Information recommendation method and device, electronic equipment, computer readable storage medium and computer program product

Abstract

The application provides an information recommendation method and device, electronic equipment, computer readable storage medium and computer program product. The method comprises: obtaining a plurality of to-be-recommended areas of a to-be-recommended store, and obtaining an intended object of the to-be-recommended store; obtaining object access data for each to-be-recommended area, and determining a to-be-recommended object of each to-be-recommended area based on the object access data of each to-be-recommended area; determining a regional recommendation degree of each to-be-recommended area based on the to-be-recommended object of each to-be-recommended area and the intended object of the to-be-recommended area; and determining an advertisement placement strategy of each to-be-recommended area for the to-be-recommended store based on the regional recommendation degree of each to-be-recommended area. Through the application, the regional characteristics can be effectively combined to realize accurate placement of advertisements.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to an information recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product. Background Technology

[0002] In local businesses (including catering and entertainment, supermarkets and retail, real estate and home improvement, home services and repair, travel and car rental, and business services), merchants consider factors such as the surrounding commercial ecosystem, consumer demographics, and pedestrian traffic when choosing a location. Geographical location is a crucial characteristic determining the overall value of customer traffic. These businesses often use physical stores as the core to radiate to customers in the surrounding area, utilizing an online-offline integrated business model to improve operational efficiency. That is, the online platform serves as a window for information display and customer acquisition, while the physical store provides actual products or services.

[0003] Effectively integrating real-world geographic information into online promotions is key to improving advertising efficiency and long-term value. Traditional advertising attribution methods can only detect online clicks and cannot link them to offline consumption data. In related technologies, the emergence of localized group-buying coupons has broken down the spatial barriers between online and offline, allowing online conversion data to be statistically analyzed through offline redemption data. Furthermore, some platforms utilize data mining and analysis techniques to correlate the redemption and usage of online coupons with offline consumption scenarios. For example, by analyzing information such as the time and location of coupon usage and matching it with transaction data in the merchant's sales system, the offline consumption conversion effect of online promotions can be determined. Summary of the Invention

[0004] This application provides an information recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product that can effectively combine regional characteristics to achieve precise advertising.

[0005] The technical solution of this application embodiment is implemented as follows:

[0006] This application provides an information recommendation method, including:

[0007] Obtain multiple recommended areas for the stores to be recommended, and obtain the intended customers for the stores to be recommended;

[0008] Obtain object access data for each of the regions to be recommended, and determine the objects to be recommended for each region based on the object access data for each region to be recommended;

[0009] Based on the objects to be recommended in each region to be recommended and the intended objects in each region to be recommended, the regional recommendation degree of each region to be recommended is determined.

[0010] Based on the regional recommendation level of each of the regions to be recommended, an advertising strategy for each of the regions to be recommended is determined for the stores to be recommended.

[0011] This application provides an information recommendation device, including:

[0012] The acquisition module is used to acquire multiple areas of stores to be recommended and to acquire the intended users of the stores to be recommended.

[0013] The determination module is used to acquire object access data for each of the regions to be recommended, and to determine the objects to be recommended for each region based on the object access data for each region to be recommended;

[0014] The determining module is further configured to determine the regional recommendation degree of each region to be recommended based on the objects to be recommended in each region to be recommended and the intended objects in each region to be recommended;

[0015] The determining module is further configured to determine the advertising strategy for the stores to be recommended in each of the recommended regions based on the regional recommendation degree of each region to be recommended.

[0016] This application provides an electronic device, including:

[0017] Memory is used to store executable instructions or computer programs.

[0018] When a processor executes computer-executable instructions or computer programs stored in the memory, it implements the information recommendation method provided in the embodiments of this application.

[0019] This application provides a computer-readable storage medium storing a computer program or computer-executable instructions, which, when executed by a processor, implements the information recommendation method provided in this application.

[0020] This application provides a computer program product, including a computer program or computer executable instructions, which, when executed by a processor, implements the information recommendation method provided in this application.

[0021] The embodiments of this application have the following beneficial effects:

[0022] By identifying multiple potential recommendation areas for stores, we can clarify their potential market scope in terms of geographical space. This allows for targeted analysis of these areas, avoiding indiscriminate promotion and improving resource utilization efficiency. Next, by identifying the target audience for each store, we can determine the target customer type, providing clear guidance for evaluating the recommendation value of different areas and developing advertising strategies. Then, by acquiring access data for each target area, we can gain a deeper understanding of the relevant data and identify the potential customers for each area, improving recommendation accuracy. Finally, by comprehensively considering the target audience and potential customers for each area, we can determine the regional recommendation level, quantifying the potential value of different areas for stores and providing crucial reference for developing advertising strategies. Finally, by determining advertising strategies based on regional recommendation levels, we can optimize the allocation of advertising resources and develop personalized advertising strategies for each area, achieving precise ad targeting and improving advertising effectiveness and recommendation efficiency. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the architecture of the information recommendation system 100 provided in an embodiment of this application;

[0024] Figure 2 This is a schematic diagram of the structure of the electronic device 500 provided in the embodiments of this application;

[0025] Figure 3 This is a flowchart illustrating the information recommendation method provided in an embodiment of this application;

[0026] Figure 4 This is a flowchart illustrating the information recommendation method provided in an embodiment of this application;

[0027] Figure 5 This is a flowchart illustrating the information recommendation method provided in an embodiment of this application;

[0028] Figure 6 This is a flowchart illustrating the information recommendation method provided in an embodiment of this application;

[0029] Figure 7 This is a flowchart illustrating the information recommendation method provided in an embodiment of this application;

[0030] Figure 8 This is a flowchart illustrating the information recommendation method provided in an embodiment of this application;

[0031] Figure 9 This is an interactive diagram of the advertising placement function provided in the embodiments of this application;

[0032] Figure 10 This is a logic diagram of the advertising delivery product provided in the embodiments of this application;

[0033] Figure 11 This is a strategy framework diagram for AOI data processing provided in the embodiments of this application;

[0034] Figure 12 This is a logic diagram for the refined exploration of AOI strategies provided in the embodiments of this application;

[0035] Figure 13 This is a logic diagram of the intelligent customer source analysis product provided in the embodiments of this application;

[0036] Figure 14 This is a schematic diagram of the question-and-answer customer source analysis product provided in the embodiments of this application;

[0037] Figure 15 This is a schematic diagram of the intelligent customer source analysis product for advertising placement provided in this application embodiment;

[0038] Figure 16 This is a scene behavior path diagram for effect measurement provided in the embodiments of this application;

[0039] Figure 17 This is a diagram illustrating the logic implementation for mining store visitor demographics provided in this application embodiment;

[0040] Figure 18 This is a technical framework diagram of the fact calculation modification provided in the embodiments of this application;

[0041] Figure 19 These are comparison diagrams showing the effects of the AOI truncation methods provided in the embodiments of this application;

[0042] Figure 20 This is a diagram illustrating the technical solution for AOI value calculation provided in an embodiment of this application;

[0043] Figure 21 This is a distance threshold difference map for intelligent customer source analysis provided in the embodiments of this application;

[0044] Figure 22 This is a technical solution diagram of the AOI sorting and selection strategy provided in the embodiments of this application;

[0045] Figure 23 This is a diagram of the payment measurement attribution technology architecture provided in the embodiments of this application;

[0046] Figure 24 This is a diagram illustrating the attribution principle of in-store measurement provided in an embodiment of this application;

[0047] Figure 25 This is a diagram of the in-store measurement and attribution technology architecture provided in the embodiments of this application;

[0048] Figure 26 This is a logic diagram of the prepayment measurement attribution implementation provided in the embodiments of this application;

[0049] Figure 27 This is a diagram of the prepayment measurement attribution technology architecture provided in the embodiments of this application. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0051] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0052] It is understood that in the embodiments of this application, data related to user information (such as object access data) is involved. When the embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0053] In this application embodiment, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0054] In the following description, the terms “first, second, ...” are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that “first, second, ...” may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0056] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0057] 1) Area of ​​Interest (AOI): An area of ​​interest refers to a geographically defined region on a map that is of interest to people. It can be the spatial range occupied by a specific geographical entity or a virtual area defined according to specific needs. It is usually used to identify and analyze geospatial information related to a specific topic. It can be regarded as a higher-level abstraction of Point of Interest (POI) and consists of polygonal fence boundaries and feature data.

[0058] 2) Location Based Services (LBS): LBS is a value-added service that obtains the location information (geographic coordinates or geodetic coordinates) of mobile terminal users through the wireless communication network of communication operators (such as GSM network, CDMA network) or external positioning methods (such as GPS), and provides corresponding services to users with the support of the Geographic Information System (GIS) platform.

[0059] 3) Post-advertising value: refers to the lasting effects of advertising, which are different from the immediate metrics of exposure and clicks, in triggering actual consumer purchase behavior and improving brand loyalty.

[0060] 4) Price Adjustment Exploration: Price adjustment exploration is when enterprises or merchants purposefully adjust the prices of goods or services and observe the market feedback, changes in sales data, changes in consumer behavior, etc., resulting from these price changes, thereby providing a basis for the formulation or optimization of subsequent pricing strategies.

[0061] 5) Location-based advertising (LBA): This is a marketing method that uses advanced positioning technology to accurately push relevant advertising content to users based on their geographical location information.

[0062] In local industries (including catering and entertainment, supermarkets and retail, real estate and home decoration, housekeeping and repair, travel and car rental, business services, etc.), merchants will comprehensively consider factors such as the surrounding business ecosystem, consumer circles, and traffic flow when choosing a location. Geographical location is an important feature that determines the overall value of customer traffic. These industries often use offline stores as the core to radiate to the surrounding area's customer base and use an online-offline integrated business model to improve operational efficiency. That is, the online platform serves as a window for information display and traffic generation, while the offline stores provide actual products or services.

[0063] Therefore, effectively integrating real-world geographic information into online promotions is key to improving advertising efficiency and long-term value. Among related technologies, location-based advertising already takes user location into account, pushing ads to nearby stores by obtaining the GPS or IP address of the user's mobile device. Subsequently, measuring the conversion effect of ads combined with geographic location information has become a new technological focus. Regarding advertising performance attribution, traditional attribution methods can only detect online click behavior and cannot correlate with offline consumption data. In related technologies, the emergence of localized group-buying coupons has broken down the spatial barriers between online and offline, allowing online conversions to be statistically analyzed through offline redemption behavior. Furthermore, some platforms utilize data mining and analysis techniques to correlate the redemption and usage of online coupons with offline consumption scenarios. For example, by analyzing information such as the time and location of coupon usage and matching it with transaction data in the merchant's sales system, the offline consumption conversion effect brought about by online promotion can be determined.

[0064] Therefore, in terms of advertising recommendations, existing LBA technology essentially relies on geofencing, which can detect customer visit data and push ads based on this data. However, it cannot dynamically identify users' frequent locations (such as residences and workplaces), easily leading to invalid exposures (e.g., pushing localized services that are not immediately needed by users during business trips or commutes). At the same time, the use of LBA technology is still relatively limited in terms of data application dimensions, lacking quantitative analysis of users' true intentions, thus making the conversion efficiency of recommendations less guaranteed.

