] The system (400) can generate an industry-specific cross-selling recommendation list by integrating flower color matching recommendation products with a customized recommendation product list based on skin suitability scores. An industry-specific cross-selling recommendation list may refer to a recommendation list composed by combining recommendations based on the customer's skin characteristics and recommendations based on the unique characteristics of the store industry. The system (400) can determine the final recommendation ranking by reflecting the skin suitability score and the color matching score, respectively, during the integration process. Through the industry-specific cross-selling recommendations, the system (400) can provide cosmetic recommendations that match the context in which the customer visited the flower shop, and can create new purchasing motivations that link flower purchases with cosmetic purchases. The cross-selling algorithm for the flower shop industry is an example, and a cross-selling algorithm based on the unique characteristics of the respective industry can be applied to other industries as well.

[0134] Figure 7 shows the information processing flow of an industry-specific cross-selling color matching algorithm according to one embodiment.

[0135] Referring to FIG. 7, the system (400) receives color data from a flower variety DB (710) and a lip and cheek product DB (711), and can generate a final recommended product (750) through a color difference calculation module (720), a recommended product selection module (730), and a gift season weighting application module (740). The algorithm illustrated in FIG. 7 is exemplified by an experience zone in the flower shop industry and may be intended to perform cross-selling recommendations by matching the unique characteristics of the store industry with the attributes of cosmetic products based on color.

[0136] The flower variety DB (710) may be a database that stores representative HSV color values ​​for major flower varieties handled in the flower shop experience zone. An HSV color value may refer to a value in a color space model that expresses color using three components: hue, saturation, and value. In the flower variety DB (710), representative HSV color values ​​for roses, tulips, and carnations may be stored for each variety. A representative HSV color value may refer to a value that expresses the most common and representative color of the flower variety as a number in the HSV color space. A single flower variety may have multiple representative colors, and in this case, multiple HSV color values ​​may be registered for each variety. The above types of flower varieties are merely examples and are not limited thereto, and may be updated according to changes in the flower varieties handled in the store.

[0137] The lip and cheek product DB (711) may be a database that stores the HSV color values ​​of each cosmetic product belonging to the lip and cheek category. The lip and cheek category may refer to cosmetic categories where color is a major attribute of the product, such as lipstick, lip tint, lip gloss, and cheek blush. The lip and cheek product DB (711) may store the product identification code, product name, color name, and the HSV color value of the corresponding color for each product. The system (400) may extract the respective HSV color values ​​from the flower variety DB (710) and the lip and cheek product DB (711) and transmit them as input data to the color difference calculation module (720).

[0138] The color difference calculation module (720) calculates the color difference between the representative HSV color value of the flower variety and the HSV color value of the lip and cheek product ( It may be a module that calculates E (Delta E). Color difference may refer to a numerical value that quantifies the visual difference between two colors. The system (400) can calculate the color difference between the representative HSV color value of each flower variety registered in the flower variety DB (710) and the HSV color value of each product registered in the lip and cheek product DB (711) through the color difference calculation module (720). The system (400) can calculate the color difference value by squaring the difference value of the hue (H) component, the difference value of the saturation (S) component, and the difference value of the lightness (V) component, summing them, and taking the square root. The system (400) can calculate the color difference between each flower variety and all products registered in the lip and cheek product DB (711), and can repeat the above calculation for multiple flower varieties. Through the above color difference calculation, the system (400) can compare the visual similarity between the color of the flower and the color of the cosmetic as an objective numerical value.

[0139] The recommended product selection module (730) may be a module that selects products whose color difference value calculated by the color difference calculation module (720) is below a preset threshold as matching recommended products. The threshold may refer to an upper limit standard value of color difference for determining whether two colors are visually similar. Through the recommended product selection module (730), the system (400) can select lip and cheek products whose color difference from the representative color of a specific flower variety is below the threshold as color matching candidate products for that flower variety. By excluding products whose color difference exceeds the threshold from the matching candidates, the system (400) can primarily classify only products that are visually harmonized with the color of the flower. The system (400) can control the strictness of the matching criteria by adjusting the threshold. If the threshold is set low, only products that are more similar to the flower color are selected, and if it is set high, products with a wider color range may be included in the selection target. The specific numerical value of the threshold may be set according to the operation policy.

