TWM588837U - Optimization system for consumer recommendation information and purchasing decision - Google Patents

Abstract

An optimization system for consuming recommendation information and purchasing decisions, mainly based on the label selection instruction of a communication device that has not been consumed (that is, after a category label is selected), filtering out other category labels that are associated with the label selection instruction to allow The communication device can be further selected from the filtered other classification tags, thereby improving the efficiency of the user browsing product information, and reducing the flow of the communication device continuously sending query requests to the system, and the operation history of selecting each classification tag is captured by the system After retrieval, a potential preference analysis will be performed to provide a consumption recommendation information (or personalized advertisement) to the communication device to find out the actual needs of consumers and improve the conversion rate of the website. At the same time, the system can use the analysis results Generate a purchasing decision proposal.

Description

Optimization system for consumption recommendation information and purchase decision

This creation is related to cloud computing, data mining (Data Mining), Data Warehouse (data warehousing) and the technical field of analysis of usage behavior, especially a kind of analysis of the user's operation history on the e-commerce platform before consumption, to Generate consumption recommendation information more in line with the actual needs of users, and provide an "optimized system of consumption recommendation information and procurement decisions" that provides optimized procurement decision suggestions.

With the development of communication technology and the development of the Internet, many users will make purchases on an e-commerce platform through computers or smart phones, and many merchants will analyze consumer consumption history, shopping cart records and commodities through analysis Browse the history, and push a consumption recommendation information on the e-commerce platform in order to increase the purchase rate, or put a dynamic advertisement on the social platform, in order to attract users to visit the e-commerce platform through the guidance of the social platform, and Add goods to the shopping cart or complete the purchase, and increase the conversion rate of the e-commerce platform.

However, whether the e-commerce platform establishes consumption recommendation information or creates a personalized dynamic advertisement in an advertising system, most of them refer to the analysis results of the consumer’s consumption history, product page browsing history and shopping cart records. Know which (and what) commodities on the platform are more popular or unsalable, in other words, the analysis results that the merchants refer to only meet the consumers who have completed the transaction. As for potential consumers or visitors who have not completed the purchase, it is difficult to understand the potential At the same time, it is difficult for the purchasing side to evaluate the purchase quantity of different commodities and their specifications from the consumption records, so there is still a problem of insufficient stock or excessive stocking. The relevant previous cases such as the Republic of China Invention Patent Publication No. TW200411496 "Target marketing system and method", China Invention Patent Publication No. CN108053282A "Combined information push method, device and terminal" and other previous cases were disclosed.

In addition, the merchant usually needs to change the settings of the advertising system or adjust the creatives and copy to test the effectiveness of ad delivery, and then evaluate the ad delivery through the ad conversion rate, return on ad spend (ROAS), and click rate (CTR) Effectiveness, in other words, the merchant must pass multiple tests to adjust the creative or copy before it can obtain better advertising performance, but in this way, the merchant will spend a higher cost of advertising.

Furthermore, when the consumer is browsing the e-commerce platform, it is usually only possible to classify products in a hierarchical tree structure (for example, if the main category is "popular women's clothing", then "tops", "shirts", "skirts", etc. Category is the sub-category of "Popular Women's Wear"), to find the products you want to buy, but if the user wants to immediately find products that meet certain conditions (such as a specific color, situation, popular element or origin), it can only One by one, from the product pages corresponding to different sub-categories (the product page is a third-level structure that belongs to the "sub-category"), find out whether there are products that meet their needs. For the products that are actually in demand, consumers continue to send query requests to the server of the e-commerce platform through the information device, which is an unnecessary waste. At the same time, the server also has to frequently extract data from the database, and Will increase the processing load of the system.

In summary, the current e-commerce system still has the above problems that need to be improved. Therefore, how to provide a product with better query efficiency, can improve the effectiveness of consumer recommendation information or personalized advertising, and can reduce the number of information devices to continue to the server The number of inquiry requests sent, an information provision system that can understand the actual or potential needs of consumers, and can optimize purchasing decisions are issues to be solved.

In order to achieve the above purpose, this creation discloses an optimization system for consumption recommendation information and purchase decision, which includes a central processing module, and a communication module, a commodity database, a classification label module, and an information retrieval module , An operation history database and a recommendation module are respectively connected to the central processing module to present information.