[0065] Regarding attribution for effectiveness measurement, while offline redemption data can cover some offline scenarios, it is limited to conversions made through group-buying coupons, resulting in limited coverage. Furthermore, there is a time lag between online coupon redemption and offline redemption, during which variables may arise. For example, consumers may abandon using coupons for various reasons, and the relevant technology struggles to capture this data promptly and adjust the assessment of advertising conversion effectiveness, leading to biased results. In other in-store consumption scenarios, the actual conversion scenario is offline, and evaluating conversion effectiveness relies on offline behavioral data returned by advertisers. However, 90% of local merchants lack the technical capability to independently report and return data, which can easily lead to a break in the attribution chain in local scenarios dominated by semi-closed-loop transactions, distorting the actual value assessment of the advertising.

[0066] Meanwhile, from an industry perspective, local industries exhibit a significant long-tail effect and a high proportion of small and medium-sized businesses. These businesses may lack a deep understanding of the regional value around their stores and be unable to effectively utilize regional characteristics, resulting in pain points such as a lack of targeting in ad settings, unreasonable ad bids, and failure to reach core potential users. There is a lack of targeted guidance tools in related technologies.

[0067] Based on this, embodiments of this application provide an information recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product, which can effectively combine regional characteristics to achieve precise advertising delivery. The information recommendation method provided in these embodiments can be implemented by a terminal alone, or by a server and a terminal working together. This will be described below.

[0068] The information recommendation method provided in this application can be applied to various scenarios that require advertising, such as the following scenarios:

[0069] 1) Commercial Promotion Scenarios: Taking a fashion and beauty store as an example, the server collects information on the target areas for this store, which are multiple bustling commercial districts in the city. The target audience is young women aged 18-35, and the target audience for advertising is young women in these areas. Based on this information, the server develops different advertising strategies for each target area and sends these strategies to partner terminals, such as advertising management terminals on major social media platforms and the backend management terminals of beauty apps. The terminals display the received advertising strategies for operators to view and execute. Taking an electronics store as an example, the server analyzes and learns that the target areas for electronics stores are mainly around technology parks and universities, with the target audience being technology enthusiasts and students. The target audience for advertising is the population around technology parks. The server develops advertising strategies highlighting the high-tech performance and innovative features of the products for the target areas around technology parks and transmits these strategies to online advertising platform terminals and electronic bulletin board management terminals in schools. The terminals display the received advertising strategies for relevant personnel to implement advertising according to the strategies.

[0070] 2) Tourism Service Scenarios: Taking seaside resort hotels as an example, the server identifies the recommended areas for these hotels as surrounding cities, cities along high-speed rail lines, and cities with a high concentration of large enterprises. The target audience includes families, couples, and corporate teams with vacation needs. The target audience data consists of people in the surrounding area. The server executes advertising strategies tailored to the specific target audience for each recommended area and transmits these strategies to advertising push terminals. The terminals then display the received advertising strategies for precise promotion. Taking historical and cultural scenic spots as an example, the server identifies the recommended areas for these scenic spots as areas with a high concentration of primary and secondary schools and universities in surrounding cities, as well as cities with active communities of history and culture enthusiasts. The target audience includes students and history and culture lovers. The target audience consists of people in the vicinity of the area. The server develops corresponding advertising strategies for recommended areas with a high concentration of schools and transmits these strategies to promotional terminals and advertising management terminals. The terminals then display the received advertising strategies to facilitate relevant personnel in implementing promotional activities.

[0071] 3) Education Enrollment Scenarios: Taking skills training institutions as an example, the server analyzes and determines that the recommended areas for skills training institutions are commercial areas, industrial areas, and areas surrounding universities in the city. The target audience is unemployed college students, employed individuals, and the unemployed. The target audience is the people in the area. The server formulates different advertising strategies for different recommended areas and transmits these strategies to advertising terminals and information publishing terminals. The terminals display the received advertising strategies to facilitate enrollment promotion for the training institutions. Taking art training schools as an example, the server analyzes and determines that the recommended areas for art training schools are cultural and art districts, high-end residential areas, and areas with a high concentration of children's activity centers in the city. The target audience is children and their parents, and art enthusiasts. The target audience is the people near the school. The server formulates different advertising strategies for different recommended areas and transmits these strategies to children's education platform terminals and information push terminals. The terminals display the received advertising strategies to help art training schools carry out enrollment work.

[0072] For example, see Figure 1 , Figure 1 This is a schematic diagram of the architecture of the information recommendation system 100 provided in this application embodiment. To support an information recommendation application, such as... Figure 1 As shown, the information recommendation system 100 includes: server 200, network 300, and terminal 400. Terminal 400 is connected to server 200 through network 300. Network 300 can be a local area network or a wide area network, or a combination of both.

[0073] In some embodiments, terminal 400 first obtains multiple recommended areas of the stores to be recommended, obtains the target audience of the stores to be recommended, and obtains the object access data for each recommended area; then, terminal 400 transmits the obtained data to server 200 through network 300. Server 200 determines the advertising strategy for each recommended area for the stores to be recommended based on the received multiple recommended areas of the stores to be recommended, the target audience of the stores to be recommended, and the object access data for each recommended area; then, server 200 transmits the advertising strategy for each recommended area for the stores to be recommended to terminal 400 through network 300 and displays it on terminal 400.

[0074] Example, Figure 1 The server 200 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal 400 can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, in-vehicle terminal, etc., but is not limited to these. The terminal 400 and server 200 can be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment.

[0075] The structure of the electronic device provided in the embodiments of this application will be further described below. Taking the electronic device as a terminal as an example, see... Figure 2 , Figure 2 This is a schematic diagram of the structure of the electronic device 500 provided in the embodiments of this application. Figure 2 The illustrated electronic device 500 includes at least one processor 510, a memory 550, at least one network interface 520, and a user interface 530. The various components in the electronic device 500 are coupled together via a bus system 540. It is understood that the bus system 540 is used to implement communication between these components. In addition to a data bus, the bus system 540 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 2 The general labeled all buses as Bus System 540.

[0076] The processor 510 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0077] User interface 530 includes one or more output devices 531 that enable the presentation of media content, including one or more speakers and / or one or more visual displays. User interface 530 also includes one or more input devices 532, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0078] The memory 550 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state storage, hard disk drives, optical disk drives, etc. The memory 550 may optionally include one or more storage devices physically located away from the processor 510.

[0079] The memory 550 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), and the volatile memory may be random access memory (RAM). The memory 550 described in this application embodiment is intended to include any suitable type of memory.

[0080] In some embodiments, memory 550 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, as illustrated below.

[0081] Operating system 551 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks;

[0082] The network communication module 552 is used to reach other computing devices via one or more (wired or wireless) network interfaces 520, exemplary network interfaces 520 including: Bluetooth, WiFi, and Universal Serial Bus (USB), etc.

[0083] Presentation module 553 is configured to enable the presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 531 (e.g., a display screen, a speaker, etc.) associated with user interface 530;

[0084] The input processing module 554 is used to detect and translate one or more user inputs or interactions from one or more input devices 532.

[0085] In some embodiments, the apparatus provided in this application can be implemented in software. Figure 2 An information recommendation device 555 stored in memory 550 is shown. This device can be software in the form of programs and plug-ins, and includes the following software modules: an acquisition module 5551, a determination module 5552, and a processing module 5553. These modules are logically linked and can therefore be arbitrarily combined or further separated according to the functions implemented. It should be noted that... Figure 2 For ease of explanation, all the above modules are shown at once, but this should not be interpreted as excluding the implementation of the information recommendation device 555 which may only include the acquisition module 5551 and the determination module 5552. The functions of each module will be described below.

[0086] The information recommendation method provided in this application will be specifically described below with reference to exemplary applications and implementations of the terminals provided in the embodiments of this application.

[0087] See Figure 3 , Figure 3 This is a flowchart illustrating the information recommendation method provided in the embodiments of this application, which will be combined with... Figure 3 The steps shown are explained.

[0088] It should be noted that, Figure 3 The method illustrated can be executed by various forms of computer programs running on the terminal. These can be dedicated information recommendation clients (e.g., shopping apps), scripts, and mini-programs, or clients that integrate information recommendation software modules, such as instant messaging software on computers or mobile phones (which recommend related articles, advertisements, etc., based on the user's reading preferences and friend interactions) or lifestyle apps (used to recommend advertisements to users). Therefore, the client-based examples used below should not be considered as limiting the embodiments of this application. Furthermore, for ease of description, no specific distinction will be made between the terminal and the client running on the terminal below.

[0089] In step 101, multiple recommended areas of the stores to be recommended are obtained.

[0090] It should be noted that multiple pre-set recommended areas for stores can be obtained directly; alternatively, multiple recommended areas for stores can be obtained by further processing the obtained historical data. For example, based on the store's historical operating data, historical sales data, and customer origin addresses, the store's main customer distribution area can be determined and used as the store's recommended areas. The method for obtaining the recommended areas for stores can be determined according to the actual situation and is not specifically limited here.

[0091] In some embodiments, see Figure 4 , Figure 4 This is a flowchart illustrating the information recommendation method provided in the embodiments of this application, such as... Figure 4 As shown, Figure 3 Step 101 shown can be achieved through... Figure 4 Steps 1011 to 1012 shown are implemented, and will be combined with Figure 4 The steps shown are explained.

[0092] In step 1011, multiple candidate regions are obtained.

[0093] It should be noted that the multiple candidate regions obtained can be related to the stores to be recommended, i.e., the distribution area of ​​the main customer base of the stores to be recommended; or they can be unrelated to the stores to be recommended, such as obtaining multiple pre-defined regions from a spatial information data warehouse. No specific limitation is made here.

[0094] As an example, a standardized spatial information data warehouse can be established based on the full map data (including administrative division codes, latitude and longitude coordinates, POI attribute labels, etc.) provided by the map service provider. Multiple candidate areas can be extracted from the standardized spatial information data warehouse. For example, if the store to be recommended is located in City A, multiple candidate areas in City A can be extracted from the standardized spatial information data warehouse.

[0095] In step 1012, the multiple candidate areas are filtered based on the location of the store to be recommended, resulting in multiple areas to be recommended for the store to be recommended.

[0096] It should be noted that multiple candidate areas can be filtered according to pre-set filtering rules. For example, a region range can be pre-set, and candidate areas within the region range can be filtered out as multiple recommended areas for stores.

[0097] In summary, obtaining multiple candidate regions provides a wider range of market choices for stores to be recommended. Filtering based on store location can further clarify the target market scope of the store, helping to accurately reach target customers, thereby improving marketing effectiveness and reducing marketing costs.

[0098] In some embodiments, see Figure 5 , Figure 5 This is a flowchart illustrating the information recommendation method provided in the embodiments of this application, such as... Figure 5 As shown, Figure 4 Step 1012 shown can be achieved through... Figure 5 At least one implementation of steps 10121 and 10122 shown will be combined Figure 5 The steps shown are explained.

[0099] In step 10121, based on the location of the store to be recommended, the administrative region to which the store to be recommended belongs is determined, and a candidate region matching the administrative region to which the store to be recommended is selected from the multiple candidate regions as the recommended region for the store to be recommended.

[0100] As an example, suppose the candidate regions include region A, region B, region C, and region D, and store A is located in city A. Through matching, it is found that region A and region B are located in city A, and region C and region D are located in city B. Then, region A and region B are selected as the regions to be recommended for store A.

[0101] In step 10122, a first region is determined with the location of the store to be recommended as the center and a set distance as the radius, and multiple candidate regions located in the first region are used as multiple recommendation regions for the store to be recommended.

[0102] As an example, suppose the candidate areas include area A, area B, area C and area D. Let the location of store A be A, and set the distance to be 20 kilometers. Let point A be the center and set the distance as the radius to determine the first area. Suppose that area A is located within the first area, area B intersects with the first area and there is a partial overlap area B1, and areas C and D are located outside the first area. Then, area A and part of area B1 of area B will be the areas to be recommended for store A.