[0140] The gift season weighting application module (740) may be a module that applies weights to matching recommended products initially selected by the recommendation product selection module (730) based on whether it is currently the gift season. The gift season may refer to a period when the demand for gifts increases, such as Valentine's Day, Parents' Day, or Christmas. The system (400) can determine whether the current month corresponds to a predefined gift season month (e.g., February, May, December) through the gift season weighting application module (740). If the current time corresponds to the gift season, the system (400) can increase the recommendation weight for the initially selected matching recommended products. Through the increase in weight, the system (400) can raise the recommendation ranking of products suitable for composing gift sets combining flowers and lip and cheek products during the gift season. If the current time does not correspond to the gift season, the system (400) can maintain the results of the initially selected products as they are without increasing the weight. The above definition of the gift season month is for illustrative purposes only and is not limited thereto; it may be changed or added according to operational policy.

[0141] The system (400) can generate a final recommended product (750) based on the results of applying gift season weights. The final recommended product (750) may refer to a list of matching recommended products sorted according to the final ranking reflecting color matching scores and gift season weights. The system (400) can integrate the final recommended product (750) with a list of customized recommended products generated based on skin suitability scores to form an industry-specific cross-selling recommendation list and transmit it to a customer terminal. Through the cross-selling color matching algorithm, the system (400) can provide cosmetic recommendations that match the context of the customer's visit to the flower shop and create a new purchasing motivation that links flower purchases with cosmetic purchases. By recommending lip and cheek products that match the bouquet to the customer, the system (400) can provide a differentiated cross-selling experience utilizing the unique characteristics of the store industry.

[0142] The cross-selling color matching algorithm illustrated in FIG. 7 is exemplified by the flower shop industry, but a cross-selling algorithm based on the unique characteristics of the industry can be applied in a similar manner to experience zones of other industries. Each of the above modules can be implemented as a software module executed by the processor (420) of the system (400), and is not limited to an independent hardware device.

[0143] According to one embodiment, the system (400) can input data on the industry classification of the store corresponding to the store identification code, the regional characteristics of the store location, and the target customer base, and retrieve a basic suitability score by industry from an industry-channel suitability matrix for each of the multiple advertising channels. The system (400) calculates a regional correction coefficient based on the regional characteristics of the store location, and can increase the correction coefficient for a regional-based community channel if the store is located in a densely populated residential area, increase the correction coefficient for a regional search-based channel if it is located in a densely populated commercial area, and increase the correction coefficient for a social media channel if it is located in an area with a large floating population of young people. The system (400) calculates a final score for each channel by reflecting the regional correction coefficient to the basic suitability score by industry, automatically selects the channel with the highest final score as the channel eligible for headquarters advertising cost support for the store, and links the execution of advertising for the automatically selected channel with the achievement of the store's performance check items, so that advertising costs can be automatically executed for the selected channel only for stores that meet the performance criteria.

[0144] The Industry-Channel Fit Matrix may refer to a mapping table that numerically predefines the suitability between a store's industry classification and multiple advertising channels. Multiple advertising channels may include location-based search channels, social media channels, and location-based community channels. Location-based search channels may refer to local business search services such as Naver Place; social media channels may refer to image-based social network services such as Instagram; and location-based community channels may refer to local lifestyle community services such as Danggeun Market. The types of advertising channels mentioned above are merely examples and are not limited thereto; new advertising channels may be added or existing channels may be changed.

[0145] The system (400) can extract a basic suitability score for each advertising channel by querying the row corresponding to the industry of the store in the industry-channel suitability matrix. The basic suitability score may refer to a value expressed as a numerical value between 0 and 1 that indicates the degree to which each advertising channel can exert marketing effects in light of the characteristics of the industry. For example, in the case of a beauty shop or a nail shop, the basic suitability score of a local search-based channel may be set high because the rate of customer inflow through local search may be high; in the case of a cafe, the basic suitability score of a local community channel may be set high because there may be many daily visits by local residents; and in the case of a fashion shop, the basic suitability score of a social media channel may be set high because image-based visual marketing may be effective. The above-mentioned setting of scores by industry is an example and may be adjusted according to the accumulation of actual marketing performance data.