Among them, the communication module mainly communicates with a communication device on the consumer side or potential consumer side; the commodity database can store multiple pieces of commodity information, and each commodity information is set to be associated with one or more classification labels; the classification label module can Store multiple classification labels, and also select a label selection command sent by selecting one of the classification labels according to the communication device, and extract other classification labels that have an association relationship with the label selection instruction from other classification labels that have not been selected to limit The communication device can only further select from the extracted other classification tags; the information extraction module can extract one or more tag selection commands sent by the communication device to select one or more classification tags before completing consumption, And generate an operation history information and store it in the operation history database; and the recommendation module can perform a potential preference analysis based on the operation history information to calculate a consumption recommendation information for a communication device that has not yet been consumed, and at the same time As a result, and a purchasing decision parameter stored in the commodity database, a purchasing decision suggestion is generated.

Furthermore, the communication module can be further connected to an advertising server, and the advertising server can group communication devices classified into a target audience on a third-party platform (such as a community platform or content aggregation) according to a target audience parameter. (Platform) Place a personalized advertisement that meets the consumption recommendation information, and the information retrieval module can also track a conversion event generated by the communication device visiting the e-commerce platform from the third-party platform, and the conversion event is associated with at least one category label.

In order to enable your reviewing committee to clearly understand the purpose, technical features and effects of this creation, the following description is accompanied by illustrations, please refer to it.

Please refer to "Picture 1", which is a system architecture diagram of this creation, which reveals an optimization system 10 for consuming recommendation information and purchasing decisions, which may itself be a physical server or in the form of a virtual machine (VM) It runs on a physical server or in the form of a virtual private server (VPS), but it is not limited to this. The optimization system 10 for consuming recommendation information and purchasing decisions includes: a central processing module 101, And a communication module 102, a commodity database 103, a classification label module 104, an information retrieval module 105, an operation history database 106 and a recommendation module 107 are respectively connected to the central processing module 101 to present information ,among them:

(1) The central processing module 101 is used to run the optimization system 10 for consumption recommendation information and purchasing decisions. It can drive the actions of the above-mentioned modules, and has functions such as logical operations, temporary storage of operation results, and storage of execution instruction positions. It can be a central processing unit (CPU). (2) The communication module 102 is used to communicate with one or more communication devices 20, and the communication module 102 may have one or a combination of Wi-Fi, 3G/4G, and Ethernet transmission interfaces. (3) The commodity database 103 may be a database host, which stores multiple pieces of commodity information, and each commodity information is set to be associated with one or more classification labels. (4) The classification label module 104 can store multiple classification labels, and can also select one of the classification labels and send a label selection command according to the communication device 20 as an index for reading the classification labels of the commodity database 103. (5) The information retrieval module 105 is used to retrieve one or more label selection commands sent by the communication device 20 to select one or more of the classification labels before completing consumption, and generate an operation history information, and It is stored in the operation history database 106, and the operation history database 106 can be a database host. (6) The recommendation module 107 is used to perform a potential preference analysis based on the aforementioned operation history information to calculate consumption recommendation information for the communication device 20, and at the same time based on the result of the potential preference analysis and a purchase decision stored in the commodity database 103 Parameters to generate a purchasing decision proposal. A. The aforementioned potential preference analysis is a collaborative filtering (CF) algorithm, more specifically, it can be further divided into a user-based collaborative recommendation algorithm (User-based-CF) and a project-based Collaborative recommendation algorithm (Item-based-CF). B. For example, the user-based collaborative recommendation algorithm means that, assuming that user A likes category labels A, B, and C, and user B likes category labels A and B, recommendation module 107 can infer that B Users may also like category label C; and item-based collaborative recommendation algorithm means that user A likes category labels A, B, C, user B likes category labels A and C, and user C likes category labels A, the recommendation module 107 can infer that the user who selects the category label A will also want to purchase the product associated with the category label C. C. The aforementioned consumption recommendation information may include a recommended product, a potential preference classification, and a combination of potential preference classifications, and the information contained in the consumption recommendation information is associated with at least one classification label stored in the classification label module 104. D. Among them, the aforementioned purchasing decision parameters may include a purchasing quantity trend, an inventory warning information, a commodity purchase price trend trend, a purchasing cost trend, a closing information, a sales quantity trend, a shopping cart record information and The aforementioned potential preferences are classified.