[0103] In summary, selecting candidate areas that match the administrative region of the store to be recommended can better meet the needs and preferences of local consumers and better integrate into the local business ecosystem, achieving complementary market demands. Determining the first area by using the location of the store to be recommended as the center and setting the distance as the radius can target potential customers within a certain range around the store as the target group, which can improve the accuracy of advertising, reach target customers more precisely, improve marketing effectiveness, and reduce marketing costs.

[0104] In step 102, the potential customers of the stores to be recommended are obtained.

[0105] It should be noted that the target audience for the recommended stores refers to individuals or groups who have a potential willingness to purchase or show a certain interest in the products and services offered by the current stores.

[0106] In some embodiments, see Figure 6 , Figure 6 This is a flowchart illustrating the information recommendation method provided in the embodiments of this application, such as... Figure 6 As shown, Figure 3 Step 102 shown can be achieved through... Figure 6 An implementation of at least one of steps 1021, 1022, and 1023 shown will be combined with Figure 6 The steps shown are explained.

[0107] In step 1021, the advertising conversion targets of the stores to be recommended are obtained as the intended targets of the stores to be recommended.

[0108] It should be noted that the advertisements for the stores to be recommended include various types of advertisements such as cost per click (CPC) ads, cost per mille (CPM) ads, and optimized cost per X (oCPX) ads, without any specific limitation here; the advertising conversion audience refers to the target audience who, during the advertising process, are influenced by the advertisement and complete the specific behavior (such as making a purchase in-store) expected by the advertiser.

[0109] As an example, suppose store A advertises on local service platforms and social media. The local service platform can record user actions such as clicking on ads, saving them, and claiming coupons. For instance, if user A sees store A's ad, clicks to view product details and opening hours, and claims a 20% discount coupon, then user A can be considered a potential conversion target. Similarly, if store A's ad on social media receives likes, comments, and shares from user B, with comments expressing interest in the product such as "Looks great, definitely going to try it this weekend," then user B can also be considered a potential conversion target. Therefore, users A and B can be considered potential customers for store A.

[0110] In step 1022, the online conversion targets of the stores to be recommended are obtained as the intended targets of the stores to be recommended.

[0111] It should be noted that online conversion targets refer to the target group in the internet environment that transforms from a potential state to a specific behavioral state desired by marketers (such as adding products to the shopping cart, completing registration, etc.); the online activity scenarios and scope are based on internet platforms, such as various websites, social media, e-commerce platforms, online advertising channels, mini-programs, etc.

[0112] As an example, suppose store B conducts online marketing campaigns on e-commerce platforms and social media. User C adds several new clothes to their shopping cart on the e-commerce platform but hasn't completed the order payment yet. User C can be considered an online conversion target. User D follows store B's official social media account, likes and saves several posts, and asks for product links in the comments. User D can also be considered an online conversion target. Therefore, users C and D can be considered potential customers for store B.

[0113] In step 1023, the offline visitors of the stores to be recommended are obtained as the intended visitors of the stores to be recommended.

[0114] It should be noted that offline visitors refer to users who visit the vicinity of the store, users who visit similar stores, and users who visit the store directly.

[0115] As an example, taking store C as an example, if user E visits store C directly, then user E can be considered an offline visit target; if user F makes a purchase near store C, then user F can also be considered an offline visit target; if user G visits store D, which is similar to store C, then user G can also be considered an offline visit target. Therefore, users E, F, and G can be considered as potential customers of store C.

[0116] In summary, by acquiring advertising conversion targets, we can accurately identify potential customers with a high willingness to buy. Using them as target audiences can improve the efficiency of marketing resource utilization and avoid indiscriminate promotion. By acquiring online conversion targets, we can break through geographical limitations and reach more potential customers, thereby expanding the marketing scope. By acquiring offline visit targets, we can collect more accurate and detailed market information. Using them as target audiences can promote immediate consumption and bring more potential customers to the store.

[0117] In step 103, object access data for each of the regions to be recommended is obtained, and based on the object access data for each of the regions to be recommended, the objects to be recommended for each region are determined.

[0118] It should be noted that the objects to be recommended in the recommended area refer to individuals or groups within a specific recommended area who are likely to participate in the interaction, based on relevant access data and other information. The objects corresponding to the access data can be directly used as the objects to be recommended in the recommended area, or the objects corresponding to the access data can be filtered to obtain the objects to be recommended in the recommended area. No specific limitation is made here.

[0119] As an example, suppose city A is divided into region A, region B, region C, and region D. By counting the user consumption records in each region, we can obtain the access data for each region. For example, if the access data for region A includes 100 users, then we can directly use the 100 users in region A as the recommended users for that region. Alternatively, based on the access data for region A, we can filter out users with more than 10 consumption records as the recommended users for region A.

[0120] In step 104, the regional recommendation degree of each region to be recommended is determined based on the objects to be recommended in each region to be recommended and the intended objects in each region to be recommended.

[0121] It should be noted that regional recommendation score is an indicator that measures the recommendation value of a region. It reflects the recommendation value of the region for a store. The higher the regional recommendation score, the better the effect of the store's advertising in that region, and the higher the conversion rate.

[0122] In some embodiments, step 104 described above can be implemented by performing the following processing for each region to be recommended: determining a first number of objects to be recommended in the region; determining a second number of potential objects in the region; and obtaining a region recommendation score that is negatively correlated with the first number and positively correlated with the second number. Thus, by calculating the corresponding region recommendation score by separately counting the first number of objects to be recommended and the second number of potential objects in each region, the recommendation value of each region can be quantified, allowing stores to allocate marketing resources more rationally and achieve precise targeting of marketing resources.

[0123]

[0124] Where P is the regional recommendation degree of the region to be recommended, N is the first number of objects to be recommended in the region to be recommended, and n is the second number of objects to be recommended in the region to be recommended.

[0125] As an example, based on the access data statistics, the first number of objects to be recommended in region A is n1, and the second number of objects intended for region A is N1. According to the formula (1) above, the regional recommendation degree of region A can be calculated as follows:

[0126] In step 105, based on the regional recommendation level of each region to be recommended, an advertising strategy for each region to be recommended is determined for the stores to be recommended.

[0127] In some embodiments, see Figure 7 , Figure 7 This is a flowchart illustrating the information recommendation method provided in the embodiments of this application, such as... Figure 7 As shown, Figure 3 Step 105 shown can be achieved through... Figure 7 Steps 1051 to 1052 shown are implemented, and will be combined with Figure 7 The steps shown are explained.

[0128] In step 1051, based on the regional recommendation degree of each region to be recommended, the multiple regions to be recommended are sorted in descending order to obtain a sequence of regions to be recommended.

[0129] As an example, suppose the regions to be recommended include regions A, B, C, and D. The region recommendation score of region A is 0.7, that of region B is 0.4, that of region C is 0.6, and that of region D is 0.5. Sort the regions to be recommended in descending order according to their region recommendation scores, and the sequence of regions to be recommended is {region A, region C, region D, region B}.

[0130] In some embodiments, step 1051 described above can be implemented as follows: Obtain regions with a regional recommendation score greater than zero from a plurality of regions to be recommended; sort the regions with a regional recommendation score greater than zero in descending order based on the regional recommendation score to obtain a descending order sorting result; when the number of regions to be recommended included in the descending order sorting result is less than a first quantity threshold, use the descending order sorting result as the sequence of regions to be recommended; when the number of regions to be recommended included in the descending order sorting result is not less than the first quantity threshold, truncate the descending order sorting result based on the first quantity threshold to obtain the sequence of regions to be recommended. Thus, by sorting and filtering the regions to be recommended to obtain a sequence of regions to be recommended, and by conducting advertising and marketing activities based on this sequence, potential customers can be reached more accurately, thereby improving the accuracy and effectiveness of marketing.

[0131] As an example, suppose the regions to be recommended include regions A, B, C, D, E, F, and G. Region A has a recommendation score of 0.2, region B has a recommendation score of 0.3, region C has a recommendation score of 0, region D has a recommendation score of 0.6, region E has a recommendation score of 0.4, region F has a recommendation score of 0.8, and region G has a recommendation score of 0.7. The first threshold is 5. First, regions with a recommendation score greater than zero are filtered out. Then, the filtered regions are sorted in descending order of recommendation score, resulting in the descending order of {region F, region G, region D, region E, region B, region A}. The number of regions to be recommended in the descending order is 6, which is greater than the first threshold of 5. Therefore, the first 5 regions to be recommended are selected from the descending order as the sequence of regions to be recommended, i.e., {region F, region G, region D, region E, region B}.

[0132] In step 1052, the advertising strategy for each of the recommended regions in the recommended region sequence is determined for the recommended store.

[0133] It should be noted that the advertising strategy for a region to be recommended can be obtained by mapping the region's recommendation score. For example, if the region's recommendation score is 0.1, then strategy 1 corresponds to a region with a recommendation score of 0.1, and the advertising strategy for region A is strategy 1. Alternatively, the advertising strategy for a region to be recommended can be determined by its position in the sequence of regions to be recommended. For example, if region B is in the top 10% of the sequence, then strategy 2 corresponds to the top 10% of regions to be recommended, and the advertising strategy for region B is strategy 2.

[0134] In some embodiments, when the advertising strategy is a bidding strategy targeting the stores to be recommended, step 1052 above can be implemented as follows: The sequence of regions to be recommended is hierarchically divided based on the stratification ratio to obtain the stratum to which each region belongs; for each region to be recommended, the bidding strategy corresponding to the stratum to which the region belongs is used as the advertising strategy for that region targeting the stores to be recommended. Thus, by dividing the regions into strata according to the stratification ratio, regions to be recommended at different value levels can be clearly distinguished. Using different advertising strategies for regions at different value levels can achieve a reasonable allocation of advertising resources and improve resource utilization efficiency.

[0135] As an example, suppose the sequence of regions to be recommended includes 400 regions. The sequence of regions to be recommended is divided into layers according to the layer ratios of 10%, 40%, and 50%. Advertising strategy 1 (e.g., increasing advertising costs by 40%) can be implemented for the first layer (the top 40 regions), advertising strategy 2 (e.g., increasing advertising costs by 20%) can be implemented for the second layer (the 41st to the 200th regions), and advertising strategy 3 (e.g., reducing advertising costs by 10%) can be implemented for the third layer (the bottom 200 regions).

[0136] In some embodiments, the intended objects of the region to be recommended have multiple object types, and the regional recommendation score of the region to be recommended includes the regional sub-recommendation score corresponding to each object type. See [link to relevant documentation] Figure 8 , Figure 8 This is a flowchart illustrating the information recommendation method provided in the embodiments of this application, such as... Figure 8 As shown, Figure 7 Step 1052 shown can be achieved through... Figure 8 Steps 10521 to 10522 shown are implemented by combining Figure 8 The steps shown are explained.

[0137] In step 10521, the following processing is performed for each of the regions to be recommended: obtaining the weight corresponding to each of the object types; based on the weight of each object type, performing a weighted summation of the sub-recommendation scores of the regions to be recommended and the multiple regions corresponding to the multiple object types to obtain the new region recommendation score of the region to be recommended.