[0146] The system (400) can calculate a regional correction factor based on the regional characteristics of the store location. The regional correction factor refers to a correction value intended to reflect that the effectiveness of a specific advertising channel may vary depending on regional factors such as the population composition, commercial area type, and characteristics of the floating population in the area where the store is located. The system (400) can increase the correction factor for a regional-based community channel if the store is located in a densely populated residential area. This is because, in densely populated residential areas, information exchange among local residents may be active, so the marketing reach effect through a regional-based community channel may be high. The system (400) can increase the correction factor for a regional search-based channel if the store is located in a densely populated commercial area. This is because, in densely populated commercial areas, the exposure effect of a regional search-based channel may be high, as the rate of customers visiting after searching for a specific service may be high. The system (400) can increase the correction factor for a social media channel if the store is located in an area with a large floating population of the younger age group. This is because the younger demographic may have a high rate of information search and consumption decision-making through social media, which can lead to significant marketing effects via social media channels. The criteria for calculating the aforementioned regional adjustment coefficient are merely an example and are not limited thereto; they may vary depending on changes in the classification criteria for regional characteristics or the addition of new regional types.

[0147] The system (400) can calculate a final score for each channel by reflecting a regional correction coefficient to the basic suitability score by industry. The system (400) can calculate a final score for each advertising channel by multiplying the value obtained by adding the regional correction coefficient of the corresponding channel to the basic suitability score. The system (400) can automatically select the channel with the highest calculated final score as the channel eligible for headquarters advertising cost support for the corresponding store. By selecting only the single channel with the highest score as the target for support, rather than multiple channels, the system (400) can concentrate the limited advertising budget on the most effective channel. By automatically selecting a channel by combining the store industry and regional characteristics, the system (400) can automate the process of a marketing manager manually determining a channel for each store.

[0148] The system (400) can link the execution of advertisements for automatically selected channels with whether the store's performance check items have been achieved. The system (400) can automatically execute advertising costs for selected channels only for stores that meet the passing criteria for the performance check items, and can withhold the execution of advertising costs for stores that do not meet the performance criteria, regardless of the channel selection results. The system (400) performs channel selection and advertising cost execution based on independent judgment criteria, but by combining the condition of whether performance is achieved at the final execution stage, it can control that advertising costs for the optimal channel are invested only in stores that maintain operational quality. Through the above linkage, the system (400) can manage advertising efficiency and store operational quality simultaneously.

[0150] FIG. 8 illustrates the information processing process of an automatic guidance method linked with a carrier AI assistant according to one embodiment.

[0151] Referring to FIG. 8, the system (400) can perform a series of processes such as: when a customer (810) calls a store (840), the carrier AI assistant (820) responds; the carrier AI assistant (820) automatically provides a greeting message generated by the marketing optimization module (831) in the integrated management server (830) to the customer (810); the message is dynamically updated according to the season and promotion; and after the call ends, a promotional text message containing a link to an experience zone QR code (850) is automatically sent to the customer (810) terminal.

[0152] In operation 801, when a customer (810) calls the store (840), a carrier AI assistant (820) can respond. The carrier AI assistant (820) may refer to an AI-based phone response service provided by a carrier for small business owners, and may include, for example, SKT AI Call Assistant, KT AI Call Assistant, and LG U+ AI Phone Assistant. The types of carrier AI assistants are merely examples and are not limited to this, and may vary depending on the carrier services that can be linked. The carrier AI assistant (820) can automatically receive calls received on the store's (840) phone number and provide voice guidance with a pre-set greeting message. By utilizing the carrier AI assistant (820), the system (400) can automatically handle customer service 24 hours a day, even during times when the store operator cannot directly answer the phone.

[0153] In operation 802, the marketing optimization module (831) within the integrated management server (830) can automatically generate a greeting message optimized for each store industry and transmit it to the AI ​​assistant API (820). The system (400) can automatically generate a greeting message by inputting the store's industry classification, store name, current promotion information, and seasonal information corresponding to the store identification code through the marketing optimization module (831). The system (400) can complete the final greeting message by loading a basic template corresponding to the store's industry from the industry-specific greeting template database and dynamically inserting the store name and promotion information. An industry-specific greeting template may refer to a basic structure of a greeting message written in advance to suit the characteristics of each industry. For example, in the case of a hair salon, a template including a thank-you message for the reservation along with information that skin type diagnosis and cosmetic trials are available may be used; in the case of a flower shop, a template including information on lip and cheek gift sets that match the flowers may be used; and in the case of a cafe, a template including information that free cosmetic trials are available when using the cafe may be used. The above industry-specific templates are merely examples and are not limited thereto, and may be updated as store types are added or marketing strategies are changed. The system (400) checks whether the generated greeting message is within a predetermined number of characters (e.g., 120 characters), and if it exceeds this, it can perform automatic abbreviation. The system (400) can transmit the completed greeting message via a carrier AI assistant API.

