Search

KR-20260065383-A - How to provide promotions for custom products based on matching scores

KR20260065383AKR 20260065383 AKR20260065383 AKR 20260065383AKR-20260065383-A

Abstract

The present invention relates to a method for providing customized advertising information based on a matching score, comprising a service and method for searching for said cosmetics based on the product name, product outline, and user's skin condition, and calculating matching information for said cosmetics using pre-set weights according to skin type based on the total ingredients, individual ingredients, and user reviews of said cosmetics, wherein the service and method include: a skin diagnosis step that classifies the user's skin type and outputs it as a keyword by matching one of the pre-classified skin types based on a user's survey or skin photograph; a cosmetic evaluation and quantification step that sets evaluation items based on pre-extracted classified keywords of said cosmetics, evaluates the effectiveness of each item based on the product name, ingredients, and reviews of said cosmetics searched, and quantifies the score by step or keyword; a matching score quantification step that forms each persona for cosmetic evaluation items that match the customer's skin keyword, compares them, and represents the matching score; and a product exposure step that receives matching information calculated from said service server when searching for a new product, and provides some benefit including free samples/discount coupons, etc., when a higher matching score is obtained.

Inventors

  • 최준호

Assignees

  • 최준호

Dates

Publication Date
20260508
Application Date
20241101

Claims (5)

  1. A service and method for calculating matching information for said cosmetics by searching for said cosmetics based on the product name, product outline, and user's skin condition, and using pre-set weights according to skin type based on the total ingredients, each ingredient, and user reviews of said cosmetics, A skin diagnosis step that classifies the user's skin type by matching it to one of the pre-classified skin types based on the user's survey or skin photographs, and outputs the result as a keyword; A cosmetic evaluation and quantification step that sets evaluation items based on previously extracted classified keywords of cosmetics, evaluates the effectiveness of each item based on the product name, ingredients, and reviews of the searched cosmetics, and quantifies scores by stage or keyword; A matching score generation step for cosmetic evaluation items that match the customer's skin keywords, wherein each persona is formed and compared to represent a matching score; A product exposure stage characterized by receiving matching information calculated from the service server when searching for a new product, and providing benefits including free samples/discount coupons, etc., when a higher matching score is obtained; and a method for providing customized advertising information based on a matching score.
  2. In Article 1, A method for providing customized advertising information based on a matching score, wherein the matching information includes a matching rate indicating the degree to which the cosmetic is suitable for the user's skin, guide information predetermined according to the matching rate, and the entire ingredients, key ingredients, harmful ingredients, and notified ingredients of the cosmetic.
  3. In Article 1, The 'skin diagnosis step' is a method of providing customized advertising information based on a matching score, wherein a mobile application is activated, a survey and a user's face are captured through the activated mobile application to obtain survey information and a face image, and a first deep learning model that is pre-trained based on the obtained information and face image is used to estimate the user's skin condition.
  4. In claim 1, the 'cosmetic evaluation and quantification step' is a method of providing customized advertising information based on a matching score that includes a natural language processing-based service that sets keyword-specific items, which serve as classification criteria for the first deep learning model of claim 3, as evaluation criteria, and estimates keyword-specific scores of cosmetics based on customer review data and product ingredients.
  5. In paragraph 1, A numerical providing system that, in the ‘matching score conversion step,’ activates a mobile application, captures a user’s face through the activated mobile application to obtain a face image, and estimates the user’s skin condition using a first deep learning model that has been pre-trained based on the obtained face image. Claim 6, in the first thereof, wherein the ‘product exposure step’ is a method for providing customized advertising information based on a matching score, which includes a service that induces customer interest based on product information scored according to Claim 5 and provides benefits, including the provision of free product samples, when searching for a product with a higher score.

Description

How to provide promotions for custom products based on matching scores The present invention relates to a system that provides customized cosmetic recommendations and promotions by analyzing the suitability of cosmetic ingredients to a user's skin condition. The present invention analyzes user skin information collected through surveys and skin photography using a deep learning model and recommends products suitable for the user's skin condition based on the results. Cosmetic ingredients and user review data are analyzed using Natural Language Processing (NLP) technology to establish evaluation criteria for each cosmetic product and assign scores. A matching score is calculated by comparing user skin information with cosmetic evaluation data, and cosmetics with a high degree of suitability are recommended based on this score. User participation is encouraged by providing customized benefits, such as free samples or discount coupons, for products with a high degree of matching. Regarding matching-based customized advertising, the applicant's prior art is Korean Patent No. 10-2020-0082849. Modern consumers explore and purchase products and services through various online and offline channels; accordingly, companies are striving to implement customized marketing tailored to each consumer's preferences and purchasing patterns. In particular, in product promotion strategies, providing customized promotions tailored to individual consumers plays a crucial role in increasing customer engagement and purchase conversion rates, and customized marketing is especially important in the purchase of cosmetics. Existing personalized cosmetic promotions are primarily based on consumers' past purchase records or basic demographic information, and this approach often fails to adequately reflect the specific preferences or needs of individual consumers. Furthermore, most current systems have limited capabilities for recommending products or providing customized promotions by analyzing the correlation score (matching score) between the consumer and the product in real time. These system limitations prevent the suggestion of appropriate products at the right time, making it difficult to attract consumer attention and convert them into purchases. Furthermore, existing marketing and promotion systems have limitations in effectively collecting and analyzing consumer interaction data, and significant improvements are needed in accurately predicting products or services that specific consumers are likely to prefer. Accordingly, there is a demand for a method to provide customized product promotions based on individual consumers' matching scores. Based on this background technology, the present invention aims to maximize marketing efficiency by providing consumers with more appropriate and effective customized promotions based on matching scores. FIG. 1 is a flowchart of a method according to the present invention. The specific details for implementing the present invention are configured as follows. In the skin diagnosis step, when a user requests a skin diagnosis through a mobile application, data is collected through a survey and skin photography. The survey includes the user's skin type, major skin concerns, and currently used products, while the skin photography is intended to input the user's face image into a deep learning-based skin condition diagnosis model. The collected data is analyzed through a pre-trained first deep learning model to estimate the user's skin type (e.g., oily, dry, combination) and condition (e.g., sensitive, acne, wrinkles, etc.). Based on this information, the user's skin condition is diagnosed, and the diagnosis results are classified and output in the form of keywords. Cosmetic Data Collection and Evaluation Item Setting Step: The present invention collects ingredient information, product names, and user reviews of various cosmetics and stores them in a database. Each cosmetic product is classified by specific keywords to evaluate ingredients that match skin types, and these keywords are used as evaluation items to match with skin diagnosis results. User review data is analyzed using Natural Language Processing (NLP) technology, and evaluation scores are set by quantifying the effect of each ingredient on a specific skin type. For example, the system is configured such that a high score is assigned if an ingredient is suitable for sensitive skin, and a low score is assigned if an ingredient is unsuitable for sensitive skin. In the matching score calculation stage, a matching score is calculated based on user skin diagnosis results and cosmetic evaluation data. The matching score quantifies how closely the user's skin condition keywords match the cosmetic's evaluation criteria. This score serves as an indicator of suitability for the user's skin type and condition, allowing users to objectively evaluate products that suit them. For example, if a user's skin condition is sensitive and dry, products with