[0138] As an example, suppose the object types in the region to be recommended include type A and type B, with type A having a weight of 0.6 and type B having a weight of 0.4. The number of potential objects for type A is 10, and the number of potential objects for type B is 15. The number of objects to be recommended for type A is 40, and the number of objects to be recommended for type B is 30. By calculation, the sub-recommendation degree of type A is 0.25, and the sub-recommendation degree of type B is 0.5. Through weighted summation, the new regional recommendation degree of the region to be recommended is 0.35.

[0139] In step 10522, based on the new regional recommendation degree of each of the regions to be recommended in the sequence of regions to be recommended, the advertising strategy for each of the regions to be recommended in the sequence of regions to be recommended is determined for the stores to be recommended.

[0140] It should be noted that the advertising strategy for the region to be recommended can be obtained by mapping the recommendation level of the new region to the region to be recommended, or the advertising strategy for the region to be recommended can be determined by the position of the region to be recommended in the sequence of regions to be recommended. No specific limitation is made here.

[0141] In some embodiments, step 10522 described above can be implemented as follows: adjusting the sequence of regions to be recommended based on the new regional recommendation level to obtain an adjusted sequence of regions to be recommended; determining a first distance threshold corresponding to the region to be recommended based on the location of the intended object in the region to be recommended; filtering the adjusted sequence of regions to be recommended based on the first distance threshold to obtain a filtered sequence of regions to be recommended; and determining an advertising placement strategy for each region to be recommended in the filtered sequence of regions to be recommended for the stores to be recommended based on a preset strategy. Thus, adjusting the sequence of regions to be recommended based on the new regional recommendation level allows the sequence to reflect the latest value of each region in a timely manner, thereby more accurately targeting advertising resources to regions with higher current value, avoiding resource waste in regions with declining value, improving resource utilization efficiency, and achieving optimal resource allocation.

[0142] As an example, suppose the new regional recommendation score for region A is 0.5, for region B it is 0.4, and for region C it is 0.6, and the sequence of regions to be recommended is {region C, region B, region A}. Adjusting the order of the regions to be recommended based on the new regional recommendation scores yields the adjusted sequence of regions to be recommended as {region C, region A, region B}. Next, based on the location information of the target audience, a distance threshold for the regions to be recommended is determined. This can be the furthest distance within the target audience's coverage area, or the distance corresponding to the main concentrated area of ​​the target audience; no specific limitation is made here. Assuming a distance threshold d is obtained, comparison shows that region C and region B are within the distance threshold, while region A is outside the distance threshold. Therefore, the filtered sequence of regions to be recommended is {region C, region B}. Finally, according to a preset strategy, the advertising strategy for each region in the filtered sequence of regions to be recommended is determined through matching. For example, the advertising strategy for region C is strategy 1, and the advertising strategy for region B is strategy 2.

[0143] In some embodiments, the above-mentioned determination of the distance threshold corresponding to the area to be recommended based on the location of the intended object in the area to be recommended can be achieved in the following way: For each intended object of the object type, the following processing is performed: Based on the location of the intended object of the object type in the area to be recommended, a first sub-recommendation area for the store to be recommended is determined; based on the proportion of intended objects of the object type, the distance quantile corresponding to the proportion of the number of the intended objects is determined; starting from a preset distance quantile, the difference between adjacent distance quantiles is calculated sequentially; if the difference meets a preset condition, the distance corresponding to the smallest distance quantile among the adjacent distance quantiles is taken as the distance threshold of the first sub-recommendation area for the store to be recommended; wherein, the preset condition includes the difference between adjacent distance quantiles being greater than a first preset value; based on the weight of each object type, the distance thresholds of multiple first sub-recommendation areas corresponding one-to-one with multiple object types are weighted and summed to obtain the first distance threshold of the area to be recommended. In this way, by determining the first sub-recommendation area based on the location of the intended object type within the area to be recommended, the geographical range that is attractive to a specific object type can be accurately defined. By calculating the distance quantile and determining the distance threshold based on the difference, the main distribution range of potential customers can be determined. By marketing only to areas within a suitable range from the store, ineffective coverage of people who are too far away and have a low probability of purchasing is reduced, thus improving the accuracy of marketing.

[0144] As an example, suppose the object types in the region to be recommended include type A and type B. The first sub-recommendation region corresponding to the intended object of type A is region M, and the first sub-recommendation region corresponding to the intended object of type B is region N. By statistics, the distance quantiles of region M can be obtained. For example, the distance quantile of 25% corresponds to a distance of 0.5 km, the distance quantile of 50% corresponds to a distance of 1 km, and the distance quantile of 75% corresponds to a distance of 2 km. Assuming the preset distance quantile is 50%, the first preset value is 0.1 km, and the difference between adjacent distance quantiles is 1%. Assuming that adjacent distance quantiles are 54% and 55%, the distance difference between them is 0.15 km. Then, the distance of 1.2 km corresponding to the 54% distance quantile is taken as the first distance threshold of region M. Similarly, the first distance threshold of region N can be calculated as 2.1 km. Assuming that the weight of type A is 0.8 and the weight of type B is 0.2, by weighted summing of the two first distance thresholds according to the weights of the two types, the distance threshold of the region to be recommended can be obtained as 1.38 km.

[0145] In some embodiments, the above-described method of determining the advertising strategy for each of the selected recommended regions in the sequence of recommended regions based on a preset strategy can be implemented in the following ways: obtaining a preset filter type, wherein the preset filter type includes a specific object type and a specific recommended region; selecting recommended regions that meet the preset filter type from the sequence of selected recommended regions, and selecting a preset proportion of recommended regions from the selection results as candidate recommended regions; obtaining a quantile threshold, wherein the quantile threshold is obtained by comparing the number of potential customers, the number of customers to be recommended, and the number of recommended regions corresponding to different quantile intervals; selecting recommended regions with quantiles less than the quantile threshold as recommended regions for the stores to be recommended, and implementing a preset advertising strategy in the recommended regions of the stores to be recommended. In this way, by filtering the areas to be recommended according to preset types, areas with high potential can be identified. From the screening results that meet the preset filtering types, a preset proportion of areas to be recommended are selected as the target areas to be recommended, which further narrows the recommendation range. This helps stores focus on the most valuable and potential areas, thereby improving the accuracy of recommendations and enabling subsequent advertising to reach target customers more effectively.

[0146] It should be noted that the quantile threshold is obtained by comparing the number of potential users, the number of users to be recommended, and the number of regions to be recommended in different quantile intervals. For example, in the 60-70% quantile interval, the number of potential users is 10, the number of users to be recommended is 15, and the number of regions to be recommended is 20. Since these numbers are greater than the number of potential users and the number of users to be recommended, 60% is the quantile threshold.

[0147] As an example, taking a sports equipment store as an example, after market research and the store's location information, a preset filtering type is obtained. The specific target type includes people who love outdoor sports, and the specific areas to be recommended are large parks, hiking trails, and bike paths around the city. The previously filtered sequence of areas to be recommended includes {area A, ..., area Z}. According to the preset filtering type, the sequence of areas to be recommended is filtered again. Assuming that area A is outside the specific areas to be recommended, and that there are fewer people who love outdoor sports in areas B and C, the filtering result obtained according to the preset filtering type is {area D, ..., area Z}. Next, the quantile threshold is obtained. Assuming that the quantile threshold is 60%, the top 60% of the areas to be recommended (area D, ..., area P) in the sequence are selected as the areas to be recommended for the sports equipment store, and the preset advertising strategy is applied in the areas to be recommended for the sports equipment store.

[0148] In some embodiments, when the number of regions to be recommended for a store is less than a quantity threshold, after performing the above steps to designate regions with quantiles greater than the quantile threshold as regions to be recommended for the store, the following processing can also be performed: determining the distance difference between the region to be recommended for the store and other regions to be recommended; and designating other regions with distance differences less than a second distance threshold as regions to be recommended for the store. Thus, by adding other regions with distance differences less than the second distance threshold to the sequence of regions to be recommended, the influence of the store to be recommended can be extended to a wider geographical area, thereby expanding the scope and size of the potential customer base.

[0149] As an example, assuming the quantity threshold is 10, the recommended areas for store A include two areas, area A and area B, which are less than the quantity threshold. Due to later adjustments, a new area C is added. Assuming the second distance threshold is d, if the distance difference between area C and area A or the distance difference between area C and area B is less than the second distance threshold d, then area C will also be included as the recommended area for store A.

[0150] In this way, by obtaining the weight corresponding to each object type and performing a weighted summation of the sub-recommendation scores of regions corresponding to multiple object types, the influence of various object types can be comprehensively considered, thereby more accurately assessing the value of regions and providing a more reliable basis for subsequent decisions. In addition, based on the new regional recommendation score of each region to be recommended, the importance and potential value of the region to the store can be scientifically determined, and differentiated advertising strategies can be adopted based on the new regional recommendation scores of different regions to be recommended, so as to improve the effectiveness and conversion rate of advertising.

[0151] In summary, sorting multiple candidate regions in descending order of regional recommendation level clearly identifies regions with higher potential value for stores, avoiding wasted advertising resources and improving ROI. Then, based on the differences in regional recommendation levels among the ranked candidate regions, a reasonable advertising strategy can be determined for each region, ensuring that resource allocation matches regional value. Thus, employing differentiated advertising strategies based on the regional recommendation levels of different candidate regions can effectively expand brand reach and attract more potential customers.

[0152] In some embodiments, after performing step 105, the following processing may also be performed: obtaining log data of the store to be recommended and object transaction data of the store to be recommended; aggregating the log data and the object transaction data, and using the aggregation result as the payment measurement result corresponding to the advertising placement strategy. Thus, by aggregating objective log data and transaction data to obtain the payment measurement result, a reference basis is provided for adjusting the store advertising placement strategy, thereby optimizing the advertising placement strategy and resource allocation, and ultimately improving the advertising conversion rate.

[0153] It should be noted that the payment metrics are used to indicate the extent to which advertising strategies contribute to user payment behavior.

[0154] As an example, log data for the past 30 days (such as ad impressions, ad click-through rates, ad placement costs, etc.) is obtained from the store's advertising logs, and object transaction data for the past 30 days (such as the number of times users pay in-store, the amount of money paid by users in-store, etc.) is obtained from the store's transaction data. Then, the obtained log data and object transaction data are aggregated to obtain the aggregated results for each ad in the past 30 days, and the aggregated results are used as payment metrics. For example, ad A has a cost of 10,000 yuan in the past 30 days, 1,000 times users pay in-store, and a total amount of money paid by users in-store, etc.

[0155] In some embodiments, after performing step 105, the following processing may also be performed: obtaining log data of the store to be recommended and offline access data of the store to be recommended; performing a first association processing on the offline access data and the log data, and using the obtained first association result as the store visit measurement result corresponding to the advertising placement strategy. Thus, by associating log data and offline access data to obtain the store visit measurement result, an objective reference basis is provided for adjusting store advertising placement strategies and planning marketing activities, avoiding blind decision-making, thereby improving the attractiveness of advertising and the store visit conversion rate.

[0156] It should be noted that the in-store measurement results are used to indicate the extent to which advertising strategies contribute to users' in-store behavior.

[0157] As an example, obtain offline access data of the store (such as in-store visits, visits near the store, etc.) and advertising log data of the store (such as advertising costs, advertising exposure rates, etc.); then, perform a first association processing on the log data and offline access data to obtain the first association result, for example, calculate the store arrival rate based on the number of users exposed by the offline access data and the number of users exposed by the advertisement, and calculate the access cost based on the number of visits driven by the advertisement and the total cost of the advertisement; finally, use the first association result as the store arrival measurement result, for example, the store arrival rate is 70% and the access cost is 50 yuan / person.