[0154] In operation 803, the carrier AI assistant (820) can receive a greeting message and provide automatic voice guidance to the customer (810). The carrier AI assistant (820) can convert the greeting message received from the system (400) into voice and automatically provide guidance to the customer (810) who made the call. By automatically guiding all customers who call the store through the carrier AI assistant (820) about the existence of the experience zone and how to use it, the system (400) can achieve continuous marketing effects without additional cost investment from the headquarters.

[0155] In operation 804, the system (400) can determine whether the current month corresponds to a gift season, dynamically replace the greeting message with a season-specific message, and resend it to the AI ​​assistant API (820). The system (400) can activate a gift season flag if the current month corresponds to a predefined gift season month (e.g., February, May, December). When the gift season flag is activated, the system (400) can automatically replace the existing greeting message with a message specialized for the gift season. A gift season-specific message may refer to a greeting message containing marketing messages suitable for the gift season, such as gift recommendations, special discounts, or set configurations. The system (400) can update the greeting by resending the replaced message to the carrier AI assistant API. When the gift season ends, the system (400) can automatically restore the greeting message to a message based on a basic template. The definition of the gift season months above is merely an example and is not limited thereto; it may be changed or added according to promotion schedules or operational policies.

[0156] In operation 805, the system (400) can automatically send a promotional text message containing a link to an experience zone QR code (850) to the customer (810) terminal after the call ends. The system (400) can automatically send a promotional text message when the call between the customer (810) and the store (840) ends by linking with the telecommunications carrier's store information notification text message service. Store information notification text messages refer to an additional service for small business owners provided by the telecommunications carrier, which automatically sends promotional information of the store to the sender in the form of a text message after the call ends. The system (400) can include a link to the store's experience zone QR code (850) in the promotional text message. Through the link in the received promotional text message, the customer (810) can check product information in the experience zone in advance or perform a skin type diagnosis in advance before visiting the store. By linking with the store information notification text message, the system (400) can create an additional point of contact to induce customers who inquire by phone to visit the experience zone, and can build anticipation for the experience even before the customer visits the store. The above store information notification text message service may be applied only if carrier integration is possible.

[0157] According to one embodiment, the system (400) can perform cross-industry skin profile correction based on the history of the same customer visiting multiple different industry experience zones. Specifically, the system (400) can retrieve experience and purchase data accumulated in an integrated database for each customer to extract an industry classification history corresponding to the store identification code of the experience zone the customer previously visited. The system (400) can classify and analyze behavioral data by industry, including the results of skin type diagnosis performed by the customer in multiple different industry experience zones, whether a recommended product was selected, whether an actual purchase was made, and whether a repurchase occurred after purchase.

[0158] Cross-industry skin profile correction may mean updating each component value of the customer skin profile vector by comprehensively reflecting behavioral data accumulated by the customer in experience zones of different industries. The system (400) may increase the reliability weight for the component of the skin profile vector associated with the attribute of the product when the customer actually purchases the recommended product in a specific industry experience zone. Conversely, if the customer does not purchase the recommended product or selects a different product, the system (400) may decrease the reliability weight of the corresponding component and correct the profile vector in the direction of the attribute of the product actually selected by the customer.

[0159] The system (400) can perform the above correction based on the weighted average of behavioral data by industry. The system (400) can calculate an industry-specific contribution coefficient proportional to the number of purchase conversions and repurchase cases in each industry experience zone. By applying the industry-specific contribution coefficient as a weight and integrating behavioral data accumulated from multiple industries, the system (400) can generate a skin profile vector that more precisely reflects the actual skin characteristics and product preference patterns of customers, which are difficult to identify using only visit data from a single industry experience zone.

[0160] The system (400) can update and store the corrected skin profile vector in a customer-specific integrated database. The system (400) can generate a customized recommended product list based on the corrected skin profile vector when the customer subsequently visits any experience zone and scans a QR code. Through the cross-industry skin profile correction, the system (400) can implement a self-reinforcing recommendation structure in which the accuracy of recommendations gradually improves as the number of customer visits to the experience zone increases. The system (400) may perform the correction cycle in real-time whenever new customer purchase data is generated, or it may perform it in batches according to a preset cycle (e.g., weekly or monthly). The correction cycle is merely an example and is not limited thereto, and may vary depending on the operational policy.