Please refer to "Figure 2", which is a schematic diagram of the system information flow of the creation, and please refer to the system implementation flow chart of "Figure 3". This creation exposes an implementation method of an optimization system for consumer recommendation information and purchase decision, It includes the following steps: (1) Setting the classification label ( step S1) : After a classification label module 104 is driven by a central processing module 101, one or more classification labels associated with multiple pieces of product information are set and stored in a Commodity database 103; (2) Present selectable classification labels on the e-commerce platform ( step S2) : the central processing module 101 can be connected to an e-commerce platform through the communication module 102 The front-end interface presents a selection of multiple classification tags (T1~T6), where the e-commerce platform itself can run on a web server (Web Server); (3) Analysis of tag selection commands ( step S3) : classification tag mode Group 104 uses a communication device 20 to select a label selection command generated by selecting one of the classification labels T1 as an index for reading the product database 103, and extracts the label selection from other classification labels (T2~T6) that have not been selected yet. Instruct other classification tags (T2, T4, T5) that have an association relationship to restrict the communication device 20 to only further select from the extracted other classification tags (T2, T4, T5); more specifically, the classification tag module 104 The basis for judging whether the aforementioned classification labels (T1~T6) are related to each other lies in confirming the product information (commodity A) associated with the selected classification label T1 according to the classification label set for each product information in step S1 In step S1, is it set to be associated with other classification tags (T2, T4, T5), as shown in the schematic diagram of the classification tag association of "Figure 4", if so, the communication device 20 is not allowed to The front-end interface selects irrelevant classification tags (T3, T6); (4) Retrieval operation history ( step S4) : the central processing module 101 drives an information retrieval module 105, so that the communication device 20 can complete In step S3, one or more classification tags (T1~T6) are selected and one or more classification tag commands sent are retrieved, and an operation history information is generated and stored in an operation history database 106; ( 5) Perform potential preference analysis ( step S5) : The central processing module 101 drives a recommendation module 107 to perform a potential preference analysis based on the aforementioned operation history information, and calculates a consumption recommendation information for the communication device 20, based on The result of the potential preference analysis and a purchasing decision parameter stored in the commodity database 103 generate a purchasing decision suggestion.

Please refer to "Figure 5", which is a schematic diagram of the data table of the usage history database of the creation, and please refer to "Figure 2" and "Figure 4" as shown in "Figure 5". The operation history information stored in the operation history database 106 may include a selection priority, a selection frequency, a selection frequency percentage, and a browsing time generated by the communication device 20 due to the selection of each classification label (T1~T6), according to Therefore, the system can determine the potential behavioral preferences of consumers (both members or visitors) based on at least one table T of the operation history database 106. For example, if the user first selects "Party" and then selects "Long Top" Classification label, but the history database 106 does not record that the user has browsed any products, did not add any products to the shopping cart, or did not take direct purchases, it may represent products corresponding to "party" and "long shirt" Has been out of stock, or not sold in accordance with the size/specification that the user wants to buy, and at the same time, it can also be used with the one-turn rate information or one inventory information stored in the product database 103 to understand the purchase of the purchase side Is the amount insufficient?

To continue, please continue to refer to "Figure 6", which is a schematic diagram of the classification label association of another embodiment (1) of this creation. This embodiment is related to the technical categories disclosed in "Figure 1" to "Figure 5" The main difference is that the multiple classification tags (T1~T6) stored in the classification tag module 104 can be divided into multiple main classification tags (M1~M3...) and multiple sub-classification tags (T1'~T6') , And each sub-category label (T1'~T6') belongs to at least one of the main category labels (M1~M3...), more specifically, in this embodiment when step S3 shown in "Figure 3" is executed , The classification label module 104 is a label selection instruction generated by the communication device 20 selecting one of the sub-categorization labels T1', and extracts the label selection instruction from other sub-categorization labels (T2'~T6') that have not been selected yet Other sub-category tags (T2', T4', T5') with an association relationship, so as to restrict the communication device 20 to only further select from the extracted other sub-category tags (T2', T4', T5'). The classification label module 104 of the embodiment determines whether each sub-classification label (T1'~T6') is related to each other, which is similar to the determination method in step S3, and will not be repeated here; After the embodiment is implemented, when the information extraction module 105 executes step S4, it can further extract the main classification label (M1) to which one or more sub-classification labels (T1', T2', T4', T5') belong. , M3), so that when the recommendation module 107 executes step S5, the name and the number of times corresponding to the main classification labels (M1, M3) can be included in the calculation basis for performing potential preference analysis.