[0158] In some embodiments, after performing step 105, the following processing may also be performed: obtaining log data of the store to be recommended and online pre-payment transaction data of the store to be recommended; performing a second correlation processing on the online pre-payment transaction data and the log data, and using the obtained second correlation result as the pre-payment measurement result corresponding to the advertising placement strategy. Thus, by correlating log data and online pre-payment transaction data, the contribution of advertising in guiding users from seeing the advertisement to generating purchase intent and completing pre-payment can be clearly understood. Furthermore, based on the pre-payment measurement result, stores can make targeted adjustments to their advertising placement strategies, thereby optimizing marketing strategies and improving the pre-payment conversion rate of advertisements.

[0159] It should be noted that the prepayment metric is used to quantify the effectiveness of marketing campaigns, advertising, or other business strategies in guiding users to the prepayment stage before formal payment.

[0160] As an example, obtain the store's advertising logs for the past 30 days (such as ad click-through rate, ad exposure rate, ad placement cost, etc.) and online prepayment transaction data from the store's related clients (such as prepayment time and prepayment amount in mini-programs, etc.); then, perform a second association processing on the log data and the online prepayment transaction data to obtain the second association result for each ad in the past 30 days, and use the second association result as the prepayment measurement result, for example, ad A's consumption cost in the past 30 days is 50,000 yuan, and the total prepayment amount by users is 20,000 yuan, etc.

[0161] In some embodiments, taking a fitness studio as an example, multiple recommended areas are determined using Geographic Information System (GIS) data and the studio's catchment area, such as surrounding office building districts, residential areas, and areas near universities. Then, based on the fitness studio's positioning and services, the target audience is identified as people with fitness needs, who want to improve their physical fitness, and who pursue a healthy lifestyle, including office workers, students, and nearby residents. Next, target audience access data is collected from multiple channels, and the collected data for each recommended area is input into a large language model. The large language model analyzes the data for each recommended area to obtain the target audience for each recommended area. The system identifies the target audience for each recommended area. Then, based on the target audience and the intended audience for each area, it calculates the regional recommendation score for each area. Finally, the regional recommendation score for each area is input into a large language model, which generates an advertising strategy for each area of ​​the fitness studio. For example, for office building commercial areas, it suggests using elevator ads within office buildings, pop-up ads in office software, and sponsorship ads for surrounding business activities. The advertising content should emphasize the fitness studio's suitability for office workers and recommend increasing advertising efforts during weekdays before work, lunch breaks, and after get off work. The recommended investment cost is 50,000 yuan.

[0162] The following describes an exemplary application of the embodiments of this application in a real-world application scenario. This exemplary application describes the specific implementation process of the information recommendation method in an advertising delivery scenario.

[0163] In local industries (including catering and entertainment, supermarkets and retail, real estate and home decoration, housekeeping and repair, travel and car rental, business services, etc.), merchants will comprehensively consider factors such as the surrounding business ecosystem, consumer circles, and traffic flow when choosing a location. Geographical location is an important feature that determines the overall value of customer traffic. These industries often use offline stores as the core to radiate to the surrounding area's customer base and use an online-offline integrated business model to improve operational efficiency. That is, the online platform serves as a window for information display and traffic generation, while the offline stores provide actual products or services.

[0164] Therefore, effectively integrating real-world geographic information into online promotions is key to improving advertising efficiency and long-term value. Among related technologies, location-based advertising already takes user location into account, pushing ads to nearby stores by obtaining the GPS or IP address of the user's mobile device. Subsequently, measuring the conversion effect of ads combined with geographic location information has become a new technological focus. Regarding advertising performance attribution, traditional attribution methods can only detect online click behavior and cannot correlate with offline consumption data. In related technologies, the emergence of localized group-buying coupons has broken down the spatial barriers between online and offline, allowing online conversions to be statistically analyzed through offline redemption behavior. Furthermore, some platforms utilize data mining and analysis techniques to correlate the redemption and usage of online coupons with offline consumption scenarios. For example, by analyzing information such as the time and location of coupon usage and matching it with transaction data in the merchant's sales system, the offline consumption conversion effect brought about by online promotion can be determined.

[0165] Therefore, in terms of advertising recommendations, existing LBA technology essentially relies on geofencing, which can detect customer visit data and push ads based on this data. However, it cannot dynamically identify users' frequent locations (such as residences and workplaces), easily leading to invalid exposures (e.g., pushing localized services that are not immediately needed by users during business trips or commutes). At the same time, the use of LBA technology is still relatively limited in terms of data application dimensions, lacking quantitative analysis of users' true intentions, thus making the conversion efficiency of recommendations less guaranteed.

[0166] Regarding attribution for effectiveness measurement, while offline redemption data can cover some offline scenarios, it is limited to conversions made through group-buying coupons, resulting in limited coverage. Furthermore, there is a time lag between online coupon redemption and offline redemption, during which variables may arise. For example, consumers may abandon using coupons for various reasons, and the relevant technology struggles to capture this data promptly and adjust the assessment of advertising conversion effectiveness, leading to biased results. In other in-store consumption scenarios, the actual conversion scenario is offline, and evaluating conversion effectiveness relies on offline behavioral data returned by advertisers. However, 90% of local merchants lack the technical capability to independently report and return data, which can easily lead to a break in the attribution chain in local scenarios dominated by semi-closed-loop transactions, distorting the actual value assessment of the advertising.

[0167] Meanwhile, from an industry perspective, local industries exhibit a significant long-tail effect and a high proportion of small and medium-sized businesses. These businesses may lack a deep understanding of the regional value around their stores and be unable to effectively utilize regional characteristics, resulting in pain points such as a lack of targeting in ad settings, unreasonable ad bids, and failure to reach core potential users. There is a lack of targeted guidance tools in related technologies.

[0168] At the product level, this application applies the advertising post-effect value assessment system (referred to as "AOI value") based on "geographic location information" and "user transaction interests" to the advertiser's placement scenario, and links the underlying advertising bidding strategy to productize this system; at the same time, combined with the big model, the information recommendation method of this application is applied to application scenarios such as advertising placement analysis and placement recommendation. In addition, the effectiveness of advertising placement strategy is measured through the implementation of multi-level attribution, thereby achieving the completeness and systematization of the overall advertising post-effect value optimization method.

[0169] In some embodiments, see Figure 9 , Figure 9 This is an interactive diagram of the advertising placement function provided in the embodiments of this application, such as... Figure 9 As shown, this application uses the advertising placement end as the carrier in product implementation, and builds an "Exploration of High-Value Range" advertising placement strategy function 201 based on "AOI value", which can serve the advertising placement of advertisers and is continuously iterated through advertising data verification and supplementation.

[0170] See here. Figure 10 , Figure 10 This is a logic diagram of the advertising delivery product implementation provided in the embodiments of this application, such as... Figure 10 As shown, advertisers can... Figure 9 On the ad placement platform, enable the function switch corresponding to "High-Value Targeting Strategy" 201, and optionally fill in the "High-Value Range" area, where "High-Value" refers to...

[0171] See also Figure 10 When the advertiser does not fill in the region, the system will filter AOIs within the ad targeting set by the advertiser, and truncate and sort them based on the AOI value calculation. Finally, it will explore price adjustments for the audience in different AOIs based on the advertiser's original bid. When the advertiser manually enters or imports regions in batches, the system will explore price adjustments simultaneously by combining the manually entered regions and the AOIs filtered by ad targeting.

[0172] In some embodiments, see Figure 11 , Figure 11 This is a strategy framework diagram for AOI data processing provided in the embodiments of this application, such as... Figure 11 As shown, the underlying framework for AOI data processing in this application is mainly divided into AOI automated production, industry target audience building, and AOI audience refinement exploration.

[0173] In the AOI automated production module, the basic data for AOIs mainly comes from map data interfaces, and the AOI data is processed as follows: First, AOI audience data tags are generated by combining user visit information; then, filtering is performed based on the targeting range set by the advertisement; finally, AOIs that have been explicitly stated by the advertiser are excluded. This results in a list of AOIs whose value needs to be calculated for each advertisement push.

[0174] In the industry-specific target audience building module, the data sources for target audience mining mainly include users converted from ads, users who opened mini-programs, users reported back by advertisers, and data on visitors to physical stores. Considering the suitability and credibility of target audiences in advertising campaigns, this application uses advertising industry system data to mine and build target audiences according to secondary industry dimensions (e.g., local services - Western fast food, automotive - car manufacturers, etc.).

[0175] In the AOI population refinement exploration module, see next. Figure 12 , Figure 12 This is a fine-grained exploration logic diagram of the AOI strategy provided in the embodiments of this application, such as... Figure 12 As shown, firstly, based on the intersection of AOI data generated by the AOI automated production module and industry target audience data, referring to the above formula (1), the audience concentration of the corresponding high-value audience in different AOIs can be calculated, thus obtaining the "AOI value". Next, the AOIs corresponding to the advertisement are sorted according to the calculated AOI value, and refined exploration is carried out. This is directly applied to the advertisement bid to modify the bid, affecting the bidding ranking and cost accounting. When there is no AOI data, no exploration operation is performed. When there is AOI data, the advertisement is placed. The strategy can filter out the AOIs that are effective. For example, the advertising strategy will not be effective for AOIs with an intent audience concentration of 0. The advertising strategy will only be effective for AOIs with an intent audience concentration of not 0 (i.e., effective AOIs). When there are effective AOIs and the number of effective AOIs is greater than 2000, the AOIs are sorted by weighted order based on the number of audiences and the concentration of high-value audiences. The top 2000 effective AOIs are truncated from high to low for detailed exploration. When there are effective AOIs and the number of effective AOIs is not greater than 2000, all effective AOIs are taken for detailed exploration.

[0176] Here, the specific rules for exploring price adjustments are as follows: Based on the AOI value from high to low, the audience within each AOI is divided into three tiers: 10%, 40%, and 50%. For the high-tier 10% (ranked top 10%), the exploration operation is conducted at 1.4 times the ad bid; for the mid-tier 10% (ranked 10%-50%), the exploration operation is conducted at 1.3 times the ad bid; and for the low-tier 50% (ranked bottom 50%), the exploration operation is conducted at 1.2 times the ad bid. Simultaneously, the audience within the AOIs explicitly defined by the advertiser is explored at 1.5 times the ad bid.

[0177] In some embodiments, LBS-based businesses are highly dependent on location information. Figure 12 The underlying logic of the refined exploration of AOI strategies presented provides a measurement system for evaluating the post-advertising value of geographical areas. This system combines advertising data to measure the potential customer value of the area surrounding a store in the real world. This measurement system can not only automate and refine advertising bidding exploration within the system, but also assist advertisers in manually formulating targeting, bidding, and other advertising execution strategies, and even in analyzing business strategies such as customer acquisition, user activation, and store opening. See also Figure 13 , Figure 13 This is a logic diagram of the intelligent customer source analysis product provided in the embodiments of this application, such as... Figure 13 As shown, this application combines a large language model with merchant store location information to generate intelligent recommendation results for advertisers to analyze customer sources.

[0178] In the process of intelligent question-answering analysis, see Figure 14 , Figure 14 This is a schematic diagram of the question-and-answer customer source analysis product provided in the embodiments of this application, such as... Figure 14 As shown, based on the intent understanding and underlying customer source identification capabilities of the large language model, it can analyze and intelligently recommend the distribution of customers around the store by combining the store address, and realize question-and-answer-style high-value AOI analysis of the store.