[0161] According to one embodiment, the system (400) can perform recommendations for store linkages based on the analysis of customer movement patterns between experience zones. Specifically, the system (400) can analyze experience data accumulated in an integrated database for each customer to extract combinations of experience zone stores that multiple customers have continuously visited within a predetermined period. The system (400) can generate a store linkage visit frequency matrix by aggregating the frequency of occurrence of the extracted store combinations.

[0162] The inter-store linked visit frequency matrix may refer to a mapping table that numerically expresses the frequency of consecutive visits occurring between any two experience zone stores within a predetermined period. The system (400) can identify pairs of stores in the matrix where the linked visit frequency is above a preset standard as pairs of stores with a high degree of linkage. Based on the degree of linkage, the system (400) can provide customers who have visited a specific experience zone with information about other experience zone stores with a high degree of linkage, along with product information recommended in the experience zone of the store in question.

[0163] The system (400) may additionally consider the industry classification, location information, and target customer base data of the two stores when performing inter-store linkage recommendations. The system (400) may perform linkage recommendations only when the physical distance between the two stores is within a predetermined range, thereby including only stores that customers can actually visit by walking or short-distance travel as recommendation targets. If the industry classifications of the two stores are different, the system (400) may increase the weight of the linkage recommendation by reflecting the cross-selling effect between industries. For example, it may linkage recommend a skincare product experience at a nearby cafe experience zone to a customer who visited a flower shop experience zone, or linkage recommend a lip and cheek product experience at a nearby hair salon experience zone to a customer who visited a nail shop experience zone. The combination of the above inter-store linkage recommendations is merely an example and is not limited thereto, and can be dynamically updated based on accumulated customer movement data.

[0164] When transmitting information about linked recommendation target stores to a customer terminal, the system (400) may provide the estimated travel distance to the store, the current operating status of the store, and top recommended product information calculated based on the customer's corrected skin profile in the store's experience zone. By guiding additional experience opportunities at the point of exit after the customer has completed a purchase through linked recommendations, the system (400) can extend the customer journey, which previously ended with a single store visit, to multiple experience zones. Through the linked recommendations between the stores, the system (400) can mutually promote customer inflow between stores within the experience zone network and achieve the effect of increasing the number of QR sales transactions for individual stores.

[0165] The system (400) can update the inter-store linkage visit frequency matrix according to a preset period. The system (400) can continuously update the inter-store linkage as new customer visit data accumulates, stop linkage recommendations for store pairs where the linkage has decreased, and start linkage recommendations for store pairs where the linkage has newly increased. The update period can be set according to an operation policy.

[0167] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave in order to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0168] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