Please refer to "Figure 7", which is a system architecture diagram of another embodiment (2) of the creation, and please refer to the system implementation flowchart of another embodiment (2) of "Figure 8", this implementation The example is similar to the technology disclosed in the "Picture 1" to "Picture 5" embodiments. The main difference is that the communication module 102 can also communicate with an advertisement server 30, and after step S5 is executed, it can be continued Perform the following steps: (2) Create dynamic advertising parameters ( step S6) : After the communication module 102 establishes communication with the advertising server 30, create and store a merchant terminal number, a product catalog, and corresponding product information on the advertising server 30 One or more product identification codes, a product link URL, a product image, one or more product classification labels, and a target audience parameter, and establish a target audience group based on operation history information of multiple communication devices 20. A. The aforementioned target audience grouping refers to the operation history information collected according to the information extraction module 105. The recommendation module 107 calculates an audience similarity between consumers and uses the audience similarity similarly The user group is classified into the same target audience group; in addition, the advertising server 30 can also calculate the aforementioned target audience group based on the application behavior of multiple communication devices 20 on a social platform or a content aggregation platform. By reading consumption recommendation information and a member database 108 of the optimization system 10 for purchasing decisions, the member information of the community platform or content aggregation platform can be combined with the e-mail, phone number, and serial platform stored in the member database 108 The member IDs are matched to achieve more precise delivery of personalized advertisements when step S7 is executed. B. The aforementioned target audience parameters may include: viewed products, added to the shopping cart, incomplete purchases, added or excluded users who have purchased a specific product combination, ever returned to the e-commerce platform, added or excluded URLs that have visited specific websites Parameters such as keywords, a specific time interval, top ten best-selling commodities, spot products, zero yards, clearance discounts, etc. C. More specifically, the target audience parameters can be set to "visited e-commerce platform in the past 30 days", "added to the shopping cart but not completed the purchase", "viewed a specific product but not added to the shopping cart", etc. , But not limited to this. D. The aforementioned advertising server 30 may be an advertising server running a social platform (such as Facebook), or a content aggregation platform (such as Google Ads, Yahoo Gemini, or Bing Ads), but not limited thereto. E. The information retrieval module 105 of this step must be embedded with an event tracking code (Event Tracking Code) with a custom parameter, so that after the operation history information is collected, the advertisement server 30 can directly receive the For the operation history information of the event tracking code, if the advertising server 30 is the advertising server of the Facebook platform as an example, the aforementioned event tracking code may be a base code (Facebook that is adapted to a Facebook Pixel supported by the Facebook platform Pixel Code), the event tracking code must be embedded in the website code of one or more website pages (commercial link URL) to be tracked, but not limited to this. F. In the preferred embodiment of this embodiment, the information retrieval module 105 of this step may further embed a global site code (global site tag) to track the URL of the webpage browsed by the user after visiting the e-commerce platform With the title, the ad server 30 can use this information to create the aforementioned target audience group. (3) Place personalized ads on a third-party platform ( step S7) : According to the target audience parameters, the advertising server 30 browses an e-commerce platform, a community platform, or a content aggregation platform, and is classified into the target audience group The communication device 20, on a third-party platform (such as a community platform or content aggregation platform), delivers a personalized advertisement that meets the consumption recommendation information, and for a communication device 20 that has a browsing e-commerce platform or community platform or content aggregation platform, The information retrieval module 105 can also track a conversion event generated by the communication device 20 visiting the e-commerce platform through the social platform or the content aggregation platform, and the conversion event is associated with at least one category label (T1~T6). (4) Among them, the aforementioned conversion event may refer to the event that the communication device 20 visits the e-commerce platform through the personalized advertisement placed on the third-party platform, or completes the purchase after visiting the e-commerce platform, or visits the electronic The event of adding a shopping cart after a business platform, the event of completing a member registration after visiting an e-commerce platform, or the event of browsing a specific product link after visiting an e-commerce platform, but not limited to this.