[0179] During the intelligent recommendation process on the delivery side, see Figure 15 , Figure 15 This is a schematic diagram of the intelligent customer source analysis product for advertising placement provided in this application embodiment, such as... Figure 15 As shown, when combined with the Advector advertising platform, the ad placement assistant can identify the store locations that advertisers use / associate with their ads in real time. Based on the store locations, it can identify customer sources and intelligently sort and truncate them, provide placement suggestions, and support the "one-click adoption" function. After clicking, the high-value range exploration switch can be turned on, and the recommended AOI will be automatically filled in (which is more store-specific than the system's automatic exploration).

[0180] In some embodiments, see Figure 16 , Figure 16 This is a scene behavior path diagram for effect measurement provided in the embodiments of this application, such as... Figure 16 As shown, to verify the application effect of the AOI strategy, this application builds an effect measurement capability tailored to local business scenarios for online promotion-offline in-store transactions. This makes the measurement of conversion effect of LBS-based advertising closer to the real effect, allowing the optimization effect of this application to be scientifically measured, while helping advertisers understand the true value of advertising for better business operations. The main conversion effect measurement scenarios include three categories: payment measurement, in-store measurement, and pre-payment measurement.

[0181] The payment measurement scenario primarily targets users who, after seeing a promotional ad, do not convert online but instead visit the store and use WeChat Pay to make a purchase. For this scenario, data such as the number of times users scanned the store's payment QR code and initiated electronic payment, the number of users, and the amount spent can be matched with the exposure data of the associated store's ads to measure the ad's effectiveness. The in-store measurement scenario primarily targets users who, after seeing a promotional ad, do not convert online but visit the store directly without using electronic payment. For this scenario, actual offline visit data can be matched with ad exposure data to measure ad effectiveness. The pre-payment measurement scenario primarily targets merchants offering discounts, coupons, etc., and providing mini-program ordering. Online users see the ad, open the mini-program, and initiate a pre-payment page. For this scenario, pre-payment behavior within the mini-program can be matched with ad exposure data to measure ad effectiveness.

[0182] At the technical level, this application is mainly implemented through three main steps: "data source production - application implementation - effect monitoring". In terms of data source production, the value of AOI is derived from the value measurement of geographic location, advertising domain and Tencent domain closed-loop data. Before being applied to the advertising strategy, the mining and production of the corresponding data source need to be automated. In terms of application implementation, the differences in the technical solutions of the two scenarios of advertising bid adjustment and intelligent customer source analysis are mainly considered. At the same time, the effect observation of the advertising strategy of this application needs to be based on the effect measurement of LBS-type advertising.

[0183] In some embodiments, AOI value is a measure of the distribution concentration of high-value people in a specific area. Its essence is to achieve a quantitative assessment of regional commercial value through multi-dimensional data fusion. The comprehensive calculation of AOI value in this application mainly includes two core elements: objective regional traffic value and target population value, namely, the visitors and potential visitors of the AOI.

[0184] In the automated production of AOI data, a pipeline for full-process automated data processing is first constructed to form an integrated analysis framework of "geographic information + object behavior". For offline AOI data, a standardized spatial data warehouse can be established relying on the full amount of AOI geographic data provided by map service providers (including administrative division codes, latitude and longitude coordinates, POI attribute tags, etc.). After obtaining the AOI information, it is necessary to associate advertisements and AOIs through two methods: administrative division targeting or point-radius targeting. Among them, the AOI information table contains the administrative division where the AOI is located, so the AOI information can be filtered by administrative division. It should be noted that in administrative division targeting, filtering may only reach the provincial / municipal dimension. If the administrative division of the AOI is at the district dimension, matching and filtering shall be performed by taking the corresponding number of digits according to the administrative division code. The rule for point-radius targeting is to restrict the area defined by a preset point (lat, lon) and a preset radius (200 meters < r < 25 kilometers). The AOI information table contains AOI coordinates, and filtering can be performed by comparing the relationship between the point coordinates and the area. Next, based on the GPS / WiFi positioning data stream of the user's mobile terminal combined with map information, the corresponding AOI visit data is obtained. The current AOI visit tags have covered more than 1.5 million AOI areas for 900 million users. On the basis of the visit data, a spatiotemporal clustering algorithm with a sliding time window (for example, 90 days) is adopted to statistically analyze the top 5 AOIs with the highest user visit frequency, and semantic annotation is performed based on the AOI classification system (for example, "business office category", "catering service category", etc.). After obtaining the above information, the user and AOI information are written into the online database to support millisecond-level real-time queries, providing data support for value calculation and bidding optimization.

[0185] In the production of intended crowd mining, the limitations of traditional advertisement conversion data are broken through, and richer crowd data is integrated by leveraging data advantages, which mainly includes advertisement conversion data, advertiser first-party data, mini-program conversion data, and store visit data. Here, the intended crowd refers to the crowd that has actually converted.

[0186] The advertisement conversion data refers to the converted crowd of historical advertisements. By considering the differences in object behaviors in different types of advertisements, this application takes the click crowd in CPC and CPM advertisements and the optimized target conversion crowd in oCPX advertisements as the intended crowd, mines the secondary industry and conversion crowd information corresponding to the advertisements according to the above advertisement data screening method, and updates the data on a daily frequency.

[0187] First-party behavioral data refers to user conversion behavior data voluntarily returned by advertisers. In the first-party data table, each user action has a corresponding tag (i.e., action_type). Different industries have different advertising conversion paths, and the corresponding action_types will also differ. Therefore, it is necessary to differentiate the first-party audience by industry, that is, to identify users corresponding to the action tags after the "common optimization goals" of the second-level industry advertising. For example, the common conversion path for the automotive industry is "exposure-click-lead generation-store visit-test drive," with a focus on the "lead generation" goal. The first-party behavioral data would be users whose "store visit" / "test drive" behavior is tagged with the action_type tag, as returned by industry advertisers. Similarly, the common conversion path for the Western fast food industry is "exposure-click-store visit-order," with a focus on "store visit" / "order" as the optimization goal. The first-party behavioral data would be users whose "order" behavior is tagged with the action_type tag, as returned by industry advertisers.

[0188] Mini-program conversion data refers to the opening data of registered mini-programs under the secondary industry. The platform aggregates users' mini-program opening behavior according to the secondary industry of account opening and then sends it to the advertising side as a high-value audience of that secondary industry for weighted calculation.

[0189] Store visit data refers to the data on the number of people visiting a store in person. See [link / reference] Figure 17 , Figure 17 This is a diagram illustrating the logic implementation for mining store visitor demographics provided in this application's embodiments, such as... Figure 17 As shown, in the selection of data sources, considering the limited range of POIs and the very sparse corresponding visit information, recommendations are made by combining more coarse-grained visit behavior, specifically users who visit similar stores and users who visit the vicinity of the stores.

[0190] In some embodiments, the application implementation layer of the technology mainly includes two parts: refined exploration of advertising and intelligent customer source analysis. Both are based on the calculation, filtering, sorting and application of AOI value based on data sources. However, due to the uniqueness of their respective scenarios, there will be certain differences in the details of the technical implementation.

[0191] For further exploration of refined advertising, see [link / reference]. Figure 18 , Figure 18 This is a technical framework diagram of the fact calculation and price modification provided in the embodiments of this application, such as... Figure 18As shown, after comparing various optimization methods, this application uses real-time calculation for price adjustments, mainly including four modules: offline advertising data calculation, advertising data access to the Adtable, audience data access to online platforms, and automated price adjustments. When the advertising strategy takes effect, considering that confidence levels are aggregated and calculated according to the secondary industry of the asset library, based on the target audience mining data results and user-AOI correlation described above, online advertising and AOI data can be linked, and target users can be associated through the CPV category to which the advertising belongs. After completing the advertising calculation, the advertising data needs to be accessed to the advertising data table. That is, by developing a Protocol Buffers (proto) file to define the data structure and format, the specifications for data organization and transmission are determined. Then, a program is written to write the data to the specified storage location. Next, the data team releases the developed proto file. After the data team releases the proto file, code is written to implement the data access function, bringing external data into the system. Code is also written to enable the sorting server to mount and process the advertising data, completing operations such as sorting the advertising data. In addition, audience data also needs to be uploaded online for retrieval and query.

[0192] In some embodiments, see Figure 19 , Figure 19 This is a comparison diagram of the effects of the AOI truncation method provided in the embodiments of this application, such as... Figure 19 As shown, in the selection of AOIs for price adjustments, if only AOI intent concentration is considered for sorting, higher-ranked AOIs tend to have smaller coverage, resulting in a smaller actual coverage of the final advertising strategy. Therefore, this application incorporates two factors: total number of AOI users (affecting the request effectiveness rate) and AOI intent concentration. It truncates the AOIs by total number of users and conducts tiered testing based on AOI intent concentration. Finally, based on the implementation framework, the price adjustment operation for the advertisement is completed.

[0193] For LBS-based intelligent customer source analysis, recommendations are based on finer-grained location information, specifically POI information of store locations, and are not subject to geographical limitations such as advertising targeting. Therefore, this application has made adjustments to the implementation of the AOI strategy for advertising placement, specifically: 1) adjusting the weights of different data sources when calculating AOI value; 2) designing an adaptive distance threshold calculation method; and 3) adjusting the rules for AOI ranking and selection used in recommendation scenarios.

[0194] See Figure 20 , Figure 20 This is a diagram illustrating the technical solution for AOI value calculation provided in an embodiment of this application, such as... Figure 20As shown, offline access data is obtained from fine to coarse granularity according to store POI-grid and same store category-grid, and the finer the granularity, the stronger the correlation with the store; online behavior data includes mini-program behavior data, advertising conversion audience and one-party audience data. Among them, mini-program behavior data is obtained by sorting users of related mini-programs according to the mini-programs associated with the store, and selecting the top N users as data to be processed. Similarly, the top N users in the advertising conversion audience are selected as data to be processed. For one-party audience data, the top N users in the one-party data returned by the advertiser are selected as data to be processed. For the above data sources, this application uses information confidence to measure the amount of information of each granularity of visit / behavior information. Here, the higher the confidence, the higher the amount of visit / behavior information, and the less space needs to be left for the lower priority granularity; conversely, the lower the confidence, the higher the AOI value. The AOI value is calculated according to the data information obtained from different data sources, which can be specifically calculated according to formula (2).

[0195] S i =s i α i +(1-α i )S i+1 (2)

[0196]

[0197] Where, α i The weight (i.e., confidence level) of the AOI from the i-th source can be calculated using formula (3), s i Let represent the user concentration of AOI in the i-th source, and n represent the number of users in the i-th source.

[0198] It should be noted that when querying the data source of a store, the audience of mini-program users and advertising conversions may not belong to that store, and their priority will be reduced. The priority of the above five data sources is: store visits > store advertising conversions > mini-program usage > grid + category visits > grid visits.

[0199] Unlike advertising strategy applications where advertisers manually adjust targeting limits, in intelligent recommendation scenarios, merchants may want to understand the reach of their stores through recommendations. However, in reality, the distribution of AOIs from different categories, cities, and regions may vary, thus requiring different distance thresholds to be set. Furthermore, the distance distribution of AOIs from different data sources also differs (for example, the distance distribution of visiting AOIs and AOIs of people who convert from ads differs greatly).