Claims

Claim 1 In an offline store-based inventory-free cosmetics trial order delivery integrated management system, a memory for storing instructions; a communication interface; The system includes a processor, and the instructions, when executed by the processor, when the system accesses an experience zone installed in another store by scanning a QR code of the experience zone, loads a store profile based on a store identification code included in the QR code, receives skin data from the customer terminal to create a customer skin profile, classifies and analyzes behavioral data by industry, including whether the customer actually purchased recommended products in each experience zone based on the history of the same customer visiting multiple different industry experience zones, corrects the profile vector in the direction of the attributes of the product actually selected by the customer by increasing the confidence weight for the component of the skin profile vector associated with the product's attributes when the recommended product is actually purchased, and decreasing the confidence weight for the component when the recommended product is not purchased or another product is selected, and performs cross-industry skin profile correction by integrating behavioral data accumulated in the multiple different industries by applying an industry-specific contribution coefficient proportional to the number of purchase conversions and repurchase cases in each industry experience zone as a weight, and performs similarity analysis between the corrected customer skin profile and product attribute information, while reflecting the industry characteristics of the store corresponding to the store identification code for each product A system that calculates a suitability score, generates a customized recommended product list based on the suitability score and transmits it to the customer terminal, processes payment for the product selected by the customer, transmits a delivery instruction to a logistics center upon completion of payment, calculates a commission per store based on the store identification code and performs automatic settlement, and controls the storage of data generated during the experience and purchase process. Claim 2 In claim 1, when the instructions are executed by the processor, the system loads an access URL containing a store identification code, a product identification code, and an experience zone type code encoded in the QR code into the web browser of the customer terminal; queries a store profile database based on the store identification code to load the industry classification, location information, commission rate, and target customer base information of the corresponding store; quantifies the skin type, skin tone, major concerns, and age group information received from the customer, respectively; calculates a seasonal correction value based on the current month and converts it into a multidimensional vector including a skin type value, a skin tone value, a weight for each concern item, an age group value, and a seasonal correction value to generate a customer skin profile vector; calculates the cosine similarity between the customer skin profile vector and the product attribute vector corresponding to the product attribute information; and to the value obtained by applying a first weight to the cosine similarity, the value obtained by applying a second weight to the seasonal correction value for each product according to the current season, the value obtained by applying a third weight to the industry affinity between the store industry and the product category, and the purchase history bonus based on the customer's past purchase history A system that calculates a skin suitability score for each product by summing the values ​​to which the fourth weight is applied and normalizing the sum of the first to fourth weights to be 1.0, and controls the generation of a customized recommended product list along with the skin suitability scores of each product by selecting a predetermined number of top products in order of highest skin suitability scores. Claim 3 In claim 1, when the above instructions are executed by the processor, the system automatically determines whether each experience zone store has achieved a plurality of performance check items, wherein it determines whether the number of experience photos transmitted by the store to the headquarters server within a specified period is above a preset standard, whether the number of new reviews of the store collected through an external regional search platform API is above a preset standard, whether the number of community platform postings within a specified period is above a preset standard, whether the headquarters inspection score and the negative review rate each satisfy a preset standard, whether the number of unresponsive cases for reservation customers is below a preset standard, and whether the performance of customer social network service authentication postings and merchandise distribution is above a preset standard, and if the sum of the number of passed items among the plurality of performance check items is above a preset passing standard, it automatically approves the payment of advertising fees for the current month for the store, and uses data on the store's industry classification corresponding to the store identification code, the regional characteristics of the store location, and the target customer base as input to query the basic suitability score by industry from the industry channel suitability matrix for each of the plurality of advertising channels, and the region of the store location Based on characteristics, a regional correction coefficient is calculated by increasing the correction coefficient for region-based community channels if the store is located in a densely populated residential area, increasing the correction coefficient for region-based search channels if it is located in a densely populated commercial area, and increasing the correction coefficient for social media channels if it is located in an area with a high volume of young people; the single channel with the highest final score calculated by reflecting the regional correction coefficient to the basic suitability score by industry is automatically selected as the channel eligible for headquarters' advertising cost support, and advertising costs are automatically executed for the selected single channel only for stores that meet the passing criteria; if the store falls below the passing criteria, a warning notification is automatically sent to the store owner including both the current and reference values ​​for each item that is deficient.A system that automatically classifies stores into multiple grades based on the monthly QR sales volume per store upon the end of a designated pilot period, and controls the determination of differential measures, including contract extension, conversion to store self-funding, or termination of the pilot, according to the classified grade. Claim 4 A system that controls the above instructions, when executed by the processor, to verify the business classification of a store corresponding to the store identification code, and if the business of the store is a flower shop, to extract the representative HSV color value of the flower variety handled by the store from the flower variety database, to extract the HSV color value of each product from the product database belonging to the lip and cheek category, to calculate the color difference between the representative HSV color value of the flower variety and the HSV color value of each product, wherein the color difference value is calculated by squaring the difference values ​​of hue, saturation, and brightness, summing them, and taking the square root, to select a product whose color difference value is below a preset threshold as a flower color matching recommended product, to determine whether the current time corresponds to the gift season, and if it corresponds to the gift season, to increase the recommendation weight for the matching recommended product, and to integrate the flower color matching recommended product with the custom recommended product list based on the suitability score to generate an industry-specific cross-selling recommendation list and transmit it to the customer terminal. Claim 5 In claim 1, the above instructions, when executed by the processor, are a system that takes data on the industry classification of a store corresponding to the above store identification code, the regional characteristics of the store location, and the target customer base as input, retrieves a basic suitability score by industry from an industry-channel suitability matrix for each of a plurality of advertising channels, calculates a regional correction coefficient based on the regional characteristics of the store location, increases the correction coefficient for a regional-based community channel if the store is located in a densely populated residential area, increases the correction coefficient for a regional-search-based channel if it is located in a densely populated commercial area, increases the correction coefficient for a social media channel if it is located in an area with a large floating population of young people, calculates a final score by channel by reflecting the regional correction coefficient in the above basic suitability score by industry, automatically selects the channel with the highest final score as the channel eligible for headquarters advertising cost support for the store, and controls the execution of advertising for the automatically selected channel in conjunction with whether the store meets preset performance criteria, thereby controlling the automatic execution of advertising costs for the selected channel only for stores that meet the performance criteria.

Images