According to the above, after this embodiment is implemented, personalized advertisements can be placed on the social platform (such as Facebook's dynamic wall) or content aggregation platform (such as presented on the Google search engine or Google's partners) browsed by the communication device 20 Platform), in order to achieve behavior analysis based on pre-consumption, to achieve the purpose of remarketing advertising, and can alleviate the problem that the conventional merchants can only spend more time on advertising costs by adjusting the creative materials or copywriting multiple times. Increase conversion rate (CVR) and reduce consumer acquisition cost (CAC).

Please refer to "Figure 9", which is a system architecture diagram of another embodiment (3) of the creation, as shown in the figure, this embodiment and "Figure 1" ~ "Figure 5", or "Figure 7" ~ The technology of the embodiment disclosed in "Figure 8" is similar, the main difference is that the consumption recommendation information and the procurement decision optimization system 10 of this embodiment can further include a consumption history data linked to the central processing module 101 to present information Library 109, which stores a consumption history information D5 of a plurality of communication devices 20 that have completed consumption, according to this embodiment, after step S3 of "Figure 3" or "Figure 8" is executed, the information is retrieved The module 105 can first retrieve the cancellation fee history information D5 from the consumption history database 109, so that when step S5 is executed, it is used as a calculation basis for the recommendation module 107 to perform potential preference analysis. Therefore, in this embodiment, the user can be used for consumption The previous operational behavior and the consumption history of the users who completed the consumption are included in the consideration of potential preference analysis.

In summary, after the implementation of this creation, at least it can help consumers to browse and find suitable product information on the e-commerce platform, and reduce the number of communication devices that continue to send query requests to the server. Beneficial effects, can better understand the actual or potential needs of consumers, and provide consumers with consumer recommendation information that meets their needs, or allow consumers to guide personalized ads placed on third-party platforms to meet their needs. At the same time, the product page of the e-commerce platform can also achieve the beneficial effect of optimizing purchasing decisions.

The above are only the preferred embodiments of this creation, and are not intended to limit the scope of the implementation of this creation; anyone who is familiar with this skill will make equal changes and modifications without departing from the spirit and scope of this creation. It should be covered by the patent scope of this creation.

In summary, this creative department has patent requirements such as "industrial utility", "novelty" and "progressiveness"; the applicant has filed an application for a new type of patent with the Bureau of Law in accordance with the provisions of the Patent Law.

10‧‧‧Optimized system for consumption recommendation information and purchasing decision

101‧‧‧Central Processing Module

102‧‧‧Communication module

103‧‧‧Commodity database

104‧‧‧ Classification label module

T1~T6‧‧‧ Classification label

M1~M3‧‧‧Main classification label

T1’~T6’‧‧‧Sub-category label

105‧‧‧ Information extraction module

106‧‧‧Operation history database

T‧‧‧ Data Sheet

107‧‧‧Recommended module

108‧‧‧ Member database

109‧‧‧Consumer History Database

D5‧‧‧Consumer history information

20‧‧‧Communication device

30‧‧‧ad server

D1‧‧‧ selectable classification label

D2‧‧‧ Tag selection command

D3‧‧‧Other classification tags associated with the selected tag

D4‧‧‧Consumption recommendation information based on pre-consumption behavior

D4’‧‧‧Consumption recommendation information generated by incorporating consumption history

S1‧‧‧Set category label

S2‧‧‧ presents selectable classification labels on e-commerce platform

S3‧‧‧Analyze the label selection command

S4‧‧‧ Capture operation history

S5‧‧‧ perform potential preference analysis

S6‧‧‧Create dynamic advertising parameters

S7‧‧‧Personalized advertising on third-party platforms

Figure 1 is a system architecture diagram for this creation.

Figure 2 is a schematic diagram of the system information flow for this creation.

Figure 3 is a flowchart of the system implementation for this creation.

Figure 4 is a schematic diagram of the classification labels associated with this creation.