[0200] Therefore, it is necessary to adaptively learn the optimal distance constraints from the visitor information. First, calculate the visitor distance quantiles. Using the number of visitors as weights, the distance of the AOI is statistically weighted, and the distance quantiles are calculated. For example, the 10% distance quantile means that 10% of the visitors are covered within that distance. Similarly, the distance quantiles from 0% to 100% of the visitor distance can be calculated sequentially. Next, find the optimal distance threshold among the distance quantiles and calculate the difference between two adjacent distance quantiles. This difference is equivalent to the first derivative of the distance quantile. In the actual calculation process, starting from a preset distance quantile position (e.g., the 50% distance quantile, i.e., covering at least 50% of users), the difference between two adjacent differences is calculated. See [link to calculation]. Figure 21 , Figure 21 This is a distance threshold difference map for intelligent customer source analysis provided in the embodiments of this application, such as... Figure 21 As shown, the difference in the distance is equivalent to the second derivative of the distance quantile. A suitable threshold is defined as the distance that meets preset conditions. These conditions include a ratio greater than 3 for the difference between two adjacent distance quantiles and an absolute value of the difference between two adjacent distance quantiles not less than 3 kilometers. This means that a large difference in the distance quantiles, with a sudden increase in the difference distance, may indicate a new outlier AOI. The absolute value of the difference between distance quantiles not less than 3 kilometers is because the absolute values ​​of distance quantiles at the top of the hierarchy may be relatively small, but the ratio of their differences is large, and this data still needs to be retained. Alternatively, the absolute value of the difference between two adjacent distance quantiles not less than 20 kilometers can avoid data omissions caused by large differences in the distance quantiles at the bottom of the hierarchy but insufficient differences in their ratios. Finally, a confidence-weighted calculation is used to obtain the final distance threshold. Furthermore, a global maximum distance is set for different categories and the overall system, and the adaptively calculated distance threshold cannot exceed this global maximum distance.

[0201] See Figure 22 , Figure 22 This is a technical solution diagram of the AOI sorting and selection strategy provided in the embodiments of this application, such as... Figure 22As shown, after calculating the AOI value and adaptive distance threshold, the AOIs that match the store locations associated with LBS-type ads can be obtained. Next, the AOIs need further filtering to determine the final recommended AOIs. The technical path mainly includes sorting, filtering, and selection, specifically including the following strategy rules: Sorting strategy: AOI information from each data source is added to the queue in the order of store visits, store ad conversions, mini-program usage, grid + category visits, and grid visits. Then, the sorting index of the AOI within each granularity is calculated according to preset rules. The preset rules are to calculate the comprehensive concentration... The weighting is calculated as follows: score + number of visitors / behaviors + category priority (the weight of each category is calculated based on the category distribution of the customer source AOI) + proportion of working and living groups; the filtering strategy is based on category (e.g., parking lots, ancillary facilities, tourist scenarios, institutional groups, etc.), area (e.g., industrial parks with unusually large areas other than airports, etc.), and low-quality APIs (e.g., statistical analysis of AOI visitor count and quantiles, using the 95th percentile value (AOI visitor count, number of visitors / behaviors) as the filtering threshold for low-quality AOIs); the selection strategy evaluates the performance gain of each data source AOI to determine whether to continue adding AOIs to be recommended. For example, calculate the concentration score thresholds for the number of visitors, the number of AOIs, and the number of AOIs at the 0%-100% quantiles. In the 10 quantiles from 0% to 100%, compare the values ​​of a (number of visitors), b (number of AOIs), and c (number of AOIs). If c>a and c>b, it means that the quality of the AOIs is relatively low, and there is no need to continue selecting AOIs. Then, c is taken as the lowest score threshold. The API clustering strategy is to exclude new AOIs when there are enough selected AOIs and the distance difference between the new AOI and the already selected AOIs is greater than the distance threshold.

[0202] In some embodiments, for the application of AOI strategies, this application needs to improve the technical methods for measuring the effectiveness of LBS-type advertising in order to verify the effectiveness of AOI strategies in advertising placement exploration.

[0203] See in payment metric attribution. Figure 23 , Figure 23 This is a diagram of the payment measurement attribution technology architecture provided in the embodiments of this application, such as... Figure 23As shown, by collaborating with electronic payment platforms, the system retrieves electronic payment actions and results to measure ad conversion effectiveness. First, advertisers need to authorize the acquisition of merchant account information and data. Next, the account, merchant information, ad impressions, and click data configured on the ad side are sent to the electronic payment side. Subsequently, the electronic payment side uses the above information to attribute the user's payment data to ad impressions / clicks within the past 30 days based on the last click action, generating attribution results aggregated across ad dimensions. Finally, the ad side will organize and display the above attribution results as data reports.

[0204] See in the attribution of in-store visit metrics. Figure 24 , Figure 24 This is a diagram illustrating the attribution principle of in-store measurement provided in an embodiment of this application, such as... Figure 24 As shown, the metrics for attributing store visits include store visit rate and cost per visit; see [link to relevant documentation]. Figure 25 , Figure 25 This is a diagram of the in-store measurement and attribution technology architecture provided in the embodiments of this application, such as... Figure 25 As shown, after advertisers bind the store's location when placing ads, they can obtain the store's LBS information and mine the user's LBS information based on the map to obtain the visiting audience. They can then match the in-store behavior of these users with the exposure / click behavior of online ads to include this converted audience in the conversion effect measurement.

[0205] See in the attribution of prepayment metrics. Figure 26 , Figure 26 This is a logic diagram of the prepayment measurement attribution implementation provided in the embodiments of this application, such as... Figure 26 As shown, prepayment metrics require advertisers to create and authorize data sources. After authorization, the advertising side collaborates with the electronic payment platform to link prepayment behavior with ad impressions / clicks; see also Figure 27 , Figure 27 This is a diagram of the prepayment measurement attribution technology architecture provided in the embodiments of this application, which will be based on... Figure 27 The steps described implement the prepayment measurement process.

[0206] In summary, this application breaks through the traditional advertising effectiveness evaluation paradigm, creatively integrating geographic location coordinates with real-time transaction scenarios in a multi-dimensional way. It constructs a dynamic evaluation system for the post-effect value of AOI (Area of ​​Interest) through weight analysis and target group index calculation, maximizing the value of closed-loop data. Furthermore, by integrating map coordinate location description interfaces with massive amounts of active and in-store user data from mini-programs, it forms a global observation network covering over 3 million AOI locations and radiating over 900 million UV users, achieving spatial mapping between advertising reach and user's actual consumption behavior. This application also overcomes the challenge of real-time processing of hundreds of billions of data points, designing and implementing a binary recommendation strategy framework based on transaction interest and location attributes. Through computational empowerment, it increases the system's response speed from daily to hourly (150 billion computational operations). Simultaneously, based on a large language model-based intent recognition engine (training data of over 1.1 billion UVs) and a customer profiling system, it achieves question-and-answer intelligent recommendation interaction. This application also improves the advertising recommendation experience by making it closer to user decision-making through strategic innovation, thereby enhancing the post-effect. A / B tests have verified that using this strategy, the core advertising metric, Return on Investment (ROI), is significantly improved. Investment (ROI) increased by 7% and effective lead rate increased by 6%. At the same time, a consumption-after-effect dual circulation model with commercial ecosystem significance was established, realizing a win-win situation for both customers and the platform. In addition, this application expanded advertising to general in-store / home delivery scenarios, forming a reusable scenario solution library (e.g., automotive industry: effective lead conversion rate increased by 48%; real estate and home furnishing: lead conversion cost reduced by 57%; local services: order ROI increased by 10%). In terms of intelligent analysis, it combined placement recommendations to help advertisers with their placement, and scenario-based AI agent applications to help merchants with their operations, and continues to explore more customer source analysis scenario innovations.

[0207] The following description continues to illustrate the exemplary structure of the information recommendation device 555 provided in the embodiments of this application as a software module. In some embodiments, such as... Figure 2 As shown, the software modules stored in the information recommendation device 555 in the memory 550 may include: an acquisition module 5551 and a determination module 5552.

[0208] The acquisition module 5551 is used to acquire multiple recommendation areas for stores to be recommended and to acquire the intended targets of the stores to be recommended; the determination module 5552 is used to acquire object access data for each of the recommendation areas and, based on the object access data for each of the recommendation areas, to determine the recommended objects for each of the recommendation areas; the determination module 5552 is also used to determine the regional recommendation degree for each of the recommendation areas based on the recommended objects and the intended targets of each of the recommendation areas; the determination module 5552 is also used to determine the advertising placement strategy for each of the recommendation areas for the stores to be recommended based on the regional recommendation degree for each of the recommendation areas.

[0209] In some embodiments, the acquisition module 5551 is further configured to acquire multiple candidate regions; and filter the multiple candidate regions based on the location of the store to be recommended to obtain multiple recommended regions for the store to be recommended.

[0210] In some embodiments, the acquisition module 5551 is further configured to perform at least one of the following processes: based on the location of the store to be recommended, determine the administrative region to which the store to be recommended belongs, and filter out candidate regions that match the administrative region to which the store to be recommended belongs from the plurality of candidate regions, as the recommended regions of the store to be recommended; determine a first region with the location of the store to be recommended as the center and a set distance as the radius, and use the plurality of candidate regions located in the first region as the plurality of recommended regions of the store to be recommended.

[0211] In some embodiments, the acquisition module 5551 is further configured to perform at least one of the following processes: acquiring the advertising conversion objects of the store to be recommended as the intended objects of the store to be recommended; acquiring the online conversion objects of the store to be recommended as the intended objects of the store to be recommended; acquiring the offline visit objects of the store to be recommended as the intended objects of the store to be recommended.

[0212] In some embodiments, the determining module 5552 is further configured to perform the following processing for each of the regions to be recommended: determining a first number of objects to be recommended in the region to be recommended; determining a second number of intended objects in the region to be recommended; and obtaining a region recommendation degree that is negatively correlated with the first number and positively correlated with the second number.

[0213] In some embodiments, the determining module 5552 is further configured to sort the multiple regions to be recommended in descending order based on the regional recommendation degree of each region to be recommended, to obtain a sequence of regions to be recommended; and to determine the advertising placement strategy for each region to be recommended in the sequence of regions to be recommended for the stores to be recommended.

[0214] In some embodiments, the determining module 5552 is further configured to: obtain regions with a region recommendation degree greater than zero from a plurality of regions to be recommended; perform descending sorting processing on the regions with a region recommendation degree greater than zero to obtain a descending sorting result; when the number of regions to be recommended included in the descending sorting result is less than a first quantity threshold, use the descending sorting result as the sequence of regions to be recommended; when the number of regions to be recommended included in the descending sorting result is not less than the first quantity threshold, perform truncation processing on the descending sorting result based on the first quantity threshold to obtain the sequence of regions to be recommended.

[0215] In some embodiments, when the advertising strategy is a bidding strategy for stores to be recommended, the determining module 5552 is further configured to divide the sequence of regions to be recommended into layers based on the layering ratio of the sequence of regions to be recommended, to obtain the layer to which each region to be recommended belongs; and for each region to be recommended, to use the bidding strategy corresponding to the layer to which the region to be recommended belongs as the advertising strategy for the region to be recommended for the stores to be recommended.

[0216] In some embodiments, the intended objects of the region to be recommended have multiple object types, and the regional recommendation degree of the region to be recommended includes the regional sub-recommendation degree corresponding to each object type. The determining module 5552 performs the following processing for each region to be recommended: obtaining the weight corresponding to each object type; based on the weight of each object type, performing a weighted summation of the regional sub-recommendation degrees corresponding one-to-one with the multiple object types for the region to be recommended, to obtain a new regional recommendation degree for the region to be recommended; and based on the new regional recommendation degree of each region to be recommended in the sequence of regions to be recommended, determining the advertising placement strategy for the stores to be recommended in each region to be recommended in the sequence of regions to be recommended.

[0217] In some embodiments, the determining module 5552 is further configured to adjust the sequence of regions to be recommended based on the new regional recommendation degree to obtain an adjusted sequence of regions to be recommended; determine a first distance threshold corresponding to the region to be recommended based on the location of the intended object of the region to be recommended; filter the adjusted sequence of regions to be recommended based on the first distance threshold to obtain a filtered sequence of regions to be recommended; and determine an advertising placement strategy for the stores to be recommended in each region of the filtered sequence of regions to be recommended based on a preset strategy.

[0218] In some embodiments, the information recommendation device 555 described above further includes a processing module 5553.

[0219] The acquisition module 5551 is further configured to acquire the log data of the store to be recommended and the object transaction data of the store to be recommended; the processing module 5553 is configured to aggregate the log data and the object transaction data, and use the aggregation result as the payment measurement result corresponding to the advertising placement strategy.

[0220] In some embodiments, the acquisition module 5551 is further configured to acquire log data of the store to be recommended and offline access data of the store to be recommended; the processing module 5553 is further configured to perform a first association processing on the offline access data and the log data, and use the obtained first association result as the store visit measurement result corresponding to the advertising placement strategy.

[0221] In some embodiments, the acquisition module 5551 is further configured to acquire log data of the store to be recommended and online pre-payment transaction data of the store to be recommended; the processing module 5553 is further configured to perform a second association processing on the online pre-payment transaction data and the log data, and use the obtained second association result as the pre-payment measurement result corresponding to the advertising placement strategy.

[0222] It should be noted that the description of the apparatus in this application embodiment is similar to the description of the method embodiment above, and has similar beneficial effects as the method embodiment; therefore, it will not be repeated. For technical details not covered in the information recommendation apparatus provided in this application embodiment, please refer to... Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 ,or Figure 8 The meaning is understood in accordance with the description of any of the accompanying drawings.

[0223] This application provides a computer program product comprising a computer program or computer-executable instructions stored in a computer-readable storage medium. The processor of an electronic device reads the computer-executable instructions from the computer-readable storage medium and executes the computer-executable instructions, causing the electronic device to perform the information recommendation method described in this application.

[0224] This application provides a computer-readable storage medium storing computer-executable instructions or a computer program. When the computer-executable instructions or the computer program are executed by a processor, the processor will execute the information recommendation method provided in this application, for example, such as... Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 ,or Figure 8 The information recommendation method is shown.

[0225] In some embodiments, the computer-readable storage medium may be a memory such as ferroelectric random access memory (FRAM), ROM, programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); or it may be a device that includes one or any combination of the above-mentioned memories.

[0226] In some embodiments, computer-executable instructions may take the form of programs, software, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

[0227] As an example, computer-executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple co-located files (e.g., files that store one or more modules, subroutines, or code sections).

[0228] As an example, computer-executable instructions can be deployed to execute on a single electronic device, or on multiple electronic devices located at one location, or on multiple electronic devices distributed across multiple locations and interconnected via a communication network.

[0229] In summary, by obtaining multiple recommended areas for stores through the embodiments of this application, the potential market scope of the stores in geographical space can be clearly defined. This allows for targeted subsequent analysis of these recommended areas, avoiding blind promotion and improving resource utilization efficiency. Next, obtaining the target audience for the stores to be recommended helps determine the target customer types, providing clear guidance for evaluating the recommendation value of different areas and developing advertising strategies. Subsequently, by obtaining the access data of each recommended area, relevant data information of the objects within each area can be deeply understood, and the target audience for each recommended area can be determined based on the access data, identifying potential customers in each area and improving the accuracy of recommendations. Then, by comprehensively considering the target audience and potential audience in each recommended area, the regional recommendation level of each area can be determined, quantifying the potential value of different areas for stores and providing important reference for developing subsequent advertising strategies. Finally, determining the advertising strategy based on the regional recommendation level allows for the optimized allocation of advertising resources. Combining the regional recommendation level with the development of personalized advertising strategies for each area enables precise advertising placement, thereby improving the effectiveness of advertising and maximizing the store's marketing goals.

[0230] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. An information recommendation method, characterized in that, The method includes: Obtain multiple recommended areas for the stores to be recommended, and obtain the intended customers for the stores to be recommended; Obtain object access data for each of the regions to be recommended, and determine the objects to be recommended for each region based on the object access data for each region to be recommended; Based on the objects to be recommended in each region to be recommended and the intended objects in each region to be recommended, the regional recommendation degree of each region to be recommended is determined. Based on the regional recommendation level of each of the regions to be recommended, an advertising strategy for each of the regions to be recommended is determined for the stores to be recommended.

2. The method according to claim 1, characterized in that, The process of obtaining multiple recommendation areas for stores to be recommended includes: Obtain multiple candidate regions; Based on the location of the store to be recommended, the multiple candidate areas are filtered to obtain multiple areas to be recommended for the store.

3. The method according to claim 2, characterized in that, The step of filtering the multiple candidate areas based on the location of the store to be recommended, to obtain multiple recommended areas for the store to be recommended, includes: Perform at least one of the following processes: Based on the location of the store to be recommended, the administrative region to which the store to be recommended belongs is determined, and a candidate region that matches the administrative region to which the store to be recommended belongs is selected from the multiple candidate regions and used as the recommended region for the store to be recommended. Using the location of the store to be recommended as the center and a set distance as the radius, a first region is determined, and multiple candidate regions located in the first region are used as multiple recommendation regions for the store to be recommended.

4. The method according to claim 1, characterized in that, The process of obtaining the intended recipients of the recommended stores includes: Perform at least one of the following processes: The advertising conversion targets of the stores to be recommended are obtained as the target audience of the stores to be recommended; The online conversion targets of the stores to be recommended are obtained as the intended targets of the stores to be recommended; The offline visitors of the stores to be recommended are obtained as the intended visitors of the stores to be recommended.

5. The method according to claim 1, characterized in that, The process of determining the regional recommendation degree of each region based on the target objects and the intended objects of each region includes: For each of the regions to be recommended, the following processing is performed: Determine the first number of objects to be recommended in the region to be recommended; Determine a second number of intended objects for the region to be recommended; Obtain the regional recommendation score that is negatively correlated with the first number and positively correlated with the second number.

6. The method according to claim 1, characterized in that, The step of determining the advertising strategy for each store in each region to be recommended based on the regional recommendation score of each region to be recommended includes: Based on the regional recommendation degree of each region to be recommended, the multiple regions to be recommended are sorted in descending order to obtain a sequence of regions to be recommended. Determine the advertising strategy for each of the recommended regions in the recommended region sequence for the recommended stores.

7. The method according to claim 6, characterized in that, The step of sorting multiple regions to be recommended in descending order based on the regional recommendation score of each region to be recommended, to obtain a sequence of regions to be recommended, includes: From the multiple regions to be recommended, obtain the regions with a recommendation degree greater than zero; The regions to be recommended with a regional recommendation score greater than zero are sorted in descending order based on the regional recommendation score to obtain the descending order ranking result. When the number of regions to be recommended included in the descending sort result is less than a first quantity threshold, the descending sort result is taken as the sequence of regions to be recommended. When the number of regions to be recommended included in the descending sort result is not less than the first quantity threshold, the descending sort result is truncated based on the first quantity threshold to obtain the sequence of regions to be recommended.

8. The method according to claim 6, characterized in that, When the advertising strategy is a bidding strategy targeting the stores to be recommended, determining the advertising strategy for each of the stores to be recommended in the sequence of recommended regions includes: The recommended region sequence is divided into layers based on the layer ratio of the recommended region sequence to obtain the layer to which each recommended region belongs; For each region to be recommended, the bidding strategy corresponding to the level to which the region to be recommended belongs is used as the advertising strategy for the region to be recommended for the stores to be recommended.

9. The method according to claim 6, characterized in that, The intended objects of the region to be recommended have multiple object types, and the regional recommendation degree of the region to be recommended includes the regional sub-recommendation degree corresponding to each of the object types. The step of determining the advertising strategy for each of the recommended regions in the recommended region sequence for the recommended store includes: For each region to be recommended, the following processing is performed: obtain the weight corresponding to each object type; based on the weight of each object type, perform a weighted summation of the sub-recommendation scores of the region to be recommended and multiple regions corresponding to multiple object types to obtain the new region recommendation score of the region to be recommended. Based on the new regional recommendation degree of each of the regions to be recommended in the sequence of regions to be recommended, an advertising placement strategy for each of the regions to be recommended in the sequence of regions to be recommended is determined for the stores to be recommended.

10. The method according to claim 9, characterized in that, The step of determining the advertising strategy for each store in the proposed recommendation area sequence based on the new regional recommendation score of each proposed recommendation area in the proposed recommendation area sequence includes: The recommended region sequence is adjusted based on the new region recommendation degree to obtain the adjusted recommended region sequence; Based on the location of the intended object in the region to be recommended, a first distance threshold corresponding to the region to be recommended is determined; The adjusted region sequence to be recommended is filtered based on the first distance threshold to obtain the filtered region sequence to be recommended. Based on a preset strategy, the advertising strategy for each of the recommended regions in the filtered recommended region sequence is determined for the recommended stores.

11. The method according to claim 1, characterized in that, After determining the advertising strategy for each store in each region based on its regional recommendation score, the method further includes: Obtain the log data of the store to be recommended and the object transaction data of the store to be recommended; The log data and the object transaction data are aggregated, and the aggregation result is used as the payment metric result corresponding to the advertising delivery strategy.

12. The method according to claim 1, characterized in that, After determining the advertising strategy for each store in each region based on its regional recommendation score, the method further includes: Obtain the log data of the stores to be recommended and the offline access data of the stores to be recommended; The offline access data and the log data are subjected to a first association process, and the obtained first association result is used as the in-store measurement result corresponding to the advertising delivery strategy.

13. The method according to claim 1, characterized in that, After determining the advertising strategy for each store in each region based on its regional recommendation score, the method further includes: Obtain the log data of the stores to be recommended and the online prepayment transaction data with the stores to be recommended; The online prepayment transaction data and the log data are subjected to a second association process, and the resulting second association result is used as the prepayment measurement result corresponding to the advertising placement strategy.

14. An information recommendation device, characterized in that, The device includes: The acquisition module is used to acquire multiple areas of stores to be recommended and to acquire the intended users of the stores to be recommended. The determination module is used to acquire object access data for each of the regions to be recommended, and to determine the objects to be recommended for each region based on the object access data for each region to be recommended; The determining module is further configured to determine the regional recommendation degree of each region to be recommended based on the objects to be recommended in each region to be recommended and the intended objects in each region to be recommended; The determining module is further configured to determine the advertising strategy for the stores to be recommended in each of the recommended regions based on the regional recommendation degree of each region to be recommended.

15. An electronic device, characterized in that, include: Memory is used to store executable instructions or computer programs. A processor, when executing computer-executable instructions or computer programs stored in the memory, implements the information recommendation method according to any one of claims 1 to 13.

16. A computer-readable storage medium storing computer-executable instructions or a computer program, characterized in that, When the computer-executable instructions or computer program are executed by a processor, the information recommendation method according to any one of claims 1 to 13 is implemented.

17. A computer program product comprising computer-executable instructions or a computer program, characterized in that, When the computer-executable instructions or computer program are executed by a processor, the information recommendation method according to any one of claims 1 to 13 is implemented.

Images