Figure 5 is a schematic diagram of the data table of the usage history database for this creation.

Fig. 6 is a schematic diagram of classification label association in another embodiment (1) of the creation.

Fig. 7 is a system architecture diagram of another embodiment (2) of the creation.

Figure 8 is a flowchart of the system implementation of another embodiment (2) of this creation.

Figure 9 is a schematic diagram of the system architecture and information flow of another embodiment (3) of the creation.

10‧‧‧Optimized system for consumption recommendation information and purchasing decision

101‧‧‧Central Processing Module

102‧‧‧Communication module

103‧‧‧Commodity database

104‧‧‧ Classification label module

105‧‧‧ Information extraction module

106‧‧‧Operation history database

107‧‧‧Recommended module

20‧‧‧Communication device

D1‧‧‧ selectable classification label

D2‧‧‧ Tag selection command

D3‧‧‧Other classification tags associated with the selected tag

D4‧‧‧Consumption recommendation information based on pre-consumption behavior

Claims (9)

An optimization system for consumption recommendation information and purchase decision for generating a consumption recommendation information corresponding to a communication device, the system includes: a central processing module, a communication module, a commodity database, a classification label module, An information retrieval module, an operation history database, and a recommendation module are respectively connected to the central processing module for information, and the central processing module is used to run the optimization system of the consumption recommendation information and purchase decision, and drive the above The operation of each module; the communication module is used to communicate with the communication device; the commodity database stores multiple pieces of commodity information, each of which is set to be associated with one or more classification labels; the classification label module The group is used to store a plurality of the classification labels, and a label selection instruction sent according to the communication device selecting one of the classification labels, and extracting the label selection instruction from the other unselected classification labels Other classification labels of the association relationship, so as to restrict the communication device to be able to select only from the other extracted classification labels; the information extraction module is used for the communication device to complete one or more of the One or more of the label selection commands sent by the classification label for selection are extracted, and an operation history information is generated and stored in the operation history database; and the recommendation module is used to execute a potential based on the operation history information Preference analysis, to calculate the consumption recommendation information for the communication device, and generate a purchase decision suggestion based on the result of the potential preference analysis and a purchase decision parameter stored in the commodity database. For example, the system for optimizing consumption recommendation information and purchasing decisions in the first patent application, where the operation history information is a selection priority order, a selection number, a selection number percentage, and a browsing for each classification label of the communication device One or a combination of time, and the operation history information further includes a view number of the product information. For example, the system for optimizing consumption recommendation information and purchasing decisions in the first patent application, in which the communication module is also used to communicate with an advertising server, which is used to store a product catalog, a or Multiple commodity identification codes and a target audience parameter also store a target audience group generated based on the operation history information of multiple communication devices. For example, the optimization system for consumer recommendation information and purchase decision in the third patent application, where the advertising server is also used to browse an e-commerce platform, a community platform or a content aggregation platform based on the target audience parameters, And the communication device classified into the target audience group is placed with a personalized advertisement that matches the consumption recommendation information. For example, the system for optimizing consumption recommendation information and purchasing decisions in the fourth patent application, where the information extraction module is also used to track a conversion event generated by the communication device visiting the e-commerce platform through a third-party platform, The information extraction module is embedded with an event tracking code with a custom parameter, and the conversion event is associated with at least one of the classification tags. For example, the system for optimizing consumption recommendation information and purchasing decisions in the first or fourth patent application scope, which also includes a consumption history database linked to the central processing module for information, which stores much of the completed consumption A piece of consumption history information of the communication device. For example, the system for optimizing consumption recommendation information and purchasing decisions in the 6th range of patent applications, in which the information retrieval module is also used to retrieve the consumption history information from the consumption history database and send it to the recommendation module, The basis for calculating the consumption recommendation information and the purchase decision recommendation by including the recommendation module. For example, the optimization system for consumer recommendation information and purchase decision in the first patent application, in which the communication module is also used for the central processing module to connect to an e-commerce platform, which can be displayed on a front-end interface of the e-commerce platform Multiple classification labels for the communication device to select. For example, the optimization system for consumer recommendation information and purchase decision in the third patent application, in which the advertising server runs a community platform or a content aggregation platform.

Images